Can’t Save Final VurtualDub2 File as an MP4 (only AVI)

[Avagadro1]

..

[ProWo]

PCM audio is not supported. Save the video without audio as mp4. You can later convert the audio track to another format (ac3, aac) and mux it to the new video.

[Avagadro1]

..

[ProWo]

You can remux your avi to mp4 with my clever FFmpeg-GUI.

[davexnet]

You can certainly capture in PCM/WAV this is advised.

However, later, when you do something in Virtualdub you have to convert the audio to something else
Audio/full processing mode/compression/FFmpeg AAC (use the configure button to raise the bitrate to 160
and try that)

The alternative is to disable the audio in Vdub and deal with it later
Audio/no audio

[Avagadro1]

..

[ProWo]

Main, Multiplex, set your avi as videosource, set the same avi as audiosource, set a target filename, set mp4 as output container, click Multiplex. Done.

[Avagadro1]

..

[DB83]

You do realise that your workflow removes all the advantages in capturing in a lossless format in the first place.


Your mp4 is now lossy and will have long GOPs, maybe as large as 450 frames, which make further editing, if frame-accurate cuts are important, difficult if not impossible without even more quality loss.


Mp4 really should be the final delivery format after ALL editing and not a bridge between one editor and another.


BTW You confuse vdub/vdub2 with mp4/mp4 v2. Only mp4 v2 allows for various video/audio codecs whereas the original container is restrictive in its options in this regard.

[Avagadro1]

..

[davexnet]

Continue to capture as you're doing.
Explore the other codecs in the Virtualdub2 compression options, for example lossless FFmpeg FFV1
(set as 8bit,4:2:2 chroma) You can encode a small file to see if your editor will accept it.

If not, you could continue to use H264 and turn the CRF down to 1, set YUV 4:2:2 in the configure window.
It's not lossless, but maintains much of the quality

What H264 settings were you using?

[Avagadro1]

..

[davexnet]

Use mediainfo in text view to check the color

Chroma subsampling                       : 4:2:2

Set it like this

[Avagadro1]

..

[davexnet]

Well you should check if you're able to set chroma sampling of 4:2:2 in your H.264 capture
if you want to try that again.

It'll be interlaced capture in a progressive encoded file, but as long as you handle it right
it may work out

CRF is control rate factor, a single pass rate control mode . The lower the number, higher the quality, biggest files sizes

[Avagadro1]

..

[DB83]

I never implied or stated that you should leave the lossless huffyuv capture out of the workflow. I was merely illustrating that the more encodes you do from there the more quality you lose.


In fact you now state that your final aim is dvd so that means yet another even lossier encode from your mp4.


But if dvd is your aim then capture as huffyuv, do your filtering in vdub2 but output not as mp4 but as dvd-compliant mpeg2. As long as you are doing no more editing then you do not even need Premier to create the dvd. Several free programs exist to do just that and also burn. If you want a copy of the filtered video for PC playback then by all means also output another copy as mp4 but, again, aim not to use that for dvd authoring.

[Avagadro1]

..

[DB83]

Unfortunately I have no experience with Premiere Elements. But I would have thought that it accepts mpeg2 which will be much easier to edit than long-GOP mp4

[Avagadro1]

..

[DB83]

Easy to read this


https://en.wikipedia.org/wiki/Group_of_pictures


than for me to try to explain it

[Avagadro1]

..

[Avagadro1]

..

[davexnet]

The best method is to capture lossless (the bits straight out of the ADC) and manipulate the video
(anything that causes quality loss) as little as possible before rendering to your delivery format

In the scenario we discussed earlier your were talking about the original Huffy capture and how to convert it
to a format acceptable to your NLE

I suggested H264 with CRF 1 @ 4:2:2 (CRF 0 is lossless but not available in Virtualdub2). CRF 1 is not lossless
but preserves much of the quality. I also mentioned a possible alternative in Virtualdub2 the FFmpeg FFV1
lossless codec - to try a short clip to see if it was compatible with your NLE.

At the end of the day the choice is yours. We point out the best practices and things to consider,
but you do what you like - don't have to justify it to us

[poisondeathray]

In an earlier discussion, I noted that my intended workflow (VHS to digital) must take into account that my video editing program, Adobe Premier Elements, does not support HuffYUV or other lossless codec; I must use a supported video codec, which in this case will be H.264.

You advised me that, all things considered, I should be ok if in VirtualDub I select H264 as the compressor but in its Configuration dialog turn the CRF down to 1 (and set YUV to 4:2:2).

But did you check to see if Elements accepts the 4:2:2 variety of h264 ? It might not. Pro versions will like Premiere Pro, but 4:2:2 tends not to be a consumer oriented format. Also check - it might accept the lossless version of x264 and treat it like YUV as Premiere does (in vdub2 x264 configuration , select single pass - lossless)

Also, you can set the aspect ratio with SAR width and SAR height. You can use 10:11 for 720x480 4:3 VHS

Also, you can set --tff or --bff in the extra command line box to enter the field order, so you get proper interlaced encoding and signaling

So if capture at bit-rate 8000 seems to produce the optimum video quality, and that corresponds to about “10” in the VirtualDub H264 CRF slider, why capture a slider setting “1”, which will result in a much larger file?

Because the quality is higher, and you were obsessed with lossless format, perhaps for additional filtering and processing. It was already mentioned in your other threads CRF1 is near lossless, higher quality than even Prores XQ4444 or Cineform Filmscan3 . You won't be able to tell the difference if the end result is DVD. QP1 will actually yield slightly higher quality and filesize than CRF1 .

(ii) If one were intending to do lots of post-capture editing, the potential advantage of a far higher capture rate, say a VirtualDub H264 CRF of 5 or lower (bit-rate of 16K to 32K), is that it would result in many more Key frames, and thus editing would be much more accurate.

Not keyframes. Keyframes are where the IDR or "I" frames are placed. Keyframes are related to the GOP length. A max keyframe interval of "1" means every frame is a keyframe. This affects seeking latency, shorter values imply shorter max GOP length, and editing is "snappier" . But everything should be "accurate" in terms of the correct frame - long gop or not - otherwise you have bigger problems. There are other settings to make it decode faster in the the editor too, you can use checkmark tune fastdecode and zero latency. Generally, the less compressed, the larger the filesize at given CRF or QP level , but this helps editor seeking performance. More compressed implies smaller filesizes, but it takes longer to seek and decode, so editor performance is slower