<?xml version="1.0"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en">
	<id>https://wiki.selur.net/index.php?action=history&amp;feed=atom&amp;title=Vapoursynth_Temporalfix</id>
	<title>Vapoursynth Temporalfix - Revision history</title>
	<link rel="self" type="application/atom+xml" href="https://wiki.selur.net/index.php?action=history&amp;feed=atom&amp;title=Vapoursynth_Temporalfix"/>
	<link rel="alternate" type="text/html" href="https://wiki.selur.net/index.php?title=Vapoursynth_Temporalfix&amp;action=history"/>
	<updated>2026-11-20T07:52:26Z</updated>
	<subtitle>Revision history for this page on the wiki</subtitle>
	<generator>MediaWiki 1.46.2</generator>
	<entry>
		<id>https://wiki.selur.net/index.php?title=Vapoursynth_Temporalfix&amp;diff=641&amp;oldid=prev</id>
		<title>Selur: New filter page (via create-page on MediaWiki MCP Server)</title>
		<link rel="alternate" type="text/html" href="https://wiki.selur.net/index.php?title=Vapoursynth_Temporalfix&amp;diff=641&amp;oldid=prev"/>
		<updated>2026-11-16T08:53:01Z</updated>

		<summary type="html">&lt;p&gt;New filter page (via create-page on MediaWiki MCP Server)&lt;/p&gt;
&lt;p&gt;&lt;b&gt;New page&lt;/b&gt;&lt;/p&gt;&lt;div&gt;&amp;lt;div style=&amp;quot;float:right; margin:0 0 1em 1em;&amp;quot;&amp;gt;__TOC__&amp;lt;/div&amp;gt;&lt;br /&gt;
&amp;#039;&amp;#039;&amp;#039;Temporalfix&amp;#039;&amp;#039;&amp;#039; adds temporal consistency to clips upscaled with single-image AI models: lines and details that wobble or flicker from frame to frame are averaged over neighbor frames along the motion.&lt;br /&gt;
&lt;br /&gt;
&amp;#039;&amp;#039;Filtering → Vapoursynth → Frame → Stabilize&amp;#039;&amp;#039;&lt;br /&gt;
&lt;br /&gt;
[[File:Vapoursynth_Temporalfix.png|none|Temporalfix controls in Hybrid]]&lt;br /&gt;
&lt;br /&gt;
== What it does ==&lt;br /&gt;
Single-image upscalers treat each frame on its own, so fine lines often &amp;quot;wiggle&amp;quot; over time. Temporalfix stabilizes them by motion-compensated temporal averaging. There are two versions: the classic version runs on the CPU and works best for 2D animation; the Torch version uses an AI model, normally on the GPU. Frame ranges listed under &amp;#039;&amp;#039;Exclude&amp;#039;&amp;#039; are left untouched.&lt;br /&gt;
&lt;br /&gt;
== Requirements ==&lt;br /&gt;
* Classic: CPU; the optional mvgpu motion search needs Vulkan.&lt;br /&gt;
* Torch: Windows and Hybrid&amp;#039;s torch add-on. The &amp;#039;&amp;#039;Torch&amp;#039;&amp;#039; checkbox is only enabled when the torch Temporalfix package is found. The cuda and tensorrt backends need an NVIDIA GPU (tensorrt an RTX card and vs-mlrt&amp;#039;s TensorRT with Torch-TensorRT); directml needs vs-mlrt&amp;#039;s ONNX Runtime and a DirectX 12 GPU; cpu runs without a GPU (slow).&lt;br /&gt;
&lt;br /&gt;
== Settings ==&lt;br /&gt;
{| class=&amp;quot;wikitable&amp;quot;&lt;br /&gt;
! Control !! Parameter !! Range / values !! Hybrid default !! Meaning&lt;br /&gt;
|-&lt;br /&gt;
| Torch || – || on/off || off || Switches between the classic and the Torch version.&lt;br /&gt;
|-&lt;br /&gt;
| Exclude || exclude || frame ranges, e.g. &amp;lt;code&amp;gt;[10 20] [600 900]&amp;lt;/code&amp;gt; || empty || Frame ranges that are not processed.&lt;br /&gt;
|-&lt;br /&gt;
| colspan=&amp;quot;5&amp;quot; | &amp;#039;&amp;#039;Classic&amp;#039;&amp;#039;&lt;br /&gt;
|-&lt;br /&gt;
| Strength || strength || 0 – 10000 || 500 || Suppression strength; 400–700 is usually good.&lt;br /&gt;
|-&lt;br /&gt;
| Temporal Radius || tr || 1 – 25 || 6 || Frames that are averaged; higher is more stable, slower and needs more RAM.&lt;br /&gt;
|-&lt;br /&gt;
| Denoise || denoise || on/off || off || Also removes leftover grain and low-frequency flicker.&lt;br /&gt;
|-&lt;br /&gt;
| Debug || debug || on/off || off || Pink overlay on untouched areas; only used in Hybrid&amp;#039;s preview, never in the encode.&lt;br /&gt;
|-&lt;br /&gt;
| Tools || – || MV: mvgpu, mvutensils; Expr: akarin; RG: zsmooth, neo_smo; Median: zsmooth, std; TMedian: zsmooth; BoxBlur: vszip, std; PlaneStats: vszip; EdgeMasks: edgemasks || – || Implementations used by the classic version.&lt;br /&gt;
|-&lt;br /&gt;
| colspan=&amp;quot;5&amp;quot; | &amp;#039;&amp;#039;Torch&amp;#039;&amp;#039;&lt;br /&gt;
|-&lt;br /&gt;
| Strength || strength || 0.0 – 3.0 || 2.0 || Suppression strength.&lt;br /&gt;
|-&lt;br /&gt;
| Backend || backend || cuda, cpu, tensorrt, directml || tensorrt (cuda when TensorRT is missing) || Inference backend; only installed backends are listed.&lt;br /&gt;
|-&lt;br /&gt;
| GPU ID || gpu_id || 0 – 8 || 0 || GPU to use; not for the cpu backend.&lt;br /&gt;
|-&lt;br /&gt;
| Tiles || tiles || 1 – 16 || 1 || More tiles need less VRAM but are slower.&lt;br /&gt;
|-&lt;br /&gt;
| Streams || num_streams || 1 – 16 || 1 || Parallel streams; only for TensorRT.&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== Example ==&lt;br /&gt;
Default settings on a 10-bit YUV 4:2:0 source (640x356):&lt;br /&gt;
&amp;lt;pre&amp;gt;&lt;br /&gt;
# stabilizing using Temporalfix&lt;br /&gt;
clip = vs_temporalfix_classic.classic(clip, tools={&amp;#039;mv&amp;#039;: &amp;#039;mvutensils&amp;#039;, &amp;#039;expr&amp;#039;: &amp;#039;akarin&amp;#039;, &amp;#039;rg&amp;#039;: &amp;#039;zsmooth&amp;#039;, &amp;#039;median&amp;#039;: &amp;#039;zsmooth&amp;#039;, &amp;#039;tmedian&amp;#039;: &amp;#039;zsmooth&amp;#039;, &amp;#039;boxblur&amp;#039;: &amp;#039;vszip&amp;#039;, &amp;#039;planestats&amp;#039;: &amp;#039;vszip&amp;#039;, &amp;#039;edgemasks&amp;#039;: &amp;#039;edgemasks&amp;#039;})&lt;br /&gt;
&amp;lt;/pre&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== How Hybrid calls it ==&lt;br /&gt;
* Classic: works on 8–16-bit integer YUV in limited or full range (the script handles the range itself); processing is done internally in 16 bit, the motion search on an 8-bit prefiltered copy. Options are only written when they differ from the defaults.&lt;br /&gt;
* Torch: the clip is converted to RGB half precision for tensorrt and directml and to float RGB for cuda and cpu. The backend is always written, the GPU ID (except for cpu) when it is not 0; Hybrid passes its own folder for the TensorRT engines.&lt;br /&gt;
* With TensorRT the first run builds an engine, which can take minutes; it is rebuilt when strength or frame size change.&lt;br /&gt;
* Frame count and frame rate do not change.&lt;br /&gt;
&lt;br /&gt;
== Background ==&lt;br /&gt;
=== Classic version ===&lt;br /&gt;
The classic version is based on SMDegrain/G41Fun-style degraining and SpotLess, with masks for areas that do not follow the global motion and color fixes. Its author recommends cropping black borders before using it.&lt;br /&gt;
&lt;br /&gt;
=== Torch version ===&lt;br /&gt;
The model supports more backends (cpu, directml) and a GPU index; Hybrid offers cuda and TensorRT.&lt;br /&gt;
&lt;br /&gt;
== Sources ==&lt;br /&gt;
* [https://github.com/pifroggi/vs_temporalfix vs_temporalfix]&lt;br /&gt;
* [https://github.com/Selur/VapoursynthScriptsInHybrid VapoursynthScriptsInHybrid] (the script collection Hybrid uses)&lt;br /&gt;
* Hybrid&amp;#039;s tooltips&lt;br /&gt;
&lt;br /&gt;
[[Category:Vapoursynth Filters]]&lt;/div&gt;</summary>
		<author><name>Selur</name></author>
	</entry>
</feed>