This is the version history, or change log, for all the released versions of NNSuperResolution (latest official release at the top).

v4.0.0 – Auguest 2024

  • Trained all the sequence models from scratch using the NNFlowVector model variant “A” as the built-in optical flow solver (instead of a pretty basic one used previously). This has resulted in much less alignment artifacts between frames, and better and sharper performance overall. (Due to these new and better performing models we have deprecated the old “Plates (Vimeo)” models in this release).
  • Added a new sequence mode knob called “show preroll result” which is used for previewing how well the previous frame is aligned to the current frame. This is purely for debugging purposes.
  • Trained all the still models from scratch using tweaked parameters and an extended dataset. We deem this result to be better than the previous network for most still upscaling scenarios.
  • Added support for Nuke15.1 on all platforms (Linux, Windows and macOS).
  • Fixed a bug where the sequence hybrid mode was always outputting results as if the details knob was set to “1.0”. This has been fixed and the knob value is now treated correctly.

v3.4.4 – April 2024

  • Solved the malloc crashing situation, that sometimes happened when using the plugin in Nuke13.2v2 or later, but that always could be solved by specifying the environment variable NUKE_ALLOCATOR=TBB. This is not needed anymore and the plugin should always be stable.
  • Fixed a minor bug where a black 2 px line was visible at the bottom and on the right side of upscaled images when the processing was performed in a single patch (when max size was set high enough so the whole image could be processed at once).

v3.4.3 – January 2024

  • Fixed a regression bug where the RGBA mode was broken.
  • Added support for Nuke15.0 (which also is natively built against CUDA 11.8 and cuDNN 8.4.1, on Rocky8).

v3.4.2 – November 2023

  • Fixed a bug that created a 1 px wide black line between the image patches when upscaling 2x and only when the padding was set to 1.
  • Increased the default value for the knob “overlap” to 12 (previously set to 4), to improve the image quality in the patch blending regions.

v3.4.1 – November 2023

  • Corrected a bug that made the node crash when a GPU build was used to process on a machine without an installed GPU. Now it automatically falls back to using the CPU instead, as intended.

v3.4.0 – November 2023

  • More efficient releasing of used GPU memory while using the node in a live Nuke GUI session, i.e. you don’t have to restart Nuke anymore to get some of the GPU memory back. Also doing a more thorough GPU memory cleanup when the node is fully deleted (usually when you run “File/Clear” in the Nuke GUI).
  • Added native support for the NVIDIA RTX40xx series of GPUs, i.e. compute capability 8.9 (also supporting 9.0), with a version built against CUDA 11.8 and cuDNN 8.4.1.
  • Added support for Nuke14.1 (which also is natively built against CUDA 11.8 and cuDNN 8.4.1)
  • Added a check for GPU compatibility, i.e. disable the GPU processing options if the current GPU is not supported (instead of crashing).
  • Improved the error logging. It’s now much more clear if the node errors because you are running out of GPU memory.
  • Added text information knobs to the “About” tab that prints out what CUDA/cuDNN versions the current build is against. Also what compute capability that is natively supported and what compute capability your current GPU supports and info about if you will or will not use JIT-compiled kernels. We’ve also added info about compatibility against Nuke’s own AIR tools in relation to the current build configuration.
  • Added a warning to the terminal during node creation if you are relying on JIT-compiled kernels. Just so you get some kind of heads up the first time the JIT-compilation kicks in (which usually takes just above half an hour). There are also warnings about the CUDA_CACHE_MAXSIZE environment variable if it’s not set correctly. These warning can be suppressed with a new environment variable PIXELMANIA_SUPPRESS_KERNEL_WARNINGS if you want.
  • Fixed a bug so the plugin can now upscale single channel layers, for example when you have a depth channel present in a multichannel CG render and you are upscaling all channels.
  • Improved internal processing and memory handling. This improves the performance overall. According to our own tests, the performance increase is between 20-25% in sequence mode when upscaling 4x.

v3.3.0 – March 2023

  • Added support for dedicated render licenses.

v3.2.1 – Sept 2022

  • Fixed a regression where the overscan handling stopped working in sequence mode, but instead resulted in a crash. This was working in v3.0.0, but broke in the optimization release v3.2.0. It’s now fully corrected and working as it should again.

v3.2.0 – May 2022

  • Speed optimizations overall. According to our own internal testing, sequence mode is
  • now about 30% faster to render.
  • Added an option for processing in mixed precision. This is using a bit less VRAM, and is
  • a quite a lot faster on some GPU architectures that are supporting it (RTX).
  • Added an option for choosing what CUDA device ID to process on. This means you can
  • pick what GPU to use if you got a workstation with multiple GPUs installed.
  • Disabled initial heuristics to fix a slow down issue on certain GPU architectures that
  • happened on the first processing frame.
  • Optimized the build of the neural network processing backend library. The plugin binary
  • (shared library) is now a bit smaller and faster to load.
  • Built the neural network processing backend with MKLDNN support, resulting in a vast
  • improvement in rendering speed when using CPU only. According to our own testing it’s
  • sometimes even less than 25% of the render time of v3.0.0 (in sequence mode).
  • Updated the NVIDIA cuDNN library to v8.0.5 for the CUDA10.1 build. This means we are
  • fully matching what Nuke13.x is built against, which means our plugin can co-exists
  • together with CopyCat nodes as well as other AIR nodes by Foundry.
  • Built the neural network processing backend with PTX support, which means that GPUs
  • with compute capability 8.0 and 8.6, i.e. Ampere cards, can now use the CUDA10.1 build if needed (see above). The only downside is that they have to JIT compile the CUDA kernels the first time they run the plugin. Please see the documentation for more information about setting the CUDA_CACHE_MAXSIZE environment variable.
  • Internal checking that the bounding box doesn’t change between frames in sequence mode (it’s not supported having animated bboxes). Now it’s throwing an error instead of crashing.
  • Bug fixes to the “frame range knobs” handling and the “reset frame range” button
  • Better render status logging for what layer is processing to the terminal
  • Better error reporting to the terminal
  • Added support for Nuke13.2

v3.0.0 – November 2021

  • Totally retrained all sequence mode networks from scratch with a higher neural network
  • capacity, to improve the overall upscale results.
  • Trained all the 2x versions of the networks, for both still and sequence mode. Added the
  • scale knob to control the upscale factor.
  • Added the details knob to control the blend between the PSNR and GAN networks.
  • Renamed the sequence type option “Plates (legacy) [RGB]” to “Plates (Vimeo) [RGB]”.
  • Fixed the handling of formats to support upscaling of anamorphic formats.

v2.5.0 – June 2021

  • Totally retrained plate (RGB) solution with sharper and more detailed results. • CG (RGBA) upscale option available (beta).
  • Nuke Indie support (requires the latest version of Nuke12.2 or Nuke13.0).

v2.0.1 – March 2021

  • Bug fix to the progress bar that didn’t show the right thing when processing the preroll in
  • sequence mode using several image patches.
  • Bug fix to a crash that happened when switching between GPU and CPU mode in the
  • same Nuke session.

v2.0.0 – March 2021

  • Release of sequence mode.
  • Support for additional layers, i.e. if you got multiple layers and render them through with
  • a Write node set to “channels: all” the plugin will run the upscale separately on each layer. Please note that there is no alpha channel (4th layer channel) support yet in this version.
  • Renamed the status label from “Inference runs” to “Image patches” to be more clear to end users and also make sense in sequence mode.
  • Fixed a bug relating to processing images using multiple scanline compression.

v1.0.0 – December 2020

  • Initial release (still mode)