diff --git a/client/xBRZ/Changelog.txt b/client/xBRZ/Changelog.txt new file mode 100644 index 000000000..4526b86b7 --- /dev/null +++ b/client/xBRZ/Changelog.txt @@ -0,0 +1,66 @@ +xBRZ 1.8 [2019-11-28] +--------------------- +Consider ARGB outside area as transparent +Fixed ARGB scaling issue on image borders + + +xBRZ 1.7 [2019-07-04] +--------------------- +Fixed asymmetric color distance +New parameter: "Center direction bias" + + +xBRZ 1.6 [2018-02-27] +--------------------- +Added bilinear scaling +Option to skip color buffer creation +Updated license info + + +xBRZ 1.5 [2017-08-07] +--------------------- +Added RGB conversion routines + + +xBRZ 1.4 [2015-07-25] +--------------------- +Added 6xBRZ scaler +Create color distance buffer lazily + + +xBRZ 1.3 [2015-04-03] +--------------------- +Improved ARGB performance by 15% +Fixed alpha channel gradient bug + + +xBRZ 1.2 [2014-11-21] +--------------------- +Further improved performance by over 30% + + +xBRZ 1.1 [2014-11-02] +--------------------- +Support images with alpha channel +Improved color analysis + + +xBRZ 1.0 [2013-02-11] +--------------------- +Fixed xBRZ scaler compiler issues for GCC + + +xBRZ 0.2 [2012-12-11] +--------------------- +Added 5xBRZ scaler +Optimized xBRZ scaler performance by factor 3 +Further improved image quality of xBRZ scaler + + +xBRZ 0.1 [2012-09-26] +--------------------- +Initial release: +- scale while preserving small image features +- support multithreading +- support 64-bit architectures +- support processing image slices diff --git a/client/xBRZ/License.txt b/client/xBRZ/License.txt new file mode 100644 index 000000000..ea09e3734 --- /dev/null +++ b/client/xBRZ/License.txt @@ -0,0 +1,621 @@ + GNU GENERAL PUBLIC LICENSE + Version 3, 29 June 2007 + + Copyright (C) 2007 Free Software Foundation, Inc. + Everyone is permitted to copy and distribute verbatim copies + of this license document, but changing it is not allowed. + + Preamble + + The GNU General Public License is a free, copyleft license for +software and other kinds of works. + + The licenses for most software and other practical works are designed +to take away your freedom to share and change the works. By contrast, +the GNU General Public License is intended to guarantee your freedom to +share and change all versions of a program--to make sure it remains free +software for all its users. We, the Free Software Foundation, use the +GNU General Public License for most of our software; it applies also to +any other work released this way by its authors. You can apply it to +your programs, too. + + When we speak of free software, we are referring to freedom, not +price. Our General Public Licenses are designed to make sure that you +have the freedom to distribute copies of free software (and charge for +them if you wish), that you receive source code or can get it if you +want it, that you can change the software or use pieces of it in new +free programs, and that you know you can do these things. + + To protect your rights, we need to prevent others from denying you +these rights or asking you to surrender the rights. Therefore, you have +certain responsibilities if you distribute copies of the software, or if +you modify it: responsibilities to respect the freedom of others. + + For example, if you distribute copies of such a program, whether +gratis or for a fee, you must pass on to the recipients the same +freedoms that you received. You must make sure that they, too, receive +or can get the source code. And you must show them these terms so they +know their rights. + + Developers that use the GNU GPL protect your rights with two steps: +(1) assert copyright on the software, and (2) offer you this License +giving you legal permission to copy, distribute and/or modify it. + + For the developers' and authors' protection, the GPL clearly explains +that there is no warranty for this free software. For both users' and +authors' sake, the GPL requires that modified versions be marked as +changed, so that their problems will not be attributed erroneously to +authors of previous versions. + + Some devices are designed to deny users access to install or run +modified versions of the software inside them, although the manufacturer +can do so. This is fundamentally incompatible with the aim of +protecting users' freedom to change the software. The systematic +pattern of such abuse occurs in the area of products for individuals to +use, which is precisely where it is most unacceptable. Therefore, we +have designed this version of the GPL to prohibit the practice for those +products. If such problems arise substantially in other domains, we +stand ready to extend this provision to those domains in future versions +of the GPL, as needed to protect the freedom of users. + + Finally, every program is threatened constantly by software patents. +States should not allow patents to restrict development and use of +software on general-purpose computers, but in those that do, we wish to +avoid the special danger that patents applied to a free program could +make it effectively proprietary. To prevent this, the GPL assures that +patents cannot be used to render the program non-free. + + The precise terms and conditions for copying, distribution and +modification follow. + + TERMS AND CONDITIONS + + 0. Definitions. + + "This License" refers to version 3 of the GNU General Public License. + + "Copyright" also means copyright-like laws that apply to other kinds of +works, such as semiconductor masks. + + "The Program" refers to any copyrightable work licensed under this +License. Each licensee is addressed as "you". "Licensees" and +"recipients" may be individuals or organizations. + + To "modify" a work means to copy from or adapt all or part of the work +in a fashion requiring copyright permission, other than the making of an +exact copy. The resulting work is called a "modified version" of the +earlier work or a work "based on" the earlier work. + + A "covered work" means either the unmodified Program or a work based +on the Program. + + To "propagate" a work means to do anything with it that, without +permission, would make you directly or secondarily liable for +infringement under applicable copyright law, except executing it on a +computer or modifying a private copy. Propagation includes copying, +distribution (with or without modification), making available to the +public, and in some countries other activities as well. + + To "convey" a work means any kind of propagation that enables other +parties to make or receive copies. Mere interaction with a user through +a computer network, with no transfer of a copy, is not conveying. + + An interactive user interface displays "Appropriate Legal Notices" +to the extent that it includes a convenient and prominently visible +feature that (1) displays an appropriate copyright notice, and (2) +tells the user that there is no warranty for the work (except to the +extent that warranties are provided), that licensees may convey the +work under this License, and how to view a copy of this License. If +the interface presents a list of user commands or options, such as a +menu, a prominent item in the list meets this criterion. + + 1. Source Code. + + The "source code" for a work means the preferred form of the work +for making modifications to it. "Object code" means any non-source +form of a work. + + A "Standard Interface" means an interface that either is an official +standard defined by a recognized standards body, or, in the case of +interfaces specified for a particular programming language, one that +is widely used among developers working in that language. + + The "System Libraries" of an executable work include anything, other +than the work as a whole, that (a) is included in the normal form of +packaging a Major Component, but which is not part of that Major +Component, and (b) serves only to enable use of the work with that +Major Component, or to implement a Standard Interface for which an +implementation is available to the public in source code form. A +"Major Component", in this context, means a major essential component +(kernel, window system, and so on) of the specific operating system +(if any) on which the executable work runs, or a compiler used to +produce the work, or an object code interpreter used to run it. + + The "Corresponding Source" for a work in object code form means all +the source code needed to generate, install, and (for an executable +work) run the object code and to modify the work, including scripts to +control those activities. However, it does not include the work's +System Libraries, or general-purpose tools or generally available free +programs which are used unmodified in performing those activities but +which are not part of the work. For example, Corresponding Source +includes interface definition files associated with source files for +the work, and the source code for shared libraries and dynamically +linked subprograms that the work is specifically designed to require, +such as by intimate data communication or control flow between those +subprograms and other parts of the work. + + The Corresponding Source need not include anything that users +can regenerate automatically from other parts of the Corresponding +Source. + + The Corresponding Source for a work in source code form is that +same work. + + 2. Basic Permissions. + + All rights granted under this License are granted for the term of +copyright on the Program, and are irrevocable provided the stated +conditions are met. This License explicitly affirms your unlimited +permission to run the unmodified Program. The output from running a +covered work is covered by this License only if the output, given its +content, constitutes a covered work. This License acknowledges your +rights of fair use or other equivalent, as provided by copyright law. + + You may make, run and propagate covered works that you do not +convey, without conditions so long as your license otherwise remains +in force. You may convey covered works to others for the sole purpose +of having them make modifications exclusively for you, or provide you +with facilities for running those works, provided that you comply with +the terms of this License in conveying all material for which you do +not control copyright. Those thus making or running the covered works +for you must do so exclusively on your behalf, under your direction +and control, on terms that prohibit them from making any copies of +your copyrighted material outside their relationship with you. + + Conveying under any other circumstances is permitted solely under +the conditions stated below. Sublicensing is not allowed; section 10 +makes it unnecessary. + + 3. Protecting Users' Legal Rights From Anti-Circumvention Law. + + No covered work shall be deemed part of an effective technological +measure under any applicable law fulfilling obligations under article +11 of the WIPO copyright treaty adopted on 20 December 1996, or +similar laws prohibiting or restricting circumvention of such +measures. + + When you convey a covered work, you waive any legal power to forbid +circumvention of technological measures to the extent such circumvention +is effected by exercising rights under this License with respect to +the covered work, and you disclaim any intention to limit operation or +modification of the work as a means of enforcing, against the work's +users, your or third parties' legal rights to forbid circumvention of +technological measures. + + 4. Conveying Verbatim Copies. + + You may convey verbatim copies of the Program's source code as you +receive it, in any medium, provided that you conspicuously and +appropriately publish on each copy an appropriate copyright notice; +keep intact all notices stating that this License and any +non-permissive terms added in accord with section 7 apply to the code; +keep intact all notices of the absence of any warranty; and give all +recipients a copy of this License along with the Program. + + You may charge any price or no price for each copy that you convey, +and you may offer support or warranty protection for a fee. + + 5. Conveying Modified Source Versions. + + You may convey a work based on the Program, or the modifications to +produce it from the Program, in the form of source code under the +terms of section 4, provided that you also meet all of these conditions: + + a) The work must carry prominent notices stating that you modified + it, and giving a relevant date. + + b) The work must carry prominent notices stating that it is + released under this License and any conditions added under section + 7. This requirement modifies the requirement in section 4 to + "keep intact all notices". + + c) You must license the entire work, as a whole, under this + License to anyone who comes into possession of a copy. This + License will therefore apply, along with any applicable section 7 + additional terms, to the whole of the work, and all its parts, + regardless of how they are packaged. This License gives no + permission to license the work in any other way, but it does not + invalidate such permission if you have separately received it. + + d) If the work has interactive user interfaces, each must display + Appropriate Legal Notices; however, if the Program has interactive + interfaces that do not display Appropriate Legal Notices, your + work need not make them do so. + + A compilation of a covered work with other separate and independent +works, which are not by their nature extensions of the covered work, +and which are not combined with it such as to form a larger program, +in or on a volume of a storage or distribution medium, is called an +"aggregate" if the compilation and its resulting copyright are not +used to limit the access or legal rights of the compilation's users +beyond what the individual works permit. Inclusion of a covered work +in an aggregate does not cause this License to apply to the other +parts of the aggregate. + + 6. Conveying Non-Source Forms. + + You may convey a covered work in object code form under the terms +of sections 4 and 5, provided that you also convey the +machine-readable Corresponding Source under the terms of this License, +in one of these ways: + + a) Convey the object code in, or embodied in, a physical product + (including a physical distribution medium), accompanied by the + Corresponding Source fixed on a durable physical medium + customarily used for software interchange. + + b) Convey the object code in, or embodied in, a physical product + (including a physical distribution medium), accompanied by a + written offer, valid for at least three years and valid for as + long as you offer spare parts or customer support for that product + model, to give anyone who possesses the object code either (1) a + copy of the Corresponding Source for all the software in the + product that is covered by this License, on a durable physical + medium customarily used for software interchange, for a price no + more than your reasonable cost of physically performing this + conveying of source, or (2) access to copy the + Corresponding Source from a network server at no charge. + + c) Convey individual copies of the object code with a copy of the + written offer to provide the Corresponding Source. This + alternative is allowed only occasionally and noncommercially, and + only if you received the object code with such an offer, in accord + with subsection 6b. + + d) Convey the object code by offering access from a designated + place (gratis or for a charge), and offer equivalent access to the + Corresponding Source in the same way through the same place at no + further charge. You need not require recipients to copy the + Corresponding Source along with the object code. If the place to + copy the object code is a network server, the Corresponding Source + may be on a different server (operated by you or a third party) + that supports equivalent copying facilities, provided you maintain + clear directions next to the object code saying where to find the + Corresponding Source. Regardless of what server hosts the + Corresponding Source, you remain obligated to ensure that it is + available for as long as needed to satisfy these requirements. + + e) Convey the object code using peer-to-peer transmission, provided + you inform other peers where the object code and Corresponding + Source of the work are being offered to the general public at no + charge under subsection 6d. + + A separable portion of the object code, whose source code is excluded +from the Corresponding Source as a System Library, need not be +included in conveying the object code work. + + A "User Product" is either (1) a "consumer product", which means any +tangible personal property which is normally used for personal, family, +or household purposes, or (2) anything designed or sold for incorporation +into a dwelling. In determining whether a product is a consumer product, +doubtful cases shall be resolved in favor of coverage. For a particular +product received by a particular user, "normally used" refers to a +typical or common use of that class of product, regardless of the status +of the particular user or of the way in which the particular user +actually uses, or expects or is expected to use, the product. A product +is a consumer product regardless of whether the product has substantial +commercial, industrial or non-consumer uses, unless such uses represent +the only significant mode of use of the product. + + "Installation Information" for a User Product means any methods, +procedures, authorization keys, or other information required to install +and execute modified versions of a covered work in that User Product from +a modified version of its Corresponding Source. The information must +suffice to ensure that the continued functioning of the modified object +code is in no case prevented or interfered with solely because +modification has been made. + + If you convey an object code work under this section in, or with, or +specifically for use in, a User Product, and the conveying occurs as +part of a transaction in which the right of possession and use of the +User Product is transferred to the recipient in perpetuity or for a +fixed term (regardless of how the transaction is characterized), the +Corresponding Source conveyed under this section must be accompanied +by the Installation Information. But this requirement does not apply +if neither you nor any third party retains the ability to install +modified object code on the User Product (for example, the work has +been installed in ROM). + + The requirement to provide Installation Information does not include a +requirement to continue to provide support service, warranty, or updates +for a work that has been modified or installed by the recipient, or for +the User Product in which it has been modified or installed. Access to a +network may be denied when the modification itself materially and +adversely affects the operation of the network or violates the rules and +protocols for communication across the network. + + Corresponding Source conveyed, and Installation Information provided, +in accord with this section must be in a format that is publicly +documented (and with an implementation available to the public in +source code form), and must require no special password or key for +unpacking, reading or copying. + + 7. Additional Terms. + + "Additional permissions" are terms that supplement the terms of this +License by making exceptions from one or more of its conditions. +Additional permissions that are applicable to the entire Program shall +be treated as though they were included in this License, to the extent +that they are valid under applicable law. If additional permissions +apply only to part of the Program, that part may be used separately +under those permissions, but the entire Program remains governed by +this License without regard to the additional permissions. + + When you convey a copy of a covered work, you may at your option +remove any additional permissions from that copy, or from any part of +it. (Additional permissions may be written to require their own +removal in certain cases when you modify the work.) You may place +additional permissions on material, added by you to a covered work, +for which you have or can give appropriate copyright permission. + + Notwithstanding any other provision of this License, for material you +add to a covered work, you may (if authorized by the copyright holders of +that material) supplement the terms of this License with terms: + + a) Disclaiming warranty or limiting liability differently from the + terms of sections 15 and 16 of this License; or + + b) Requiring preservation of specified reasonable legal notices or + author attributions in that material or in the Appropriate Legal + Notices displayed by works containing it; or + + c) Prohibiting misrepresentation of the origin of that material, or + requiring that modified versions of such material be marked in + reasonable ways as different from the original version; or + + d) Limiting the use for publicity purposes of names of licensors or + authors of the material; or + + e) Declining to grant rights under trademark law for use of some + trade names, trademarks, or service marks; or + + f) Requiring indemnification of licensors and authors of that + material by anyone who conveys the material (or modified versions of + it) with contractual assumptions of liability to the recipient, for + any liability that these contractual assumptions directly impose on + those licensors and authors. + + All other non-permissive additional terms are considered "further +restrictions" within the meaning of section 10. If the Program as you +received it, or any part of it, contains a notice stating that it is +governed by this License along with a term that is a further +restriction, you may remove that term. If a license document contains +a further restriction but permits relicensing or conveying under this +License, you may add to a covered work material governed by the terms +of that license document, provided that the further restriction does +not survive such relicensing or conveying. + + If you add terms to a covered work in accord with this section, you +must place, in the relevant source files, a statement of the +additional terms that apply to those files, or a notice indicating +where to find the applicable terms. + + Additional terms, permissive or non-permissive, may be stated in the +form of a separately written license, or stated as exceptions; +the above requirements apply either way. + + 8. Termination. + + You may not propagate or modify a covered work except as expressly +provided under this License. Any attempt otherwise to propagate or +modify it is void, and will automatically terminate your rights under +this License (including any patent licenses granted under the third +paragraph of section 11). + + However, if you cease all violation of this License, then your +license from a particular copyright holder is reinstated (a) +provisionally, unless and until the copyright holder explicitly and +finally terminates your license, and (b) permanently, if the copyright +holder fails to notify you of the violation by some reasonable means +prior to 60 days after the cessation. + + Moreover, your license from a particular copyright holder is +reinstated permanently if the copyright holder notifies you of the +violation by some reasonable means, this is the first time you have +received notice of violation of this License (for any work) from that +copyright holder, and you cure the violation prior to 30 days after +your receipt of the notice. + + Termination of your rights under this section does not terminate the +licenses of parties who have received copies or rights from you under +this License. If your rights have been terminated and not permanently +reinstated, you do not qualify to receive new licenses for the same +material under section 10. + + 9. Acceptance Not Required for Having Copies. + + You are not required to accept this License in order to receive or +run a copy of the Program. Ancillary propagation of a covered work +occurring solely as a consequence of using peer-to-peer transmission +to receive a copy likewise does not require acceptance. However, +nothing other than this License grants you permission to propagate or +modify any covered work. These actions infringe copyright if you do +not accept this License. Therefore, by modifying or propagating a +covered work, you indicate your acceptance of this License to do so. + + 10. Automatic Licensing of Downstream Recipients. + + Each time you convey a covered work, the recipient automatically +receives a license from the original licensors, to run, modify and +propagate that work, subject to this License. You are not responsible +for enforcing compliance by third parties with this License. + + An "entity transaction" is a transaction transferring control of an +organization, or substantially all assets of one, or subdividing an +organization, or merging organizations. If propagation of a covered +work results from an entity transaction, each party to that +transaction who receives a copy of the work also receives whatever +licenses to the work the party's predecessor in interest had or could +give under the previous paragraph, plus a right to possession of the +Corresponding Source of the work from the predecessor in interest, if +the predecessor has it or can get it with reasonable efforts. + + You may not impose any further restrictions on the exercise of the +rights granted or affirmed under this License. For example, you may +not impose a license fee, royalty, or other charge for exercise of +rights granted under this License, and you may not initiate litigation +(including a cross-claim or counterclaim in a lawsuit) alleging that +any patent claim is infringed by making, using, selling, offering for +sale, or importing the Program or any portion of it. + + 11. Patents. + + A "contributor" is a copyright holder who authorizes use under this +License of the Program or a work on which the Program is based. The +work thus licensed is called the contributor's "contributor version". + + A contributor's "essential patent claims" are all patent claims +owned or controlled by the contributor, whether already acquired or +hereafter acquired, that would be infringed by some manner, permitted +by this License, of making, using, or selling its contributor version, +but do not include claims that would be infringed only as a +consequence of further modification of the contributor version. For +purposes of this definition, "control" includes the right to grant +patent sublicenses in a manner consistent with the requirements of +this License. + + Each contributor grants you a non-exclusive, worldwide, royalty-free +patent license under the contributor's essential patent claims, to +make, use, sell, offer for sale, import and otherwise run, modify and +propagate the contents of its contributor version. + + In the following three paragraphs, a "patent license" is any express +agreement or commitment, however denominated, not to enforce a patent +(such as an express permission to practice a patent or covenant not to +sue for patent infringement). To "grant" such a patent license to a +party means to make such an agreement or commitment not to enforce a +patent against the party. + + If you convey a covered work, knowingly relying on a patent license, +and the Corresponding Source of the work is not available for anyone +to copy, free of charge and under the terms of this License, through a +publicly available network server or other readily accessible means, +then you must either (1) cause the Corresponding Source to be so +available, or (2) arrange to deprive yourself of the benefit of the +patent license for this particular work, or (3) arrange, in a manner +consistent with the requirements of this License, to extend the patent +license to downstream recipients. "Knowingly relying" means you have +actual knowledge that, but for the patent license, your conveying the +covered work in a country, or your recipient's use of the covered work +in a country, would infringe one or more identifiable patents in that +country that you have reason to believe are valid. + + If, pursuant to or in connection with a single transaction or +arrangement, you convey, or propagate by procuring conveyance of, a +covered work, and grant a patent license to some of the parties +receiving the covered work authorizing them to use, propagate, modify +or convey a specific copy of the covered work, then the patent license +you grant is automatically extended to all recipients of the covered +work and works based on it. + + A patent license is "discriminatory" if it does not include within +the scope of its coverage, prohibits the exercise of, or is +conditioned on the non-exercise of one or more of the rights that are +specifically granted under this License. You may not convey a covered +work if you are a party to an arrangement with a third party that is +in the business of distributing software, under which you make payment +to the third party based on the extent of your activity of conveying +the work, and under which the third party grants, to any of the +parties who would receive the covered work from you, a discriminatory +patent license (a) in connection with copies of the covered work +conveyed by you (or copies made from those copies), or (b) primarily +for and in connection with specific products or compilations that +contain the covered work, unless you entered into that arrangement, +or that patent license was granted, prior to 28 March 2007. + + Nothing in this License shall be construed as excluding or limiting +any implied license or other defenses to infringement that may +otherwise be available to you under applicable patent law. + + 12. No Surrender of Others' Freedom. + + If conditions are imposed on you (whether by court order, agreement or +otherwise) that contradict the conditions of this License, they do not +excuse you from the conditions of this License. If you cannot convey a +covered work so as to satisfy simultaneously your obligations under this +License and any other pertinent obligations, then as a consequence you may +not convey it at all. For example, if you agree to terms that obligate you +to collect a royalty for further conveying from those to whom you convey +the Program, the only way you could satisfy both those terms and this +License would be to refrain entirely from conveying the Program. + + 13. Use with the GNU Affero General Public License. + + Notwithstanding any other provision of this License, you have +permission to link or combine any covered work with a work licensed +under version 3 of the GNU Affero General Public License into a single +combined work, and to convey the resulting work. The terms of this +License will continue to apply to the part which is the covered work, +but the special requirements of the GNU Affero General Public License, +section 13, concerning interaction through a network will apply to the +combination as such. + + 14. Revised Versions of this License. + + The Free Software Foundation may publish revised and/or new versions of +the GNU General Public License from time to time. Such new versions will +be similar in spirit to the present version, but may differ in detail to +address new problems or concerns. + + Each version is given a distinguishing version number. If the +Program specifies that a certain numbered version of the GNU General +Public License "or any later version" applies to it, you have the +option of following the terms and conditions either of that numbered +version or of any later version published by the Free Software +Foundation. If the Program does not specify a version number of the +GNU General Public License, you may choose any version ever published +by the Free Software Foundation. + + If the Program specifies that a proxy can decide which future +versions of the GNU General Public License can be used, that proxy's +public statement of acceptance of a version permanently authorizes you +to choose that version for the Program. + + Later license versions may give you additional or different +permissions. However, no additional obligations are imposed on any +author or copyright holder as a result of your choosing to follow a +later version. + + 15. Disclaimer of Warranty. + + THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY +APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT +HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY +OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, +THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR +PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM +IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF +ALL NECESSARY SERVICING, REPAIR OR CORRECTION. + + 16. Limitation of Liability. + + IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING +WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS +THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY +GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE +USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF +DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD +PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), +EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF +SUCH DAMAGES. + + 17. Interpretation of Sections 15 and 16. + + If the disclaimer of warranty and limitation of liability provided +above cannot be given local legal effect according to their terms, +reviewing courts shall apply local law that most closely approximates +an absolute waiver of all civil liability in connection with the +Program, unless a warranty or assumption of liability accompanies a +copy of the Program in return for a fee. + + END OF TERMS AND CONDITIONS diff --git a/client/xBRZ/xbrz.cpp b/client/xBRZ/xbrz.cpp new file mode 100644 index 000000000..a48ff83eb --- /dev/null +++ b/client/xBRZ/xbrz.cpp @@ -0,0 +1,1356 @@ +// **************************************************************************** +// * This file is part of the xBRZ project. It is distributed under * +// * GNU General Public License: https://www.gnu.org/licenses/gpl-3.0 * +// * Copyright (C) Zenju (zenju AT gmx DOT de) - All Rights Reserved * +// * * +// * Additionally and as a special exception, the author gives permission * +// * to link the code of this program with the following libraries * +// * (or with modified versions that use the same licenses), and distribute * +// * linked combinations including the two: MAME, FreeFileSync, Snes9x, ePSXe * +// * You must obey the GNU General Public License in all respects for all of * +// * the code used other than MAME, FreeFileSync, Snes9x, ePSXe. * +// * If you modify this file, you may extend this exception to your version * +// * of the file, but you are not obligated to do so. If you do not wish to * +// * do so, delete this exception statement from your version. * +// **************************************************************************** + +#include "xbrz.h" +#include +#include +#include +#include //std::sqrt +#include "xbrz_tools.h" + +using namespace xbrz; + + +namespace +{ +template inline +uint32_t gradientRGB(uint32_t pixFront, uint32_t pixBack) //blend front color with opacity M / N over opaque background: https://en.wikipedia.org/wiki/Alpha_compositing#Alpha_blending +{ + static_assert(0 < M && M < N && N <= 1000); + + auto calcColor = [](unsigned char colFront, unsigned char colBack) -> unsigned char { return (colFront * M + colBack * (N - M)) / N; }; + + return makePixel(calcColor(getRed (pixFront), getRed (pixBack)), + calcColor(getGreen(pixFront), getGreen(pixBack)), + calcColor(getBlue (pixFront), getBlue (pixBack))); +} + + +template inline +uint32_t gradientARGB(uint32_t pixFront, uint32_t pixBack) //find intermediate color between two colors with alpha channels (=> NO alpha blending!!!) +{ + static_assert(0 < M && M < N && N <= 1000); + + const unsigned int weightFront = getAlpha(pixFront) * M; + const unsigned int weightBack = getAlpha(pixBack) * (N - M); + const unsigned int weightSum = weightFront + weightBack; + if (weightSum == 0) + return 0; + + auto calcColor = [=](unsigned char colFront, unsigned char colBack) + { + return static_cast((colFront * weightFront + colBack * weightBack) / weightSum); + }; + + return makePixel(static_cast(weightSum / N), + calcColor(getRed (pixFront), getRed (pixBack)), + calcColor(getGreen(pixFront), getGreen(pixBack)), + calcColor(getBlue (pixFront), getBlue (pixBack))); +} + + +//inline +//double fastSqrt(double n) +//{ +// __asm //speeds up xBRZ by about 9% compared to std::sqrt which internally uses the same assembler instructions but adds some "fluff" +// { +// fld n +// fsqrt +// } +//} +// + + +#ifdef _MSC_VER + #define FORCE_INLINE __forceinline +#elif defined __GNUC__ + #define FORCE_INLINE __attribute__((always_inline)) inline +#else + #define FORCE_INLINE inline +#endif + + +enum RotationDegree //clock-wise +{ + ROT_0, + ROT_90, + ROT_180, + ROT_270 +}; + +//calculate input matrix coordinates after rotation at compile time +template +struct MatrixRotation; + +template +struct MatrixRotation +{ + static const size_t I_old = I; + static const size_t J_old = J; +}; + +template //(i, j) = (row, col) indices, N = size of (square) matrix +struct MatrixRotation +{ + static const size_t I_old = N - 1 - MatrixRotation(rotDeg - 1), I, J, N>::J_old; //old coordinates before rotation! + static const size_t J_old = MatrixRotation(rotDeg - 1), I, J, N>::I_old; // +}; + + +template +class OutputMatrix +{ +public: + OutputMatrix(uint32_t* out, int outWidth) : //access matrix area, top-left at position "out" for image with given width + out_(out), + outWidth_(outWidth) {} + + template + uint32_t& ref() const + { + static const size_t I_old = MatrixRotation::I_old; + static const size_t J_old = MatrixRotation::J_old; + return *(out_ + J_old + I_old * outWidth_); + } + +private: + uint32_t* out_; + const int outWidth_; +}; + + +template inline +T square(T value) { return value * value; } + + +#if 0 +inline +double distRGB(uint32_t pix1, uint32_t pix2) +{ + const double r_diff = static_cast(getRed (pix1)) - getRed (pix2); + const double g_diff = static_cast(getGreen(pix1)) - getGreen(pix2); + const double b_diff = static_cast(getBlue (pix1)) - getBlue (pix2); + + //euklidean RGB distance + return std::sqrt(square(r_diff) + square(g_diff) + square(b_diff)); +} +#endif + + +inline +double distYCbCr(uint32_t pix1, uint32_t pix2, double lumaWeight) +{ + //https://en.wikipedia.org/wiki/YCbCr#ITU-R_BT.601_conversion + //YCbCr conversion is a matrix multiplication => take advantage of linearity by subtracting first! + const int r_diff = static_cast(getRed (pix1)) - getRed (pix2); //we may delay division by 255 to after matrix multiplication + const int g_diff = static_cast(getGreen(pix1)) - getGreen(pix2); // + const int b_diff = static_cast(getBlue (pix1)) - getBlue (pix2); //substraction for int is noticeable faster than for double! + + //const double k_b = 0.0722; //ITU-R BT.709 conversion + //const double k_r = 0.2126; // + const double k_b = 0.0593; //ITU-R BT.2020 conversion + const double k_r = 0.2627; // + const double k_g = 1 - k_b - k_r; + + const double scale_b = 0.5 / (1 - k_b); + const double scale_r = 0.5 / (1 - k_r); + + const double y = k_r * r_diff + k_g * g_diff + k_b * b_diff; //[!], analog YCbCr! + const double c_b = scale_b * (b_diff - y); + const double c_r = scale_r * (r_diff - y); + + //we skip division by 255 to have similar range like other distance functions + return std::sqrt(square(lumaWeight * y) + square(c_b) + square(c_r)); +} + + +inline +double distYCbCrBuffered(uint32_t pix1, uint32_t pix2) +{ + //30% perf boost compared to plain distYCbCr()! + //consumes 64 MB memory; using double is only 2% faster, but takes 128 MB + static const std::vector diffToDist = [] + { + std::vector tmp; + + for (uint32_t i = 0; i < 256 * 256 * 256; ++i) //startup time: 114 ms on Intel Core i5 (four cores) + { + const int r_diff = static_cast(getByte<2>(i)) * 2; + const int g_diff = static_cast(getByte<1>(i)) * 2; + const int b_diff = static_cast(getByte<0>(i)) * 2; + + const double k_b = 0.0593; //ITU-R BT.2020 conversion + const double k_r = 0.2627; // + const double k_g = 1 - k_b - k_r; + + const double scale_b = 0.5 / (1 - k_b); + const double scale_r = 0.5 / (1 - k_r); + + const double y = k_r * r_diff + k_g * g_diff + k_b * b_diff; //[!], analog YCbCr! + const double c_b = scale_b * (b_diff - y); + const double c_r = scale_r * (r_diff - y); + + tmp.push_back(static_cast(std::sqrt(square(y) + square(c_b) + square(c_r)))); + } + return tmp; + }(); + + //if (pix1 == pix2) -> 8% perf degradation! + // return 0; + //if (pix1 < pix2) + // std::swap(pix1, pix2); -> 30% perf degradation!!! + + const int r_diff = static_cast(getRed (pix1)) - getRed (pix2); + const int g_diff = static_cast(getGreen(pix1)) - getGreen(pix2); + const int b_diff = static_cast(getBlue (pix1)) - getBlue (pix2); + + const size_t index = (static_cast(r_diff / 2) << 16) | //slightly reduce precision (division by 2) to squeeze value into single byte + (static_cast(g_diff / 2) << 8) | + (static_cast(b_diff / 2)); + +#if 0 //attention: the following calculation creates an asymmetric color distance!!! (e.g. r_diff=46 will be unpacked as 45, but r_diff=-46 unpacks to -47 + const size_t index = (((r_diff + 0xFF) / 2) << 16) | //slightly reduce precision (division by 2) to squeeze value into single byte + (((g_diff + 0xFF) / 2) << 8) | + (( b_diff + 0xFF) / 2); +#endif + return diffToDist[index]; +} + + +#if defined _MSC_VER && !defined NDEBUG + const int debugPixelX = -1; + const int debugPixelY = 58; + + thread_local bool breakIntoDebugger = false; +#endif + + +enum BlendType +{ + BLEND_NONE = 0, + BLEND_NORMAL, //a normal indication to blend + BLEND_DOMINANT, //a strong indication to blend + //attention: BlendType must fit into the value range of 2 bit!!! +}; + +struct BlendResult +{ + BlendType + /**/blend_f, blend_g, + /**/blend_j, blend_k; +}; + + +struct Kernel_3x3 +{ + uint32_t + a, b, c, + d, e, f, + g, h, i; +}; + +struct Kernel_4x4 //kernel for preprocessing step +{ + uint32_t + a, b, c, // + e, f, g, // support reinterpret_cast from Kernel_4x4 => Kernel_3x3 + i, j, k, // + m, n, o, + d, h, l, p; +}; + +/* input kernel area naming convention: +----------------- +| A | B | C | D | +|---|---|---|---| +| E | F | G | H | evaluate the four corners between F, G, J, K +|---|---|---|---| input pixel is at position F +| I | J | K | L | +|---|---|---|---| +| M | N | O | P | +----------------- +*/ +template +FORCE_INLINE //detect blend direction +BlendResult preProcessCorners(const Kernel_4x4& ker, const xbrz::ScalerCfg& cfg) //result: F, G, J, K corners of "GradientType" +{ +#if defined _MSC_VER && !defined NDEBUG + if (breakIntoDebugger) + __debugbreak(); //__asm int 3; +#endif + + BlendResult result = {}; + + if ((ker.f == ker.g && + ker.j == ker.k) || + (ker.f == ker.j && + ker.g == ker.k)) + return result; + + auto dist = [&](uint32_t pix1, uint32_t pix2) { return ColorDistance::dist(pix1, pix2, cfg.luminanceWeight); }; + + double jg = dist(ker.i, ker.f) + dist(ker.f, ker.c) + dist(ker.n, ker.k) + dist(ker.k, ker.h) + cfg.centerDirectionBias * dist(ker.j, ker.g); + double fk = dist(ker.e, ker.j) + dist(ker.j, ker.o) + dist(ker.b, ker.g) + dist(ker.g, ker.l) + cfg.centerDirectionBias * dist(ker.f, ker.k); + + if (jg < fk) //test sample: 70% of values max(jg, fk) / min(jg, fk) are between 1.1 and 3.7 with median being 1.8 + { + const bool dominantGradient = cfg.dominantDirectionThreshold * jg < fk; + if (ker.f != ker.g && ker.f != ker.j) + result.blend_f = dominantGradient ? BLEND_DOMINANT : BLEND_NORMAL; + + if (ker.k != ker.j && ker.k != ker.g) + result.blend_k = dominantGradient ? BLEND_DOMINANT : BLEND_NORMAL; + } + else if (fk < jg) + { + const bool dominantGradient = cfg.dominantDirectionThreshold * fk < jg; + if (ker.j != ker.f && ker.j != ker.k) + result.blend_j = dominantGradient ? BLEND_DOMINANT : BLEND_NORMAL; + + if (ker.g != ker.f && ker.g != ker.k) + result.blend_g = dominantGradient ? BLEND_DOMINANT : BLEND_NORMAL; + } + return result; +} + +#define DEF_GETTER(x) template uint32_t inline get_##x(const Kernel_3x3& ker) { return ker.x; } +//we cannot and NEED NOT write "ker.##x" since ## concatenates preprocessor tokens but "." is not a token +DEF_GETTER(a) DEF_GETTER(b) DEF_GETTER(c) +DEF_GETTER(d) DEF_GETTER(e) DEF_GETTER(f) +DEF_GETTER(g) DEF_GETTER(h) DEF_GETTER(i) +#undef DEF_GETTER + +#define DEF_GETTER(x, y) template <> inline uint32_t get_##x(const Kernel_3x3& ker) { return ker.y; } +DEF_GETTER(a, g) DEF_GETTER(b, d) DEF_GETTER(c, a) +DEF_GETTER(d, h) DEF_GETTER(e, e) DEF_GETTER(f, b) +DEF_GETTER(g, i) DEF_GETTER(h, f) DEF_GETTER(i, c) +#undef DEF_GETTER + +#define DEF_GETTER(x, y) template <> inline uint32_t get_##x(const Kernel_3x3& ker) { return ker.y; } +DEF_GETTER(a, i) DEF_GETTER(b, h) DEF_GETTER(c, g) +DEF_GETTER(d, f) DEF_GETTER(e, e) DEF_GETTER(f, d) +DEF_GETTER(g, c) DEF_GETTER(h, b) DEF_GETTER(i, a) +#undef DEF_GETTER + +#define DEF_GETTER(x, y) template <> inline uint32_t get_##x(const Kernel_3x3& ker) { return ker.y; } +DEF_GETTER(a, c) DEF_GETTER(b, f) DEF_GETTER(c, i) +DEF_GETTER(d, b) DEF_GETTER(e, e) DEF_GETTER(f, h) +DEF_GETTER(g, a) DEF_GETTER(h, d) DEF_GETTER(i, g) +#undef DEF_GETTER + + +//compress four blend types into a single byte +//inline BlendType getTopL (unsigned char b) { return static_cast(0x3 & b); } +inline BlendType getTopR (unsigned char b) { return static_cast(0x3 & (b >> 2)); } +inline BlendType getBottomR(unsigned char b) { return static_cast(0x3 & (b >> 4)); } +inline BlendType getBottomL(unsigned char b) { return static_cast(0x3 & (b >> 6)); } + +inline void clearAddTopL(unsigned char& b, BlendType bt) { b = static_cast(bt); } +inline void addTopR (unsigned char& b, BlendType bt) { b |= (bt << 2); } //buffer is assumed to be initialized before preprocessing! +inline void addBottomR (unsigned char& b, BlendType bt) { b |= (bt << 4); } //e.g. via clearAddTopL() +inline void addBottomL (unsigned char& b, BlendType bt) { b |= (bt << 6); } // + +inline bool blendingNeeded(unsigned char b) +{ + static_assert(BLEND_NONE == 0); + return b != 0; +} + +template inline +unsigned char rotateBlendInfo(unsigned char b) { return b; } +template <> inline unsigned char rotateBlendInfo(unsigned char b) { return ((b << 2) | (b >> 6)) & 0xff; } +template <> inline unsigned char rotateBlendInfo(unsigned char b) { return ((b << 4) | (b >> 4)) & 0xff; } +template <> inline unsigned char rotateBlendInfo(unsigned char b) { return ((b << 6) | (b >> 2)) & 0xff; } + + +/* input kernel area naming convention: +------------- +| A | B | C | +|---|---|---| +| D | E | F | input pixel is at position E +|---|---|---| +| G | H | I | +------------- +*/ +template +FORCE_INLINE //perf: quite worth it! +void blendPixel(const Kernel_3x3& ker, + uint32_t* target, int trgWidth, + unsigned char blendInfo, //result of preprocessing all four corners of pixel "e" + const xbrz::ScalerCfg& cfg) +{ + //#define a get_a(ker) +#define b get_b(ker) +#define c get_c(ker) +#define d get_d(ker) +#define e get_e(ker) +#define f get_f(ker) +#define g get_g(ker) +#define h get_h(ker) +#define i get_i(ker) + +#if defined _MSC_VER && !defined NDEBUG + if (breakIntoDebugger) + __debugbreak(); //__asm int 3; +#endif + + const unsigned char blend = rotateBlendInfo(blendInfo); + + if (getBottomR(blend) >= BLEND_NORMAL) + { + auto eq = [&](uint32_t pix1, uint32_t pix2) { return ColorDistance::dist(pix1, pix2, cfg.luminanceWeight) < cfg.equalColorTolerance; }; + auto dist = [&](uint32_t pix1, uint32_t pix2) { return ColorDistance::dist(pix1, pix2, cfg.luminanceWeight); }; + + const bool doLineBlend = [&]() -> bool + { + if (getBottomR(blend) >= BLEND_DOMINANT) + return true; + + //make sure there is no second blending in an adjacent rotation for this pixel: handles insular pixels, mario eyes + if (getTopR(blend) != BLEND_NONE && !eq(e, g)) //but support double-blending for 90° corners + return false; + if (getBottomL(blend) != BLEND_NONE && !eq(e, c)) + return false; + + //no full blending for L-shapes; blend corner only (handles "mario mushroom eyes") + if (!eq(e, i) && eq(g, h) && eq(h, i) && eq(i, f) && eq(f, c)) + return false; + + return true; + }(); + + const uint32_t px = dist(e, f) <= dist(e, h) ? f : h; //choose most similar color + + OutputMatrix out(target, trgWidth); + + if (doLineBlend) + { + const double fg = dist(f, g); //test sample: 70% of values max(fg, hc) / min(fg, hc) are between 1.1 and 3.7 with median being 1.9 + const double hc = dist(h, c); // + + const bool haveShallowLine = cfg.steepDirectionThreshold * fg <= hc && e != g && d != g; + const bool haveSteepLine = cfg.steepDirectionThreshold * hc <= fg && e != c && b != c; + + if (haveShallowLine) + { + if (haveSteepLine) + Scaler::blendLineSteepAndShallow(px, out); + else + Scaler::blendLineShallow(px, out); + } + else + { + if (haveSteepLine) + Scaler::blendLineSteep(px, out); + else + Scaler::blendLineDiagonal(px, out); + } + } + else + Scaler::blendCorner(px, out); + } + + //#undef a +#undef b +#undef c +#undef d +#undef e +#undef f +#undef g +#undef h +#undef i +} + + +class OobReaderTransparent +{ +public: + OobReaderTransparent(const uint32_t* src, int srcWidth, int srcHeight, int y) : + s_m1(0 <= y - 1 && y - 1 < srcHeight ? src + srcWidth * (y - 1) : nullptr), + s_0 (0 <= y && y < srcHeight ? src + srcWidth * y : nullptr), + s_p1(0 <= y + 1 && y + 1 < srcHeight ? src + srcWidth * (y + 1) : nullptr), + s_p2(0 <= y + 2 && y + 2 < srcHeight ? src + srcWidth * (y + 2) : nullptr), + srcWidth_(srcWidth) {} + + void readDhlp(Kernel_4x4& ker, int x) const //(x, y) is at kernel position F + { + [[likely]] if (const int x_p2 = x + 2; 0 <= x_p2 && x_p2 < srcWidth_) + { + ker.d = s_m1 ? s_m1[x_p2] : 0; + ker.h = s_0 ? s_0 [x_p2] : 0; + ker.l = s_p1 ? s_p1[x_p2] : 0; + ker.p = s_p2 ? s_p2[x_p2] : 0; + } + else + { + ker.d = 0; + ker.h = 0; + ker.l = 0; + ker.p = 0; + } + } + +private: + const uint32_t* const s_m1; + const uint32_t* const s_0; + const uint32_t* const s_p1; + const uint32_t* const s_p2; + const int srcWidth_; +}; + + +class OobReaderDuplicate +{ +public: + OobReaderDuplicate(const uint32_t* src, int srcWidth, int srcHeight, int y) : + s_m1(src + srcWidth * std::clamp(y - 1, 0, srcHeight - 1)), + s_0 (src + srcWidth * std::clamp(y, 0, srcHeight - 1)), + s_p1(src + srcWidth * std::clamp(y + 1, 0, srcHeight - 1)), + s_p2(src + srcWidth * std::clamp(y + 2, 0, srcHeight - 1)), + srcWidth_(srcWidth) {} + + void readDhlp(Kernel_4x4& ker, int x) const //(x, y) is at kernel position F + { + const int x_p2 = std::clamp(x + 2, 0, srcWidth_ - 1); + ker.d = s_m1[x_p2]; + ker.h = s_0 [x_p2]; + ker.l = s_p1[x_p2]; + ker.p = s_p2[x_p2]; + } + +private: + const uint32_t* const s_m1; + const uint32_t* const s_0; + const uint32_t* const s_p1; + const uint32_t* const s_p2; + const int srcWidth_; +}; + + +template //scaler policy: see "Scaler2x" reference implementation +void scaleImage(const uint32_t* src, uint32_t* trg, int srcWidth, int srcHeight, const xbrz::ScalerCfg& cfg, int yFirst, int yLast) +{ + yFirst = std::max(yFirst, 0); + yLast = std::min(yLast, srcHeight); + if (yFirst >= yLast || srcWidth <= 0) + return; + + const int trgWidth = srcWidth * Scaler::scale; + + //(ab)use space of "sizeof(uint32_t) * srcWidth * Scaler::scale" at the end of the image as temporary + //buffer for "on the fly preprocessing" without risk of accidental overwriting before accessing + unsigned char* const preProcBuf = reinterpret_cast(trg + yLast * Scaler::scale * trgWidth) - srcWidth; + + //initialize preprocessing buffer for first row of current stripe: detect upper left and right corner blending + //this cannot be optimized for adjacent processing stripes; we must not allow for a memory race condition! + { + const OobReader oobReader(src, srcWidth, srcHeight, yFirst - 1); + + //initialize at position x = -1 + Kernel_4x4 ker4 = {}; + oobReader.readDhlp(ker4, -4); //hack: read a, e, i, m at x = -1 + ker4.a = ker4.d; + ker4.e = ker4.h; + ker4.i = ker4.l; + ker4.m = ker4.p; + + oobReader.readDhlp(ker4, -3); + ker4.b = ker4.d; + ker4.f = ker4.h; + ker4.j = ker4.l; + ker4.n = ker4.p; + + oobReader.readDhlp(ker4, -2); + ker4.c = ker4.d; + ker4.g = ker4.h; + ker4.k = ker4.l; + ker4.o = ker4.p; + + oobReader.readDhlp(ker4, -1); + + { + const BlendResult res = preProcessCorners(ker4, cfg); + clearAddTopL(preProcBuf[0], res.blend_k); //set 1st known corner for (0, yFirst) + } + + for (int x = 0; x < srcWidth; ++x) + { + ker4.a = ker4.b; //shift previous kernel to the left + ker4.e = ker4.f; // ----------------- + ker4.i = ker4.j; // | A | B | C | D | + ker4.m = ker4.n; // |---|---|---|---| + /**/ // | E | F | G | H | (x, yFirst - 1) is at position F + ker4.b = ker4.c; // |---|---|---|---| + ker4.f = ker4.g; // | I | J | K | L | + ker4.j = ker4.k; // |---|---|---|---| + ker4.n = ker4.o; // | M | N | O | P | + /**/ // ----------------- + ker4.c = ker4.d; + ker4.g = ker4.h; + ker4.k = ker4.l; + ker4.o = ker4.p; + + oobReader.readDhlp(ker4, x); + + /* preprocessing blend result: + --------- + | F | G | evaluate corner between F, G, J, K + |---+---| current input pixel is at position F + | J | K | + --------- */ + const BlendResult res = preProcessCorners(ker4, cfg); + addTopR(preProcBuf[x], res.blend_j); //set 2nd known corner for (x, yFirst) + + if (x + 1 < srcWidth) + clearAddTopL(preProcBuf[x + 1], res.blend_k); //set 1st known corner for (x + 1, yFirst) + } + } + //------------------------------------------------------------------------------------ + + for (int y = yFirst; y < yLast; ++y) + { + uint32_t* out = trg + Scaler::scale * y * trgWidth; //consider MT "striped" access + + const OobReader oobReader(src, srcWidth, srcHeight, y); + + //initialize at position x = -1 + Kernel_4x4 ker4 = {}; + oobReader.readDhlp(ker4, -4); //hack: read a, e, i, m at x = -1 + ker4.a = ker4.d; + ker4.e = ker4.h; + ker4.i = ker4.l; + ker4.m = ker4.p; + + oobReader.readDhlp(ker4, -3); + ker4.b = ker4.d; + ker4.f = ker4.h; + ker4.j = ker4.l; + ker4.n = ker4.p; + + oobReader.readDhlp(ker4, -2); + ker4.c = ker4.d; + ker4.g = ker4.h; + ker4.k = ker4.l; + ker4.o = ker4.p; + + oobReader.readDhlp(ker4, -1); + + unsigned char blend_xy1 = 0; //corner blending for current (x, y + 1) position + { + const BlendResult res = preProcessCorners(ker4, cfg); + clearAddTopL(blend_xy1, res.blend_k); //set 1st known corner for (0, y + 1) and buffer for use on next column + + addBottomL(preProcBuf[0], res.blend_g); //set 3rd known corner for (0, y) + } + + for (int x = 0; x < srcWidth; ++x, out += Scaler::scale) + { +#if defined _MSC_VER && !defined NDEBUG + breakIntoDebugger = debugPixelX == x && debugPixelY == y; +#endif + ker4.a = ker4.b; //shift previous kernel to the left + ker4.e = ker4.f; // ----------------- + ker4.i = ker4.j; // | A | B | C | D | + ker4.m = ker4.n; // |---|---|---|---| + /**/ // | E | F | G | H | (x, y) is at position F + ker4.b = ker4.c; // |---|---|---|---| + ker4.f = ker4.g; // | I | J | K | L | + ker4.j = ker4.k; // |---|---|---|---| + ker4.n = ker4.o; // | M | N | O | P | + /**/ // ----------------- + ker4.c = ker4.d; + ker4.g = ker4.h; + ker4.k = ker4.l; + ker4.o = ker4.p; + + oobReader.readDhlp(ker4, x); + + //evaluate the four corners on bottom-right of current pixel + unsigned char blend_xy = preProcBuf[x]; //for current (x, y) position + { + /* preprocessing blend result: + --------- + | F | G | evaluate corner between F, G, J, K + |---+---| current input pixel is at position F + | J | K | + --------- */ + const BlendResult res = preProcessCorners(ker4, cfg); + addBottomR(blend_xy, res.blend_f); //all four corners of (x, y) have been determined at this point due to processing sequence! + + addTopR(blend_xy1, res.blend_j); //set 2nd known corner for (x, y + 1) + preProcBuf[x] = blend_xy1; //store on current buffer position for use on next row + + [[likely]] if (x + 1 < srcWidth) + { + //blend_xy1 -> blend_x1y1 + clearAddTopL(blend_xy1, res.blend_k); //set 1st known corner for (x + 1, y + 1) and buffer for use on next column + + addBottomL(preProcBuf[x + 1], res.blend_g); //set 3rd known corner for (x + 1, y) + } + } + + //fill block of size scale * scale with the given color + fillBlock(out, trgWidth * sizeof(uint32_t), ker4.f, Scaler::scale, Scaler::scale); + //place *after* preprocessing step, to not overwrite the results while processing the last pixel! + + //blend all four corners of current pixel + if (blendingNeeded(blend_xy)) + { + const auto& ker3 = reinterpret_cast(ker4); //"The Things We Do for Perf" + blendPixel(ker3, out, trgWidth, blend_xy, cfg); + blendPixel(ker3, out, trgWidth, blend_xy, cfg); + blendPixel(ker3, out, trgWidth, blend_xy, cfg); + blendPixel(ker3, out, trgWidth, blend_xy, cfg); + } + } + } +} + +//------------------------------------------------------------------------------------ + +template +struct Scaler2x : public ColorGradient +{ + static const int scale = 2; + + template //bring template function into scope for GCC + static void alphaGrad(uint32_t& pixBack, uint32_t pixFront) { ColorGradient::template alphaGrad(pixBack, pixFront); } + + + template + static void blendLineShallow(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<3, 4>(out.template ref(), col); + } + + template + static void blendLineSteep(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref<0, scale - 1>(), col); + alphaGrad<3, 4>(out.template ref<1, scale - 1>(), col); + } + + template + static void blendLineSteepAndShallow(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref<1, 0>(), col); + alphaGrad<1, 4>(out.template ref<0, 1>(), col); + alphaGrad<5, 6>(out.template ref<1, 1>(), col); //[!] fixes 7/8 used in xBR + } + + template + static void blendLineDiagonal(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 2>(out.template ref<1, 1>(), col); + } + + template + static void blendCorner(uint32_t col, OutputMatrix& out) + { + //model a round corner + alphaGrad<21, 100>(out.template ref<1, 1>(), col); //exact: 1 - pi/4 = 0.2146018366 + } +}; + + +template +struct Scaler3x : public ColorGradient +{ + static const int scale = 3; + + template //bring template function into scope for GCC + static void alphaGrad(uint32_t& pixBack, uint32_t pixFront) { ColorGradient::template alphaGrad(pixBack, pixFront); } + + + template + static void blendLineShallow(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<1, 4>(out.template ref(), col); + + alphaGrad<3, 4>(out.template ref(), col); + out.template ref() = col; + } + + template + static void blendLineSteep(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref<0, scale - 1>(), col); + alphaGrad<1, 4>(out.template ref<2, scale - 2>(), col); + + alphaGrad<3, 4>(out.template ref<1, scale - 1>(), col); + out.template ref<2, scale - 1>() = col; + } + + template + static void blendLineSteepAndShallow(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref<2, 0>(), col); + alphaGrad<1, 4>(out.template ref<0, 2>(), col); + alphaGrad<3, 4>(out.template ref<2, 1>(), col); + alphaGrad<3, 4>(out.template ref<1, 2>(), col); + out.template ref<2, 2>() = col; + } + + template + static void blendLineDiagonal(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 8>(out.template ref<1, 2>(), col); //conflict with other rotations for this odd scale + alphaGrad<1, 8>(out.template ref<2, 1>(), col); + alphaGrad<7, 8>(out.template ref<2, 2>(), col); // + } + + template + static void blendCorner(uint32_t col, OutputMatrix& out) + { + //model a round corner + alphaGrad<45, 100>(out.template ref<2, 2>(), col); //exact: 0.4545939598 + //alphaGrad<7, 256>(out.template ref<2, 1>(), col); //0.02826017254 -> negligible + avoid conflicts with other rotations for this odd scale + //alphaGrad<7, 256>(out.template ref<1, 2>(), col); //0.02826017254 + } +}; + + +template +struct Scaler4x : public ColorGradient +{ + static const int scale = 4; + + template //bring template function into scope for GCC + static void alphaGrad(uint32_t& pixBack, uint32_t pixFront) { ColorGradient::template alphaGrad(pixBack, pixFront); } + + + template + static void blendLineShallow(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<1, 4>(out.template ref(), col); + + alphaGrad<3, 4>(out.template ref(), col); + alphaGrad<3, 4>(out.template ref(), col); + + out.template ref() = col; + out.template ref() = col; + } + + template + static void blendLineSteep(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref<0, scale - 1>(), col); + alphaGrad<1, 4>(out.template ref<2, scale - 2>(), col); + + alphaGrad<3, 4>(out.template ref<1, scale - 1>(), col); + alphaGrad<3, 4>(out.template ref<3, scale - 2>(), col); + + out.template ref<2, scale - 1>() = col; + out.template ref<3, scale - 1>() = col; + } + + template + static void blendLineSteepAndShallow(uint32_t col, OutputMatrix& out) + { + alphaGrad<3, 4>(out.template ref<3, 1>(), col); + alphaGrad<3, 4>(out.template ref<1, 3>(), col); + alphaGrad<1, 4>(out.template ref<3, 0>(), col); + alphaGrad<1, 4>(out.template ref<0, 3>(), col); + + alphaGrad<1, 3>(out.template ref<2, 2>(), col); //[!] fixes 1/4 used in xBR + + out.template ref<3, 3>() = col; + out.template ref<3, 2>() = col; + out.template ref<2, 3>() = col; + } + + template + static void blendLineDiagonal(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 2>(out.template ref(), col); + alphaGrad<1, 2>(out.template ref(), col); + out.template ref() = col; + } + + template + static void blendCorner(uint32_t col, OutputMatrix& out) + { + //model a round corner + alphaGrad<68, 100>(out.template ref<3, 3>(), col); //exact: 0.6848532563 + alphaGrad< 9, 100>(out.template ref<3, 2>(), col); //0.08677704501 + alphaGrad< 9, 100>(out.template ref<2, 3>(), col); //0.08677704501 + } +}; + + +template +struct Scaler5x : public ColorGradient +{ + static const int scale = 5; + + template //bring template function into scope for GCC + static void alphaGrad(uint32_t& pixBack, uint32_t pixFront) { ColorGradient::template alphaGrad(pixBack, pixFront); } + + + template + static void blendLineShallow(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<1, 4>(out.template ref(), col); + + alphaGrad<3, 4>(out.template ref(), col); + alphaGrad<3, 4>(out.template ref(), col); + + out.template ref() = col; + out.template ref() = col; + out.template ref() = col; + out.template ref() = col; + } + + template + static void blendLineSteep(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref<0, scale - 1>(), col); + alphaGrad<1, 4>(out.template ref<2, scale - 2>(), col); + alphaGrad<1, 4>(out.template ref<4, scale - 3>(), col); + + alphaGrad<3, 4>(out.template ref<1, scale - 1>(), col); + alphaGrad<3, 4>(out.template ref<3, scale - 2>(), col); + + out.template ref<2, scale - 1>() = col; + out.template ref<3, scale - 1>() = col; + out.template ref<4, scale - 1>() = col; + out.template ref<4, scale - 2>() = col; + } + + template + static void blendLineSteepAndShallow(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref<0, scale - 1>(), col); + alphaGrad<1, 4>(out.template ref<2, scale - 2>(), col); + alphaGrad<3, 4>(out.template ref<1, scale - 1>(), col); + + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<3, 4>(out.template ref(), col); + + alphaGrad<2, 3>(out.template ref<3, 3>(), col); + + out.template ref<2, scale - 1>() = col; + out.template ref<3, scale - 1>() = col; + out.template ref<4, scale - 1>() = col; + + out.template ref() = col; + out.template ref() = col; + } + + template + static void blendLineDiagonal(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 8>(out.template ref(), col); //conflict with other rotations for this odd scale + alphaGrad<1, 8>(out.template ref(), col); + alphaGrad<1, 8>(out.template ref(), col); // + + alphaGrad<7, 8>(out.template ref<4, 3>(), col); + alphaGrad<7, 8>(out.template ref<3, 4>(), col); + + out.template ref<4, 4>() = col; + } + + template + static void blendCorner(uint32_t col, OutputMatrix& out) + { + //model a round corner + alphaGrad<86, 100>(out.template ref<4, 4>(), col); //exact: 0.8631434088 + alphaGrad<23, 100>(out.template ref<4, 3>(), col); //0.2306749731 + alphaGrad<23, 100>(out.template ref<3, 4>(), col); //0.2306749731 + //alphaGrad<1, 64>(out.template ref<4, 2>(), col); //0.01676812367 -> negligible + avoid conflicts with other rotations for this odd scale + //alphaGrad<1, 64>(out.template ref<2, 4>(), col); //0.01676812367 + } +}; + + +template +struct Scaler6x : public ColorGradient +{ + static const int scale = 6; + + template //bring template function into scope for GCC + static void alphaGrad(uint32_t& pixBack, uint32_t pixFront) { ColorGradient::template alphaGrad(pixBack, pixFront); } + + + template + static void blendLineShallow(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<1, 4>(out.template ref(), col); + + alphaGrad<3, 4>(out.template ref(), col); + alphaGrad<3, 4>(out.template ref(), col); + alphaGrad<3, 4>(out.template ref(), col); + + out.template ref() = col; + out.template ref() = col; + out.template ref() = col; + out.template ref() = col; + + out.template ref() = col; + out.template ref() = col; + } + + template + static void blendLineSteep(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref<0, scale - 1>(), col); + alphaGrad<1, 4>(out.template ref<2, scale - 2>(), col); + alphaGrad<1, 4>(out.template ref<4, scale - 3>(), col); + + alphaGrad<3, 4>(out.template ref<1, scale - 1>(), col); + alphaGrad<3, 4>(out.template ref<3, scale - 2>(), col); + alphaGrad<3, 4>(out.template ref<5, scale - 3>(), col); + + out.template ref<2, scale - 1>() = col; + out.template ref<3, scale - 1>() = col; + out.template ref<4, scale - 1>() = col; + out.template ref<5, scale - 1>() = col; + + out.template ref<4, scale - 2>() = col; + out.template ref<5, scale - 2>() = col; + } + + template + static void blendLineSteepAndShallow(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 4>(out.template ref<0, scale - 1>(), col); + alphaGrad<1, 4>(out.template ref<2, scale - 2>(), col); + alphaGrad<3, 4>(out.template ref<1, scale - 1>(), col); + alphaGrad<3, 4>(out.template ref<3, scale - 2>(), col); + + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<1, 4>(out.template ref(), col); + alphaGrad<3, 4>(out.template ref(), col); + alphaGrad<3, 4>(out.template ref(), col); + + out.template ref<2, scale - 1>() = col; + out.template ref<3, scale - 1>() = col; + out.template ref<4, scale - 1>() = col; + out.template ref<5, scale - 1>() = col; + + out.template ref<4, scale - 2>() = col; + out.template ref<5, scale - 2>() = col; + + out.template ref() = col; + out.template ref() = col; + } + + template + static void blendLineDiagonal(uint32_t col, OutputMatrix& out) + { + alphaGrad<1, 2>(out.template ref(), col); + alphaGrad<1, 2>(out.template ref(), col); + alphaGrad<1, 2>(out.template ref(), col); + + out.template ref() = col; + out.template ref() = col; + out.template ref() = col; + } + + template + static void blendCorner(uint32_t col, OutputMatrix& out) + { + //model a round corner + alphaGrad<97, 100>(out.template ref<5, 5>(), col); //exact: 0.9711013910 + alphaGrad<42, 100>(out.template ref<4, 5>(), col); //0.4236372243 + alphaGrad<42, 100>(out.template ref<5, 4>(), col); //0.4236372243 + alphaGrad< 6, 100>(out.template ref<5, 3>(), col); //0.05652034508 + alphaGrad< 6, 100>(out.template ref<3, 5>(), col); //0.05652034508 + } +}; + +//------------------------------------------------------------------------------------ + +struct ColorDistanceRGB +{ + static double dist(uint32_t pix1, uint32_t pix2, double luminanceWeight) + { + return distYCbCrBuffered(pix1, pix2); + + //if (pix1 == pix2) //about 4% perf boost + // return 0; + //return distYCbCr(pix1, pix2, luminanceWeight); + } +}; + +struct ColorDistanceARGB +{ + static double dist(uint32_t pix1, uint32_t pix2, double luminanceWeight) + { + const double a1 = getAlpha(pix1) / 255.0 ; + const double a2 = getAlpha(pix2) / 255.0 ; + /* + Requirements for a color distance handling alpha channel: with a1, a2 in [0, 1] + + 1. if a1 = a2, distance should be: a1 * distYCbCr() + 2. if a1 = 0, distance should be: a2 * distYCbCr(black, white) = a2 * 255 + 3. if a1 = 1, ??? maybe: 255 * (1 - a2) + a2 * distYCbCr() + */ + + //return std::min(a1, a2) * distYCbCrBuffered(pix1, pix2) + 255 * abs(a1 - a2); + //=> following code is 15% faster: + const double d = distYCbCrBuffered(pix1, pix2); + if (a1 < a2) + return a1 * d + 255 * (a2 - a1); + else + return a2 * d + 255 * (a1 - a2); + + //alternative? return std::sqrt(a1 * a2 * square(distYCbCrBuffered(pix1, pix2)) + square(255 * (a1 - a2))); + } +}; + + +struct ColorDistanceUnbufferedARGB +{ + static double dist(uint32_t pix1, uint32_t pix2, double luminanceWeight) + { + const double a1 = getAlpha(pix1) / 255.0 ; + const double a2 = getAlpha(pix2) / 255.0 ; + + const double d = distYCbCr(pix1, pix2, luminanceWeight); + if (a1 < a2) + return a1 * d + 255 * (a2 - a1); + else + return a2 * d + 255 * (a1 - a2); + } +}; + + +struct ColorGradientRGB +{ + template + static void alphaGrad(uint32_t& pixBack, uint32_t pixFront) + { + pixBack = gradientRGB(pixFront, pixBack); + } +}; + +struct ColorGradientARGB +{ + template + static void alphaGrad(uint32_t& pixBack, uint32_t pixFront) + { + pixBack = gradientARGB(pixFront, pixBack); + } +}; +} + + +void xbrz::scale(size_t factor, const uint32_t* src, uint32_t* trg, int srcWidth, int srcHeight, ColorFormat colFmt, const xbrz::ScalerCfg& cfg, int yFirst, int yLast) +{ + if (factor == 1) + { + std::copy(src + yFirst * srcWidth, src + yLast * srcWidth, trg); + return; + } + + static_assert(SCALE_FACTOR_MAX == 6); + switch (colFmt) + { + case ColorFormat::RGB: + switch (factor) + { + case 2: + return scaleImage, ColorDistanceRGB, OobReaderDuplicate>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 3: + return scaleImage, ColorDistanceRGB, OobReaderDuplicate>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 4: + return scaleImage, ColorDistanceRGB, OobReaderDuplicate>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 5: + return scaleImage, ColorDistanceRGB, OobReaderDuplicate>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 6: + return scaleImage, ColorDistanceRGB, OobReaderDuplicate>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + } + break; + + case ColorFormat::ARGB: + switch (factor) + { + case 2: + return scaleImage, ColorDistanceARGB, OobReaderTransparent>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 3: + return scaleImage, ColorDistanceARGB, OobReaderTransparent>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 4: + return scaleImage, ColorDistanceARGB, OobReaderTransparent>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 5: + return scaleImage, ColorDistanceARGB, OobReaderTransparent>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 6: + return scaleImage, ColorDistanceARGB, OobReaderTransparent>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + } + break; + + case ColorFormat::ARGB_UNBUFFERED: + switch (factor) + { + case 2: + return scaleImage, ColorDistanceUnbufferedARGB, OobReaderTransparent>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 3: + return scaleImage, ColorDistanceUnbufferedARGB, OobReaderTransparent>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 4: + return scaleImage, ColorDistanceUnbufferedARGB, OobReaderTransparent>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 5: + return scaleImage, ColorDistanceUnbufferedARGB, OobReaderTransparent>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + case 6: + return scaleImage, ColorDistanceUnbufferedARGB, OobReaderTransparent>(src, trg, srcWidth, srcHeight, cfg, yFirst, yLast); + } + break; + } + assert(false); +} + + +bool xbrz::equalColorTest(uint32_t col1, uint32_t col2, ColorFormat colFmt, double luminanceWeight, double equalColorTolerance) +{ + switch (colFmt) + { + case ColorFormat::RGB: + return ColorDistanceRGB::dist(col1, col2, luminanceWeight) < equalColorTolerance; + case ColorFormat::ARGB: + return ColorDistanceARGB::dist(col1, col2, luminanceWeight) < equalColorTolerance; + case ColorFormat::ARGB_UNBUFFERED: + return ColorDistanceUnbufferedARGB::dist(col1, col2, luminanceWeight) < equalColorTolerance; + } + assert(false); + return false; +} + + +void xbrz::bilinearScale(const uint32_t* src, int srcWidth, int srcHeight, + /**/ uint32_t* trg, int trgWidth, int trgHeight) +{ + bilinearScale(src, srcWidth, srcHeight, srcWidth * sizeof(uint32_t), + trg, trgWidth, trgHeight, trgWidth * sizeof(uint32_t), + 0, trgHeight, [](uint32_t pix) { return pix; }); +} + + +void xbrz::nearestNeighborScale(const uint32_t* src, int srcWidth, int srcHeight, + /**/ uint32_t* trg, int trgWidth, int trgHeight) +{ + nearestNeighborScale(src, srcWidth, srcHeight, srcWidth * sizeof(uint32_t), + trg, trgWidth, trgHeight, trgWidth * sizeof(uint32_t), + 0, trgHeight, [](uint32_t pix) { return pix; }); +} + + +#if 0 +//#include +void bilinearScaleCpu(const uint32_t* src, int srcWidth, int srcHeight, + /**/ uint32_t* trg, int trgWidth, int trgHeight) +{ + const int TASK_GRANULARITY = 16; + + concurrency::task_group tg; + + for (int i = 0; i < trgHeight; i += TASK_GRANULARITY) + tg.run([=] + { + const int iLast = std::min(i + TASK_GRANULARITY, trgHeight); + xbrz::bilinearScale(src, srcWidth, srcHeight, srcWidth * sizeof(uint32_t), + trg, trgWidth, trgHeight, trgWidth * sizeof(uint32_t), + i, iLast, [](uint32_t pix) { return pix; }); + }); + tg.wait(); +} + + +//Perf: AMP vs CPU: merely ~10% shorter runtime (scaling 1280x800 -> 1920x1080) +//#include +void bilinearScaleAmp(const uint32_t* src, int srcWidth, int srcHeight, //throw concurrency::runtime_exception + /**/ uint32_t* trg, int trgWidth, int trgHeight) +{ + //C++ AMP reference: https://msdn.microsoft.com/en-us/library/hh289390.aspx + //introduction to C++ AMP: https://msdn.microsoft.com/en-us/magazine/hh882446.aspx + using namespace concurrency; + //TODO: pitch + + if (srcHeight <= 0 || srcWidth <= 0) return; + + const float scaleX = static_cast(trgWidth ) / srcWidth; + const float scaleY = static_cast(trgHeight) / srcHeight; + + array_view srcView(srcHeight, srcWidth, src); + array_view< uint32_t, 2> trgView(trgHeight, trgWidth, trg); + trgView.discard_data(); + + parallel_for_each(trgView.extent, [=](index<2> idx) restrict(amp) //throw ? + { + const int y = idx[0]; + const int x = idx[1]; + //Perf notes: + // -> float-based calculation is (almost) 2x as fas as double! + // -> no noticeable improvement via tiling: https://msdn.microsoft.com/en-us/magazine/hh882447.aspx + // -> no noticeable improvement with restrict(amp,cpu) + // -> iterating over y-axis only is significantly slower! + // -> pre-calculating x,y-dependent variables in a buffer + array_view<> is ~ 20 % slower! + const int y1 = srcHeight * y / trgHeight; + int y2 = y1 + 1; + if (y2 == srcHeight) --y2; + + const float yy1 = y / scaleY - y1; + const float y2y = 1 - yy1; + //------------------------------------- + const int x1 = srcWidth * x / trgWidth; + int x2 = x1 + 1; + if (x2 == srcWidth) --x2; + + const float xx1 = x / scaleX - x1; + const float x2x = 1 - xx1; + //------------------------------------- + const float x2xy2y = x2x * y2y; + const float xx1y2y = xx1 * y2y; + const float x2xyy1 = x2x * yy1; + const float xx1yy1 = xx1 * yy1; + + auto interpolate = [=](int offset) + { + /* + https://en.wikipedia.org/wiki/Bilinear_interpolation + (c11(x2 - x) + c21(x - x1)) * (y2 - y ) + + (c12(x2 - x) + c22(x - x1)) * (y - y1) + */ + const auto c11 = (srcView(y1, x1) >> (8 * offset)) & 0xff; + const auto c21 = (srcView(y1, x2) >> (8 * offset)) & 0xff; + const auto c12 = (srcView(y2, x1) >> (8 * offset)) & 0xff; + const auto c22 = (srcView(y2, x2) >> (8 * offset)) & 0xff; + + return c11 * x2xy2y + c21 * xx1y2y + + c12 * x2xyy1 + c22 * xx1yy1; + }; + + const float bi = interpolate(0); + const float gi = interpolate(1); + const float ri = interpolate(2); + const float ai = interpolate(3); + + const auto b = static_cast(bi + 0.5f); + const auto g = static_cast(gi + 0.5f); + const auto r = static_cast(ri + 0.5f); + const auto a = static_cast(ai + 0.5f); + + trgView(y, x) = (a << 24) | (r << 16) | (g << 8) | b; + }); + trgView.synchronize(); //throw ? +} +#endif \ No newline at end of file diff --git a/client/xBRZ/xbrz.h b/client/xBRZ/xbrz.h new file mode 100644 index 000000000..d8f5d0fcb --- /dev/null +++ b/client/xBRZ/xbrz.h @@ -0,0 +1,79 @@ +// **************************************************************************** +// * This file is part of the xBRZ project. It is distributed under * +// * GNU General Public License: https://www.gnu.org/licenses/gpl-3.0 * +// * Copyright (C) Zenju (zenju AT gmx DOT de) - All Rights Reserved * +// * * +// * Additionally and as a special exception, the author gives permission * +// * to link the code of this program with the following libraries * +// * (or with modified versions that use the same licenses), and distribute * +// * linked combinations including the two: MAME, FreeFileSync, Snes9x, ePSXe * +// * You must obey the GNU General Public License in all respects for all of * +// * the code used other than MAME, FreeFileSync, Snes9x, ePSXe. * +// * If you modify this file, you may extend this exception to your version * +// * of the file, but you are not obligated to do so. If you do not wish to * +// * do so, delete this exception statement from your version. * +// **************************************************************************** + +#ifndef XBRZ_HEADER_3847894708239054 +#define XBRZ_HEADER_3847894708239054 + +#include //size_t +#include //uint32_t +#include +#include "xbrz_config.h" + + +namespace xbrz +{ +/* +------------------------------------------------------------------------- +| xBRZ: "Scale by rules" - high quality image upscaling filter by Zenju | +------------------------------------------------------------------------- +using a modified approach of xBR: +http://board.byuu.org/viewtopic.php?f=10&t=2248 +- new rule set preserving small image features +- highly optimized for performance +- support alpha channel +- support multithreading +- support 64-bit architectures +- support processing image slices +- support scaling up to 6xBRZ +*/ + +enum class ColorFormat //from high bits -> low bits, 8 bit per channel +{ + RGB, //8 bit for each red, green, blue, upper 8 bits unused + ARGB, //including alpha channel, BGRA byte order on little-endian machines + ARGB_UNBUFFERED, //like ARGB, but without the one-time buffer creation overhead (ca. 100 - 300 ms) at the expense of a slightly slower scaling time +}; + +const int SCALE_FACTOR_MAX = 6; + +/* +-> map source (srcWidth * srcHeight) to target (scale * width x scale * height) image, optionally processing a half-open slice of rows [yFirst, yLast) only +-> if your emulator changes only a few image slices during each cycle (e.g. DOSBox) then there's no need to run xBRZ on the complete image: + Just make sure you enlarge the source image slice by 2 rows on top and 2 on bottom (this is the additional range the xBRZ algorithm is using during analysis) + CAVEAT: If there are multiple changed slices, make sure they do not overlap after adding these additional rows in order to avoid a memory race condition + in the target image data if you are using multiple threads for processing each enlarged slice! + +THREAD-SAFETY: - parts of the same image may be scaled by multiple threads as long as the [yFirst, yLast) ranges do not overlap! + - there is a minor inefficiency for the first row of a slice, so avoid processing single rows only; suggestion: process at least 8-16 rows +*/ +void scale(size_t factor, //valid range: 2 - SCALE_FACTOR_MAX + const uint32_t* src, uint32_t* trg, int srcWidth, int srcHeight, + ColorFormat colFmt, + const ScalerCfg& cfg = ScalerCfg(), + int yFirst = 0, int yLast = std::numeric_limits::max()); //slice of source image + +void bilinearScale(const uint32_t* src, int srcWidth, int srcHeight, + /**/ uint32_t* trg, int trgWidth, int trgHeight); + +void nearestNeighborScale(const uint32_t* src, int srcWidth, int srcHeight, + /**/ uint32_t* trg, int trgWidth, int trgHeight); + + +//parameter tuning +bool equalColorTest(uint32_t col1, uint32_t col2, ColorFormat colFmt, double luminanceWeight, double equalColorTolerance); +} + +#endif diff --git a/client/xBRZ/xbrz_config.h b/client/xBRZ/xbrz_config.h new file mode 100644 index 000000000..a5b1739e2 --- /dev/null +++ b/client/xBRZ/xbrz_config.h @@ -0,0 +1,35 @@ +// **************************************************************************** +// * This file is part of the xBRZ project. It is distributed under * +// * GNU General Public License: https://www.gnu.org/licenses/gpl-3.0 * +// * Copyright (C) Zenju (zenju AT gmx DOT de) - All Rights Reserved * +// * * +// * Additionally and as a special exception, the author gives permission * +// * to link the code of this program with the following libraries * +// * (or with modified versions that use the same licenses), and distribute * +// * linked combinations including the two: MAME, FreeFileSync, Snes9x, ePSXe * +// * You must obey the GNU General Public License in all respects for all of * +// * the code used other than MAME, FreeFileSync, Snes9x, ePSXe. * +// * If you modify this file, you may extend this exception to your version * +// * of the file, but you are not obligated to do so. If you do not wish to * +// * do so, delete this exception statement from your version. * +// **************************************************************************** + +#ifndef XBRZ_CONFIG_HEADER_284578425345 +#define XBRZ_CONFIG_HEADER_284578425345 + +//do NOT include any headers here! used by xBRZ_dll!!! + +namespace xbrz +{ +struct ScalerCfg +{ + double luminanceWeight = 1; + double equalColorTolerance = 30; + double centerDirectionBias = 4; + double dominantDirectionThreshold = 3.6; + double steepDirectionThreshold = 2.2; + double newTestAttribute = 0; //unused; test new parameters +}; +} + +#endif diff --git a/client/xBRZ/xbrz_tools.h b/client/xBRZ/xbrz_tools.h new file mode 100644 index 000000000..5f9a10072 --- /dev/null +++ b/client/xBRZ/xbrz_tools.h @@ -0,0 +1,266 @@ +// **************************************************************************** +// * This file is part of the xBRZ project. It is distributed under * +// * GNU General Public License: https://www.gnu.org/licenses/gpl-3.0 * +// * Copyright (C) Zenju (zenju AT gmx DOT de) - All Rights Reserved * +// * * +// * Additionally and as a special exception, the author gives permission * +// * to link the code of this program with the following libraries * +// * (or with modified versions that use the same licenses), and distribute * +// * linked combinations including the two: MAME, FreeFileSync, Snes9x, ePSXe * +// * You must obey the GNU General Public License in all respects for all of * +// * the code used other than MAME, FreeFileSync, Snes9x, ePSXe. * +// * If you modify this file, you may extend this exception to your version * +// * of the file, but you are not obligated to do so. If you do not wish to * +// * do so, delete this exception statement from your version. * +// **************************************************************************** + +#ifndef XBRZ_TOOLS_H_825480175091875 +#define XBRZ_TOOLS_H_825480175091875 + +#include +#include +#include + + +namespace xbrz +{ +template inline +unsigned char getByte(uint32_t val) { return static_cast((val >> (8 * N)) & 0xff); } + +inline unsigned char getAlpha(uint32_t pix) { return getByte<3>(pix); } +inline unsigned char getRed (uint32_t pix) { return getByte<2>(pix); } +inline unsigned char getGreen(uint32_t pix) { return getByte<1>(pix); } +inline unsigned char getBlue (uint32_t pix) { return getByte<0>(pix); } + +inline uint32_t makePixel(unsigned char a, unsigned char r, unsigned char g, unsigned char b) { return (a << 24) | (r << 16) | (g << 8) | b; } +inline uint32_t makePixel( unsigned char r, unsigned char g, unsigned char b) { return (r << 16) | (g << 8) | b; } + +inline uint32_t rgb555to888(uint16_t pix) { return ((pix & 0x7C00) << 9) | ((pix & 0x03E0) << 6) | ((pix & 0x001F) << 3); } +inline uint32_t rgb565to888(uint16_t pix) { return ((pix & 0xF800) << 8) | ((pix & 0x07E0) << 5) | ((pix & 0x001F) << 3); } + +inline uint16_t rgb888to555(uint32_t pix) { return static_cast(((pix & 0xF80000) >> 9) | ((pix & 0x00F800) >> 6) | ((pix & 0x0000F8) >> 3)); } +inline uint16_t rgb888to565(uint32_t pix) { return static_cast(((pix & 0xF80000) >> 8) | ((pix & 0x00FC00) >> 5) | ((pix & 0x0000F8) >> 3)); } + + +template inline +Pix* byteAdvance(Pix* ptr, int bytes) +{ + using PixNonConst = typename std::remove_cv::type; + using PixByte = typename std::conditional::value, char, const char>::type; + + static_assert(std::is_integral::value, "Pix* is expected to be cast-able to char*"); + + return reinterpret_cast(reinterpret_cast(ptr) + bytes); +} + + +//fill block with the given color +template inline +void fillBlock(Pix* trg, int pitch /*[bytes]*/, Pix col, int blockWidth, int blockHeight) +{ + //for (int y = 0; y < blockHeight; ++y, trg = byteAdvance(trg, pitch)) + // std::fill(trg, trg + blockWidth, col); + + for (int y = 0; y < blockHeight; ++y, trg = byteAdvance(trg, pitch)) + for (int x = 0; x < blockWidth; ++x) + trg[x] = col; +} + + +//nearest-neighbor (going over target image - slow for upscaling, since source is read multiple times missing out on cache! Fast for similar image sizes!) +template +void nearestNeighborScale(const PixSrc* src, int srcWidth, int srcHeight, int srcPitch /*[bytes]*/, + /**/ PixTrg* trg, int trgWidth, int trgHeight, int trgPitch /*[bytes]*/, + int yFirst, int yLast, PixConverter pixCvrt /*convert PixSrc to PixTrg*/) +{ + static_assert(std::is_integral::value, "PixSrc* is expected to be cast-able to char*"); + static_assert(std::is_integral::value, "PixTrg* is expected to be cast-able to char*"); + + static_assert(std::is_same::value, "PixConverter returning wrong pixel format"); + + if (srcPitch < srcWidth * static_cast(sizeof(PixSrc)) || + trgPitch < trgWidth * static_cast(sizeof(PixTrg))) + { + assert(false); + return; + } + + yFirst = std::max(yFirst, 0); + yLast = std::min(yLast, trgHeight); + if (yFirst >= yLast || srcHeight <= 0 || srcWidth <= 0) return; + + for (int y = yFirst; y < yLast; ++y) + { + const int ySrc = srcHeight * y / trgHeight; + const PixSrc* const srcLine = byteAdvance(src, ySrc * srcPitch); + PixTrg* const trgLine = byteAdvance(trg, y * trgPitch); + + for (int x = 0; x < trgWidth; ++x) + { + const int xSrc = srcWidth * x / trgWidth; + trgLine[x] = pixCvrt(srcLine[xSrc]); + } + } +} + + +//nearest-neighbor (going over source image - fast for upscaling, since source is read only once +template +void nearestNeighborScaleOverSource(const PixSrc* src, int srcWidth, int srcHeight, int srcPitch /*[bytes]*/, + /**/ PixTrg* trg, int trgWidth, int trgHeight, int trgPitch /*[bytes]*/, + int yFirst, int yLast, PixConverter pixCvrt /*convert PixSrc to PixTrg*/) +{ + static_assert(std::is_integral::value, "PixSrc* is expected to be cast-able to char*"); + static_assert(std::is_integral::value, "PixTrg* is expected to be cast-able to char*"); + + static_assert(std::is_same::value, "PixConverter returning wrong pixel format"); + + if (srcPitch < srcWidth * static_cast(sizeof(PixSrc)) || + trgPitch < trgWidth * static_cast(sizeof(PixTrg))) + { + assert(false); + return; + } + + yFirst = std::max(yFirst, 0); + yLast = std::min(yLast, srcHeight); + if (yFirst >= yLast || trgWidth <= 0 || trgHeight <= 0) return; + + for (int y = yFirst; y < yLast; ++y) + { + //mathematically: ySrc = floor(srcHeight * yTrg / trgHeight) + // => search for integers in: [ySrc, ySrc + 1) * trgHeight / srcHeight + + //keep within for loop to support MT input slices! + const int yTrgFirst = ( y * trgHeight + srcHeight - 1) / srcHeight; //=ceil(y * trgHeight / srcHeight) + const int yTrgLast = ((y + 1) * trgHeight + srcHeight - 1) / srcHeight; //=ceil(((y + 1) * trgHeight) / srcHeight) + const int blockHeight = yTrgLast - yTrgFirst; + + if (blockHeight > 0) + { + const PixSrc* srcLine = byteAdvance(src, y * srcPitch); + /**/ PixTrg* trgLine = byteAdvance(trg, yTrgFirst * trgPitch); + int xTrgFirst = 0; + + for (int x = 0; x < srcWidth; ++x) + { + const int xTrgLast = ((x + 1) * trgWidth + srcWidth - 1) / srcWidth; + const int blockWidth = xTrgLast - xTrgFirst; + if (blockWidth > 0) + { + xTrgFirst = xTrgLast; + + const auto trgPix = pixCvrt(srcLine[x]); + fillBlock(trgLine, trgPitch, trgPix, blockWidth, blockHeight); + trgLine += blockWidth; + } + } + } + } +} + + +template +void bilinearScale(const uint32_t* src, int srcWidth, int srcHeight, int srcPitch, + /**/ PixTrg* trg, int trgWidth, int trgHeight, int trgPitch, + int yFirst, int yLast, PixConverter pixCvrt /*convert uint32_t to PixTrg*/) +{ + static_assert(std::is_integral::value, "PixTrg* is expected to be cast-able to char*"); + static_assert(std::is_same::value, "PixConverter returning wrong pixel format"); + + if (srcPitch < srcWidth * static_cast(sizeof(uint32_t)) || + trgPitch < trgWidth * static_cast(sizeof(PixTrg))) + { + assert(false); + return; + } + + yFirst = std::max(yFirst, 0); + yLast = std::min(yLast, trgHeight); + if (yFirst >= yLast || srcHeight <= 0 || srcWidth <= 0) return; + + const double scaleX = static_cast(trgWidth ) / srcWidth; + const double scaleY = static_cast(trgHeight) / srcHeight; + + //perf notes: + // -> double-based calculation is (slightly) faster than float + // -> pre-calculation gives significant boost; std::vector<> memory allocation is negligible! + struct CoeffsX + { + int x1 = 0; + int x2 = 0; + double xx1 = 0; + double x2x = 0; + }; + std::vector buf(trgWidth); + for (int x = 0; x < trgWidth; ++x) + { + const int x1 = srcWidth * x / trgWidth; + int x2 = x1 + 1; + if (x2 == srcWidth) --x2; + + const double xx1 = x / scaleX - x1; + const double x2x = 1 - xx1; + + buf[x] = { x1, x2, xx1, x2x }; + } + + for (int y = yFirst; y < yLast; ++y) + { + const int y1 = srcHeight * y / trgHeight; + int y2 = y1 + 1; + if (y2 == srcHeight) --y2; + + const double yy1 = y / scaleY - y1; + const double y2y = 1 - yy1; + + const uint32_t* const srcLine = byteAdvance(src, y1 * srcPitch); + const uint32_t* const srcLineNext = byteAdvance(src, y2 * srcPitch); + PixTrg* const trgLine = byteAdvance(trg, y * trgPitch); + + for (int x = 0; x < trgWidth; ++x) + { + //perf: do NOT "simplify" the variable layout without measurement! + const int x1 = buf[x].x1; + const int x2 = buf[x].x2; + const double xx1 = buf[x].xx1; + const double x2x = buf[x].x2x; + + const double x2xy2y = x2x * y2y; + const double xx1y2y = xx1 * y2y; + const double x2xyy1 = x2x * yy1; + const double xx1yy1 = xx1 * yy1; + + auto interpolate = [=](int offset) + { + /* https://en.wikipedia.org/wiki/Bilinear_interpolation + (c11(x2 - x) + c21(x - x1)) * (y2 - y ) + + (c12(x2 - x) + c22(x - x1)) * (y - y1) */ + const auto c11 = (srcLine [x1] >> (8 * offset)) & 0xff; + const auto c21 = (srcLine [x2] >> (8 * offset)) & 0xff; + const auto c12 = (srcLineNext[x1] >> (8 * offset)) & 0xff; + const auto c22 = (srcLineNext[x2] >> (8 * offset)) & 0xff; + + return c11 * x2xy2y + c21 * xx1y2y + + c12 * x2xyy1 + c22 * xx1yy1; + }; + + const double bi = interpolate(0); + const double gi = interpolate(1); + const double ri = interpolate(2); + const double ai = interpolate(3); + + const auto b = static_cast(bi + 0.5); + const auto g = static_cast(gi + 0.5); + const auto r = static_cast(ri + 0.5); + const auto a = static_cast(ai + 0.5); + + const uint32_t trgPix = (a << 24) | (r << 16) | (g << 8) | b; + + trgLine[x] = pixCvrt(trgPix); + } + } +} +} + +#endif //XBRZ_TOOLS_H_825480175091875