@@DrVektor that would be ambitious, I take it as a big compliment. For now, the framework is oriented for the creation of cross-platform application. But, I do not exclude it for the future.

You only need to add a small "boiler plate" code for reading and writing variables in your class. Let's assume you have a uint32_t variable called m_framerate, you only need to do: reading: __readSingleVariable(uint32_t, "framerate", m_framerate); writing: __writeSingleVariable(uint32_t, "framerate", m_framerate, TRUE); __readSingleVariable and __writeSingleVariable are predefined macros containing a small portion of code with the serialization abstraction layer. The last boolean is set to "TRUE" is used to flag the variable as "attribute" if the serialization is done with a markup language (like xml) and ignored if the format is binary. Unlike protobuf, you don't have to write external .proto files and additional code to handle the serialization in your project. This is what I do to serialize a 3D vector class: template <class T> bool_t tmf::Vector3<T>::readSerialized_This(IDeserializer *p_deserializer) { __readSingleVariable(T, "x", m_x); __readSingleVariable(T, "y", m_y); __readSingleVariable(T, "z", m_z); return TRUE; } template <class T> bool_t tmf::Vector3<T>::writeSerialized_This(ISerializer *p_serializer, bool_t p_markupAttributeTurn, bool_t p_headerTurn, bool_t &p_hasObject) { __writeSingleVariable(T, "x", m_x, TRUE); __writeSingleVariable(T, "y", m_y, TRUE); __writeSingleVariable(T, "z", m_z, TRUE); return TRUE; } You can read / write entire data structures, but also unknown objects derived from the class Object, which is an entire system similar to COM. You can serialize also strings, constants, arrays, lists, hash maps and so on. My serialization system supports also automatic compression in zip and lz4, which can be extended to other formats. There is also an abstraction layer to export / import from native formats.

I want to give you a gift. With my framework you can do beautiful stuff like this: tmf::IImageTexture2D *imageTexture = createImageTexture2DFromFiles("TestImage", "gfx\\test_image.png", "gfx\\test_image_alpha.png"); if (imageTexture == NULL) { tmf_error("Unable to load image:%s", "gfx\\test_image.png"); return 1; } tmf::IImage *internalImage = (tmf::IImage *)imageTexture->getImageData(); tmf::IJpegFormatInfo *colorFormatInfo = tmf_createObject<tmf::IJpegFormatInfo>(); colorFormatInfo->setFormatName("jpg"); colorFormatInfo->setQuality(80); internalImage->setColorDataFormatInfo(colorFormatInfo); colorFormatInfo->release(); tmf::ILz4CompressionInfo *alphaFormatInfo = tmf_createObject<tmf::ILz4CompressionInfo>(); alphaFormatInfo->setFormatName("lz4"); internalImage->setAlphaDataFormatInfo(alphaFormatInfo); alphaFormatInfo->release(); tmf::IZipCompressionInfo *compressionInfo = tmf_createObject<tmf::IZipCompressionInfo>(); compressionInfo->setFormatName("zip"); internalImage->setCompressionInfo(compressionInfo); compressionInfo->release(); internalImage->setDeferredFile("test_image.tmg"); tmf_saveToFile(internalImage, "test_image_meta.xml", "xml", TRUE, ""); The application is doing this: - Create a texture 2D with color and alpha channels, loaded from "test_image.png" and "test_image_alpha.png" - Get the internal image used for the texture - Set a jpeg format for the color channel - Set a lz4 compression format for the alpha channel - Compress further the entire object with zip format - Save the image content into a separate file named "test_image.tmg" in binary format - Save the image into a "test_image_meta.xml" file with xml format There is also an interface for saving the native file without changing the data during serialization / deserialization process. For example, you can keep a music in protracker format along the author's comments and converting it into audio track for playing it: the original data is kept untouched inside the file along other variables and stuff.

In the video, I tested on GeForce RTX 3070 8 GB. I tested also on some EC2 instances, like g4dn.xlarge, g5.xlarge and g6.xlarge, and it works fine.

​@@texturemindPretty cool stuff! What would be your intended use for this framework? Looks like it is good for doing pixel intense works like image editing/streaming, but I still don't quite understand the whole thing 😅 maybe it's still a bit experimental?

@@ChengduLittleA basically, the framework can be used to create anything, from pixel intensive applications, to editors or video games. In the video, you see a test program created with the framework for testing remoting protocols like RDP or VNC. The program creates frames with a number associated to them. Frames with the same number will have always the same pixels, even if the test is executed in different moment on different machines. The test is executed in the host machine and a client is connected into it. The entire desktop is observed in the client window through a remoting protocol that makes use of image compression for the transmission. The frames on the client may be captured by a script in python and stored into the client machine. When the test is finished, the client is disconnected from the host and the test is executed again on the client machine by another script in python which compares the frames with the same numbers captured in the previous step. The image are compared with an application like imagemagik to calculate the PSNR. The PSNR is used to determine the overall quality of the network transmission. The framework makes it easy to create this kind of tests because the GUI is deterministic, and to evaluate the remoting protocol you need GUI like images to simulate a customer experience. I hope this explanation clarifies the video. I'm using the framework also to create a remoting protocol by myself, but it's very early stage and full of technical problems to solve (I know it because I worked in a remoting protocol for about 13 years). I hope soon to release the first demo on KZbin. Basically, a client which connects to a host machine and controls the desktop with the mouse. The compression for now will be performed with ffmpeg libraries, in the future I will optimize it for the GPU with NvEnc and AMF. This is the first step I want to achieve. I know it could sound confusing because I'm working on about 10 projects at the same time :-)

The first you said, tile position array, the same I use to draw the tiles on screen.

Yes. This project is implemented with my own TextureMind Framework. In the last 4 months I continued the implementation of the framework, so this project can take benefits from the progresses implemented so far. I improved the resources system and most importantly I implemented all the audio part with OpenAL Soft, which was missing. As next step, I want to introduce physics engine, collision detection and animated textures with program shaders. I don't want to change too much in the original design, only small improvements, like Doom for RTX (in the future, I want to implement RTX rendering as well). Then it will be a porting for Linux and other platforms. Visit www.texturemind.com/post1543/ for more information and to test the old demo (you need a graphics card with the latest Vulkan libraries).

@@texturemind thank you very much for the update!

Hi. Thanks for your interest. The custom GUI will be used to create applications and video games. The current implementation is for Windows but I have in my plans to port the entire framework to Linux and MacOS, so the same for the GUI system. The framework has cross platform design, so it won't be difficult at all, it's something that I have already done in the past, since I have experience as Linux and MacOS. programmer as well. It's not excluded that it will be ported for other operating systems.

Brethless conversion is part of the framework so it's not an abandoned project. It is a visual studio project inside the framework solution (along other applications and tests), so I'm currently using it to debug and evolve the framework itself (i did it on purpose to avoid abandoned stuff). Currently I'm working on the GUI side of the framework, so I had no time for the other 3D projects. Breathless will be continued when I'll return back to the graphics engine. I'm planning to implement PBR (Physical Based Rendering) in my 3D engine, so all the projects will benefit from it , including Breathless. I will use GLSL program shader for some animated textures. I'm planning also to implement realtime raytracing with my GeForce RTX 3700, so it will become something like RTX Doom! i'm really greedy for results this time so it won't pass years to see something on screen. ps: I said visual studio but the framework has a strong cross-platform architecture, so in the future it will be ported for other OS / Platforms, like Linux, MacOS, Emscripten, Android, iOS, Raspberry Pi OS and probably AmigaOS.

In realtà la colpa è di tutti gli impegni che ho avuto in questi anni, ma il progetto non è stato abbandonato. Avevo pensato di riprenderlo tra qualche mese, magari cambiando la grafica che non può essere distribuita per motivi di copyright. L'ideale sarebbe quello di rilasciare un game creator.

@@texturemind Si, la Nintendo è famosa per questo genere di attività "persecutoria" nei confronti di chi sviluppa/converte giochi per macchine ormai morte commercialmente, ma c'è sempre la possibilità di fare qualcosa...l'idea di cambiare grafica non è male, ma sicuramente questo comporterà una notevole perdita di tempo. Detto questo, qualsiasi posizione vorrai prendere in merito a questo progetto sarà ben accolta, sono sempre a disposizione per quello che posso fare!😊

in any CPC, Edge Grinder uses R3 for example

It's just a technical show case, user experience won't be changed I guess. I wish to use the UI for my future projects. I will change the skin, that currently is based on windows xp (check other videos).

Hi, thanks. Yes, I didn't continue the project in the last year because I had health problems (heart issues, fortunately nothing serious), but I will continue it in a near future. I'm refactoring the hardware abstraction layer of the 3D engine to support plugin modules for the supported graphics libraries, I want to extend the future support of the engine also for Direct3D and not only Vulkan. I want also to add effects like dynamic lights, shadows and program shaders, maybe with support for raytracing. Then I think that I could release more videos and a playable demo.

@@texturemind If you need any help from Texture Artist like me im willing to help!

C++ and Vulkan libraries, with VS 2017 Community edition. It's part of an entire framework.

Your comment shows only your low grade of education. This video shows only a feature of an entire framework with a 2D / 3D engine that can handle programmable materials with visual expression nodes and scalable hardware abstraction layer with Vulkan and Cairo library implementations (virtually it can be ported for any driver or operating system). The GUI has been programmed from scratch and it's built on top of the 2D engine with support for modern program shaders. But naturally, you can see only 3 movable quads. Probably if you look at Guernica of Picasso you see only deformed shapes instead of art concept. Compliments.

Wayland is already a replacement for X11 and you can implement your own compositor. There is already a Wayland compositor in Vulkan that is Swvkc, anyway it's not the purpose of this framework. I started this framework years ago as a collection of code written by me in C++ for the development of applications. For now, it runs on Windows, but I used cross-platform principles, so it won't be difficult to port it for Linux or other platforms. Thanks for your interest.

@@texturemind Actually, I thank you for informing and giving life to such a beautiful project. i hope someday it will can hopefully it will work as responsive as windows wayland soon and without input lag.

Grazie! Sarebbe difficile fare un tutorial step by step perché non è uno tra i tanti demo dove si carica il modello 3D e lo si disegna con le Vulkan, il lavoro fa parte di un intero engine 3D con una precisa architettura interna, diversi strati di astrazione e una pipeline di rendering. Forse un giorno scriverò anche il layer di rendering per DirectX e OpenGL, così come il porting verso altre piattaforme. L'intero lavoro è una rivisitazione in chiave moderna di un altro lavoro che mi trascino dietro da circa 20 anni. Dovrei scrivere una serie di documenti tecnici sulle componenti interne, sul formato delle mesh e in generale sul framework che ho scritto e sul formato di serializzazione. Per l'implementazione, un giorno forse rilascerò il codice open source. Potrei anche scrivere un paio di tutorial sulle Vulkan.

@@texturemind So, in short it's useless.

@@2002budokan So, in short, you didn't understand a damn thing. I've said in italian that the engine cannot be explained as a "step by step" tutorial, because it's part of a complex work with an abstraction layer to support multiple graphics libraries, like DirectX and OpenGL. How in your puny understanding an abstraction layer can be useless?

Thanks. The most difficult challenge was to load the maps from the original Amiga format (GLD) and convert them from raycasting into polygons. I want to create something similar to Doom Remake 4 and I feel like I have the instruments for doing it. The 3D engine is part of a bigger work programmed by me that is called TextureMind Framework. Even if you cannot see it in this video, in that engine I have reprogrammed the same material system of Unreal Engine 4, with some improvements. For instance, I can take one shader from ShaderToys and use it as a texture for one of the materials. I have also a system to handle 3D characters with skinned meshes, bones and animations (check my other videos). The animations can be imported from many formats with the library AssImp, so I want to replace the old flat 2D sprites with modern 3D characters. I'm looking also to RTX for real-time raytracing effects.

@@texturemind I have little underestanding of what you are saying but you seem very talented from the end result.. congratulations again! The amiga gaming industry was once built with the help of talented programmers like yourself..

Thanks for the answer, I'm glad to read the comment from the developer of Amiga Breathless. I already have a solid job (SDE2 for Amazon), so I can dedicate myself to these activities in my spare time without the need of becoming a celebrity on youtube or to make money, but only for the pleasure of doing it. The creation of a remake / clone of Breathless was an old dream of when I was 14, along other titles like Gloom, Alien Breed 3D, Super Stardust, Mario 64 and so on. This work is part of another bigger project that is called TextureMind Framework, developed for these kinds of tasks, games, demos, presentations and so on. Probably I'll create also clones of past Amiga demos, like Lech by the Freezers or Fruit Kitchen by the Silents, ported into modern GPU rendering. I will upload videos along my progresses, it can be Breathless or other projects around this framework.

@@texturemind It's a nice project indeed. Keep me posted on your progresses !

Hi, thanks for the comment. This is a very old project and I don't have the entire source code anymore. Today I'm very busy with my job (SDE II for AWS, so you can imagine) but in my spare time I'm developing a new framework in C++ with more advanced features, I'm creating new 3D engine that will have advanced features like Vulkan and ray tracing support. As soon as possible I wish to create a series of videos on youtube to show this work. Even if this is an amateur project, I' don't know if it will be open source one day because of my job position, but I can release some closed-source demos and videos so you can continue to follow this channel to have more information about it. Regards.

@@texturemind I appreciate that! Looking forward to the series.

No, it's not. The tiles are taken from the original SMB and converted to 4 dithered colors (mode 1). The resolution is 256x192, which covers 16x12 tiles on screen (enough to draw World 1-1). The horizontal scrolling (and so the update of the screen) is realized drawing only two columns of bytes on the right of the screen and shifting the screen by 1 character (8 pixels = 2 bytes) changing the offset of CRTC R12 and R13 (in a recent implementation, I made a smoother scrolling changing also the CRTC R3). The physics of Mario was more complicated. I found online indications for the gravity and the acceleration, that may vary on the base of the action.