<?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_DDD</id>
	<title>Vapoursynth DDD - 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_DDD"/>
	<link rel="alternate" type="text/html" href="https://wiki.selur.net/index.php?title=Vapoursynth_DDD&amp;action=history"/>
	<updated>2026-11-19T06:30:16Z</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_DDD&amp;diff=478&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_DDD&amp;diff=478&amp;oldid=prev"/>
		<updated>2026-11-16T08:48:07Z</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;DDD&amp;#039;&amp;#039;&amp;#039; (Deep Deinterlacing) is a neural-network deinterlacer: temporally aware models rebuild every field as a full frame.&lt;br /&gt;
&lt;br /&gt;
&amp;#039;&amp;#039;Filtering → (De-)Interlace/Telecine → Deinterlacer&amp;#039;&amp;#039;&lt;br /&gt;
&lt;br /&gt;
[[File:Vapoursynth_DDD.png|none|DDD controls in Hybrid]]&lt;br /&gt;
&lt;br /&gt;
== What it does ==&lt;br /&gt;
The networks look at three neighboring fields at a time and reconstruct the missing lines of each field. The native output has one frame per field (double rate); Hybrid can drop every second frame to keep the source rate.&lt;br /&gt;
&lt;br /&gt;
== Requirements ==&lt;br /&gt;
* Windows with the torch add-on.&lt;br /&gt;
* An NVIDIA graphics card with CUDA.&lt;br /&gt;
* Without the add-on&amp;#039;s DDD models the entry is not offered.&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;
| Method || – (function) || DDD, DeF || DDD || DDD is the fastest network with the lowest quality; DeF is slower and better.&lt;br /&gt;
|-&lt;br /&gt;
| Bob || – || on/off || on || On: keep all frames (double rate). Off: keep every second frame (source rate).&lt;br /&gt;
|-&lt;br /&gt;
| Field || – || even, odd || even || Which output frames are kept when Bob is off; only enabled then.&lt;br /&gt;
|-&lt;br /&gt;
| TTA || tta || on/off || off || Test-time augmentation: slightly better quality at about four times the cost.&lt;br /&gt;
|-&lt;br /&gt;
| FP16 || fp16 || on/off || on || Half-precision inference, roughly twice as fast.&lt;br /&gt;
|}&lt;br /&gt;
&lt;br /&gt;
== Example ==&lt;br /&gt;
Default settings on an 8-bit YUV 4:2:0 interlaced source (704x576, 25 fps, top field first):&lt;br /&gt;
&amp;lt;pre&amp;gt;&lt;br /&gt;
# adjusting color space from YUV420P8 to RGBS for vsDDD&lt;br /&gt;
clip = core.resize.Bicubic(clip=clip, format=vs.RGBS, matrix_in_s=&amp;quot;470bg&amp;quot;, range_in_s=&amp;quot;limited&amp;quot;, range_s=&amp;quot;full&amp;quot;)&lt;br /&gt;
clip = vs_deepdeinterlace.DDD(clip, tff=True, fp16=True, device=&amp;quot;cuda&amp;quot;)&lt;br /&gt;
# Making sure content is perceived as frame based&lt;br /&gt;
clip = core.std.SetFrameProps(clip=clip, _FieldBased=vs.FIELD_PROGRESSIVE) # scan type: progressive&lt;br /&gt;
&amp;lt;/pre&amp;gt;&lt;br /&gt;
&lt;br /&gt;
== How Hybrid calls it ==&lt;br /&gt;
* The clip is converted to 32-bit float RGB (the only format the models accept), deinterlaced on the CUDA device, and converted back later.&lt;br /&gt;
* Field order: taken from the detected scan type or [[Overwrite input scan type]]; bottom field first runs as BFF, every other interlaced scan type as TFF.&lt;br /&gt;
* Output frame rate: double with Bob; otherwise every second frame is kept and the source rate stays.&lt;br /&gt;
* Afterwards the clip counts as progressive. No decimation.&lt;br /&gt;
* The same models can serve as the external interpolation (EdiExt) of [[Vapoursynth QTGMC|QTGMC]].&lt;br /&gt;
&lt;br /&gt;
== Background ==&lt;br /&gt;
The models come from pifroggi&amp;#039;s vs_deepdeinterlace. The package has a third, best network, DfConvEkSA (five fields, optional motion compensation). Hybrid does not offer it: it needs mmcv 2.0.0 or newer, which caused problems.&lt;br /&gt;
&lt;br /&gt;
== Sources ==&lt;br /&gt;
* [https://github.com/pifroggi/vs_deepdeinterlace vs_deepdeinterlace README]&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>