Aspect Ratio setting in VirtualDub Capture mode

[alesco]

Hi,

I would like to capture my VHS tapes via VCR using VirtualDub. I have made some tests and it will work. I have few questions about issues that I would like to make clear/fix. Thanks for your answers!

1) my tapes are NTSC, so I have set output to 720x480 Custom YUV2 format.
problem: the output is 3:2 and the picture is wider than it should be - how can I fix that (please bare in mind I want to capture lossless video and audio without reencoding/converting it besides AVI>MPEG-2 conversion). I know it is a matter of pixel shape but not sure how to fix it:/

2) what is your preferred setting for keeping audio in sync? I can use coax/antenna cable, rca and s-video as well as input.

3) Huffyuv - does it "defaultly" capture the original colour scheme? (I want to be sure it is not converted)

4) just to be sure - S-Video input is useless (does not change the quality) if the source VHS is not SVHS, right? Could it help preventing a/v synchronization problems (see question 2)) or enhance audio quality?


Thanks again.

[jagabo]

1) my tapes are NTSC, so I have set output to 720x480 Custom YUV2 format.
problem: the output is 3:2 and the picture is wider than it should be - how can I fix that (please bare in mind I want to capture lossless video and audio without reencoding/converting it besides AVI>MPEG-2 conversion). I know it is a matter of pixel shape but not sure how to fix it:/

When you open the AVI file in your editor/encoder you specify that its display aspect ratio is 4:3.

2) what is your preferred setting for keeping audio in sync?

Whatever works right with your capture device.

3) Huffyuv - does it "defaultly" capture the original colour scheme? (I want to be sure it is not converted)

It depends on what you are capturing. You should capture as YUY2.

4) just to be sure - S-Video input is useless (does not change the quality) if the source VHS is not SVHS, right?

No, that's not correct. Regular VHS encodes the chroma channels separately. S-video will keep the chroma channels separate and prevent degradation and dot crawl artifacts that come from multiplexing the chroma and luma (composite video). This is assuming that the deck isn't producing s-video with a cheap composite to s-video circuit.

Could it help preventing a/v synchronization problems (see question 2)) or enhance audio quality?

No.

[alesco]

Thank you.

2) okay I was mainly wondering about the options: -dropping frames, - inserting frames, -correct video timing etc.

3) in the configuration window for the codec in VirtualDub, there is no option to pick, just the YUV2 and RGB compression methods. Or is it automatically depending on the "Custom format" setting?

4) so if possible, I should use it, no matter of the type of the VHS?

1) my tapes are NTSC, so I have set output to 720x480 Custom YUV2 format.
problem: the output is 3:2 and the picture is wider than it should be - how can I fix that (please bare in mind I want to capture lossless video and audio without reencoding/converting it besides AVI>MPEG-2 conversion). I know it is a matter of pixel shape but not sure how to fix it:/

When you open the AVI file in your editor/encoder you specify that its display aspect ratio is 4:3.

2) what is your preferred setting for keeping audio in sync?

Whatever works right with your capture device.

3) Huffyuv - does it "defaultly" capture the original colour scheme? (I want to be sure it is not converted)

It depends on what you are capturing. You should capture as YUY2.

4) just to be sure - S-Video input is useless (does not change the quality) if the source VHS is not SVHS, right?

No, that's not correct. Regular VHS encodes the chroma channels separately. S-video will keep the chroma channels separate and prevent degradation and dot crawl artifacts that come from multiplexing the chroma and luma (composite video). This is assuming that the deck isn't producing s-video with a cheap composite to s-video circuit.

Could it help preventing a/v synchronization problems (see question 2)) or enhance audio quality?

No.

[edDV]

4) just to be sure - S-Video input is useless (does not change the quality) if the source VHS is not SVHS, right? Could it help preventing a/v synchronization problems (see question 2)) or enhance audio quality?

Not true. Both VHS and SVHS record luma and chroma separately on tape (aka color under process). S-Video keeps luma and chroma separate from tape to A/D. Composite video mixes luma and chroma (modulated subcarrier) to one shielded wire. At the capture card the luma and chroma must be separated (Y/C separation) before A/D.

However, since VHS luma is limited to 3 MHz (chroma subcarrier at 3.58 MHz), Y/C separation can be accomplished with simple notch filters. SVHS has luma bandwidth out to 4.5 MHz, hence needs a more complex comb filter for Y/C separation. This capture card filter can be avoided if S-Video is used.

[sanlyn]

4) just to be sure - S-Video input is useless (does not change the quality) if the source VHS is not SVHS, right? Could it help preventing a/v synchronization problems (see question 2)) or enhance audio quality?

I think you're confusing S-VHS and s- video. The "S" in S-VHS does not stand for "s-video". The "S" in S-VHS comes from the term "Super-VHS". S-VHS is a tape format. S-video is a type of data transmission, and its associated cables and i/o circuits.

You can play VHS, S-VHS, DV, DVD, or whatever your players use as source material, and they will all play thru composite, s-video, component, or whatever ouput your player supports. The source format and the output transmission are two different things. The overall quality of output types are rated this way:

worst: RF cable
better: composite
even better: s-video
best: component

As explained in the above post by edDV, S-VHS works best with s-video cable, but it will still work with other cable types (but expect inferior output). Almost all video will play with better quality thru s-video or component. Then there's HDMI, which is another matter. . .

[alesco]

Thanks both. Yes I must have confused the terms into one thought. So even if I have a simple VHS format, S-VIDEO is better than composite for transfering. That's what I was asking about.

The bandwith(s) schemes are really helpful.

[manono]

problem: the output is 3:2 and the picture is wider than it should be - how can I fix that

When sending to your MPEG-2 encoder, you set the DAR there (4:3).

[alesco]

Thanks. AVS is the best choice?

problem: the output is 3:2 and the picture is wider than it should be - how can I fix that

When sending to your MPEG-2 encoder, you set the DAR there (4:3).

[sanlyn]

So even if I have a simple VHS format, S-VIDEO is better than composite for transfering.

Yes, much better.

[alesco]

Okay, I have tried both, but S-Video input works only in black & white:/ The composite is working okay though. Any ideas?

Also, I am still a bit suffering from the out of sync audio in virtualdub, so any advice welcome.

[sanlyn]

Something wrong there. We might need more info here. What player and capture device are you using?

[alesco]

Hauppauge HVR 713X BDA (Hauppauge 1100 PCI). VCR is Panasonic NVSJ422. SCART>Reduction>S-Video/Composite>Hauppauge>VirtualDub. I thought it is because it is a PAL VCR but the problem is the same with NTSC and PAL VHSs.

[edDV]

Does S-Video play in color to a monitor?

Hauppauge cards need to be set to include S-Video capture in the WinTV Channel Manager Sources tab.


And then select S-Video in Menu Channel Select.

[edDV]

Or after setting up the source in WinTV-Channel Manager, select it in VDub

[manono]

Thanks. AVS is the best choice?

Frameserving via an AviSynth script? I think so. Others think differently.

[alesco]

In WinTV is doesn't play in colour as well. It is running through S-Video input. I have tried various formats and also 'auto' option but no colours.

The Hauppauge number is 1110 to be correct.

[alesco]

Okay think I will use composite for now. I have two more questions:

1) I have selected yuv2 in custom format setting, but huffyuv encodes the avi to RGB. I wanted to have yuv2 to retain the best possible quality. How do I change the huffyuv format? Or is it automatic?

2) Any tips for A/V sync setting? now I have it like this:
- drops/insert frames are off
- resyn is off
- correct video timing for .... is unchecked
- automatically disable resync .... is checked since I use composite straight to my TV Card.

Seems that it could work, but I think the sync is a bit off (that can be fixed afterwards). Not sure if it is constant, though. I have capped a 10 minute vid.

Thanks

[2Bdecided]

Hauppauge HVR 713X BDA (Hauppauge 1100 PCI). VCR is Panasonic NVSJ422. SCART>Reduction>S-Video/Composite>Hauppauge>VirtualDub. I thought it is because it is a PAL VCR but the problem is the same with NTSC and PAL VHSs.

I've never heard of that VCR, but unless it's an S-VHS machine it probably doesn't have an S-video output.

If it is an S-VHS machine, they'll be a menu somewhere to switch the SCART output between composite and S-video.

Just using a SCART adapter/lead/cable with S-video won't give you an S-video output unless the VCR is sending it - if the VCR just sends composite, such an adapter will give you composite on the S-video luma pin and nothing on the S-video chroma pin. i.e. noisy black and white if you try to capture it as S-video.

Cheers,
David.

[jagabo]

1) I have selected yuv2 in custom format setting, but huffyuv encodes the avi to RGB. I wanted to have yuv2 to retain the best possible quality. How do I change the huffyuv format? Or is it automatic?

HuffYUV compresses whatever you send to it. It has the option to convert RGB to YUY2 before compression. But if you're capturing YUY2 it should be receiving YUY2.

2) Any tips for A/V sync setting? now I have it like this:

Disable audio playback while capturing. Audio -> Enable Audio Playback (off). All the different A/V sync options are there because different devices require different settings. So it's hard to give specific advice about each setting. And of course, follow all the advice about how to avoid dropped frames:

https://forum.videohelp.com/threads/104098-Why-does-your-system-drop-frames

[alesco]

Alright.

Now when I have an AVI, I want to convert it to MPEG-2, in 4:3 and max bitrate and lossless audio. Any ideas what to use? Avisynth? HC? Tmpegenc? Also, I want to cut off some scenes as well, so I would rather have AVI>MPEG-2>Cut>DVD Author

[manono]

It's easier to cut when it's still a lossless AVI because every frame is a keyframe.

Max Bitrate and lossless audio? Lossless audio is PCM WAV (probably just the way you capped it). But using that will limit your max bitrate because the audio bitrate is so large.

Me, of the choices you mentioned, I'd use an AviSynth script opened in HCEnc.

[sanlyn]

Alright.

Now when I have an AVI, I want to convert it to MPEG-2, in 4:3 and max bitrate and lossless audio. Any ideas what to use? Avisynth? HC? Tmpegenc? Also, I want to cut off some scenes as well, so I would rather have AVI>MPEG-2>Cut>DVD Author

Most serious hobbyists would use a separate piece of software for each of the tasks you mention. Browse the Tools section of videohelp for more choices than you could possibly want, some free, some not-free.

You do a lot of cutting in Avisynth and VirtualDub using AVI's. For any processing or editing, keep your audio in lossles (PCM) format and lossless video compression (huffdyuv or Lagarith). You don't want compression of any kind until you get to your final MPEG rendering. For audio, the two standard formats for MPEG/DVD are PCM or Dolby AC3. There's no reason to keep your final audio lossless, but many people do; you'll save about 10% disc space using Dolby.

For encoding, I'd recommend only two apps. The first is the free HC Encoder, but it does have a learning curve (what doesn't?). More user-friendly is the old standby, TMPGenc Plus 2.5 (still a mere $38!). It's SD only, although can output 16:9 SD format with it. Recommended because it's easy to use, doesn't take 6 months of study, is still one of the best encoders around at mid-to-high bitrates, outputs many acreen formats, does a neat job of things like 3:2 pulldown, etc., but mainly because it has excellent color correction controls and a ghost filter you won't believe (very old-fashioned noise filter, though. Avoid using it. Avisynth and VirtualDub are much better for denoising). You'll find comments on TMPGenc Plus here. HC Encoder is mentioned all over this forum. There are other worthy encoders, but the good ones cost a lot more than either of these recommendations.

I would not recommend TMPGenc's newer MPEG encoders, for any reason. They are inferior to the old 2.5, but they do handle HD material. HC Encoder would be a better choice.

You can "edit" (cut) in many authoring programs, including TMPGEnc Authoring works, but most of these authoring programs will cut on key frames only. There are free frame-level editors around, but I've been using TMPGENC MPEG Editor v3 for a while now; it has excellent performance and many extras. It's for MPEG editing/joining/effects only, not for AVI, but I find it indispensable for many tasks. Pretty good Dolby encoder, too.

There are also free MPEG authoring programs. and you might already have seen the not-free Authoring Works 5.

You'll find plenty of comments in this forum concerning all these tools. None are perfect. Look before you commit.

[alesco]

thanks manono, yes I was thinking about this combination.
I have capped it in PCM yes. I want to leave it lossless.

so the process will be:
1) cutting the avi in VirtualDub
2) splitting audio and video (necessary?)
3) creating an AviSynth script
4) running it through HCEnc
5) muxing the mpeg-2 with wav (or is it included in the script?)

It would be great if you provided me with any guide, thanks.

[sanlyn]

I don't see any mention of filtering, denoising, color correction, etc. Almost every piece of VHS could use some of that. When you get to the rendering stage I think you'll see what I mean. But many people don't bother. If the latter is the case, why go to AVI in the first place? You can record directly to MPEG and use an MPEG editor for cutting. Most people would go into the AVI phase because of corrections that even a prisitine VHS usually calls for. But thar's a matter of preference, not one of picky "requirements". Let us know how it's going.

[alesco]

thanks for your input Sanlyn.

I wanted to do a most possible quality dvd transfer of tapes.. so I picked lossless avi and audio, then when I encode it into MPEG-2 with AviSynth, I should get best possible results (hopefully). I would like to leave the picture as much original as possible so the use of filters was not my choice. Not sure if it would improve the quality besides colour/brightness/contrast enhancement, which I can set afterwards in the player or remaster the original dvd.

Also, I tried capturing straight to MPEG-2 in WinTV but the picture was bad, with horizontal lines flashing in it.

[sanlyn]

If you do find some cleanup is needed, work in AVI before you encode .

[manono]

It would be great if you provided me with any guide, thanks.

So, you have an AVI with the audio already in it? Then the cutting would go something like this:

AVISource("Video.avi")
Trim(100,500) ++ Trim(700,1000) ++ Trim(1500,1800)...

The trims tell which frames you're keeping. Open the original script in VDub(Mod) and scroll through finding the places to be removed, and then set up the trims to keep what you want. the '++' is an aligned splice which will keep the audio in synch.

After that's all done I'd demux the audio to be reused in the authoring stage. Then send the cut up AVI to HCEnc:

AVISource("VideoCut.avi",False)
ConvertToYV12()

The 'False' removes the audio from the AVI and the 'ConvertToYV12()' converts the colorspace as I don't think HCEnc accepts RGB input. That line may not be necessary, depending on the colorspace of your HuffYUV AVI.

Make sure your max video bitrate isn't set above 8000 or maybe 8400 or so.

http://avisynth.org/mediawiki/Trim
http://avisynth.org/mediawiki/Splice

[alesco]

Thanks. Hope the colour converting is not necessary since the avi is in rgb (not sure why, probably determined by the equipment).

I will report back to let you know how it worked.

[manono]

Thanks. Hope the colour converting is not necessary since the avi is in rgb

If using HCEnc and if the AVI is really in RGB (put 'Info' in the script if you're not sure), you'll have to convert as I don't believe HCEnc even accepts RGB input. You can try, but it'll most likely throw an error. All DVDs are YV12.

[sanlyn]

alesco, that's something of a mysery to me that you specify YUY2 for output but get RGB.