USB Capture for VHS and DV8

[Leanoric]

Getting rid of DV codec is not just a benefit, it is a necessity. If you want to de-interlace, denoise, color correct, upscale and more, you would have to do it outside the DV codec. When you use NLE software to edit a DV footage "supposedly in the DV domain", the software decodes the DV into an uncompressed video, does the processing and encode back to DV, it just doesn't tell you that. But we all know script software are better and more transparent than NLE software, this is what the OP is doing.

So dv captures should be encoded straight to huffyuv, and then any post processing work that needs to be done should be on the huffyuv encode?


Ive been storing my dv captures with the dv codec and havent noticed any problems post processing them with that codec so far?

[lollo]

Ive also been using the default RGB24 pixel format to encode all of my captures so far. Is this going to be a problem?

You are introducing a lossy conversion to RGB, small: https://forum.videohelp.com/threads/359657-If-I-convert-from-YUV-to-RGB-I-loss-quality

Respect the input range 16-235 before colorspace conversion: http://avisynth.nl/index.php/Convert

So dv captures should be encoded straight to huffyuv

Not at all. Only if you need an intermediate file in the middle of your post-processing (which is quite uncommon), you save it as lossless (i.e. HuffYUV)

[DB83]

TBH (and my last post in this thread unless...) I am still struggling with the OP's work flow (if I follow other comments directly)

So did the OP do the de-interlacing, cropping, resizing blah.. having merely loading the original DV BEFORE the ouput to huffyuv (whatever the other settings)


And, if this is so, are there no visual issues ? (which I read could be - in the day I would transfer as DV yet use software to 'convert' to DVD without any obvious problem and had no feedback from my 'customers' that there were problems -)


Apologies for the 'pedantic' nature of this post yet I believe we all should be true for the benefit of those that follow and, hopefully, read this rather than 'reinvent' the wheel again.

[Leanoric]

Ive also been using the default RGB24 pixel format to encode all of my captures so far. Is this going to be a problem?

You are introducing a lossy conversion to RGB, small: https://forum.videohelp.com/threads/359657-If-I-convert-from-YUV-to-RGB-I-loss-quality

Respect the input range 16-235 before colorspace conversion: http://avisynth.nl/index.php/Convert

This is disheartening to hear as I have encoded many hours of captured tape using the RGB pixel format.

Does this tape need to be recaptured and will there actually be a noticeable difference to the naked eye if I recapture and encode with the correct setting?


[Attachment 75525 - Click to enlarge]

Can you let me know what which option to select please?

[lollo]

Small/ no visible loss, no need to recapture.
4:2:2 YUY2.

[dellsam34]

I don't think it's worth recapture unless you are anal about it. Notice when you choose 4:2:2 YUV2 the options on the right for YCbCr will be active and when you are outputing SD leave everything on "No change", If outputing HD (1440x1080) or higher tick Rec.709, I usually leave component range alone but some members may have a suggestion for HD.

[dellsam34]

So dv captures should be encoded straight to huffyuv, and then any post processing work that needs to be done should be on the huffyuv encode?

Not necessarily HuffYUV, any lossless form. If cutting scenes or rearranging them you don't need to strip DV, But any other processing would need the DV to be stripped away especially if encoding to another format is the goal. If the software you are using doesn't tell you, it doesn't mean it's doing it within DV. But I'm not well versed into software, I know there are members here more qualified than me to speak in the matter.

[Leanoric]

Ok thats good to hear, I wasnt looking forward to have to recapture all thoses tapes.


Ive actually processed several dv captures (cropped, deinterlaced etc) and encoded to mp4 using vdub2 and hybrid, that was without removing the dv codec. I didnt encounter any problems and they all seem to look fine, that said, Im no expert and you guys might see a ton of issues with them.

[lollo]

Ive actually processed several dv captures (cropped, deinterlaced etc) and encoded to mp4 using vdub2 and hybrid, that was without removing the dv codec.

Then you do need to save any intermediate file I suppose. Opening the file in VirtualDub2 will decode DV video to apply its filtering (cropping and deinterlacing). Then you encode to mp4 still in VirtualDub2.

dellsam34 in his flow sometimes needs to save an intermediate file for further purposes. Different scenario.

BTW, if you use an AviSynth approach and a mp4 encoding with a command line as in one of my previous post, you can't introduce any error/mis-processing as with the VirtualDub2 (useless) GUI, where you should have been checked/understood all options prior to use it.

BTW2: ALWAYS keep the original capture!!! as you have experimented even a simple processing may introduce side effects

[Leanoric]

Ive actually processed several dv captures (cropped, deinterlaced etc) and encoded to mp4 using vdub2 and hybrid, that was without removing the dv codec.

Then you do need to save any intermediate file I suppose. Opening the file in VirtualDub2 will decode DV video to apply its filtering (cropping and deinterlacing). Then you encode to mp4 still in VirtualDub2.

dellsam34 in his flow sometimes needs to save an intermediate file for further purposes. Different scenario.

BTW, if you use an AviSynth approach and a mp4 encoding with a command line as in one of my previous post, you can't introduce any error/mis-processing as with the VirtualDub2 (useless) GUI, where you should have been checked/understood all options prior to use it.

BTW2: ALWAYS keep the original capture!!! as you have experimented even a simple processing may introduce side effects

OK thanks, I'll do it the correct way from now on. Maybe one day Ill give avisynth a go but I dont think Im quite up to the challenge yet. I do keep all original captures though, on 2 different hard drives.


BTW, should I also remove the dv codec before using programs like dvdstyler?

[lollo]

I do keep all original captures though, on 2 different hard drives.

So why did you talk about re-capturing? The loss you introduced with the RGB conversion was in post processing. If you have the original YUV 4:2:2 lossless interlaced capture restart from that.

BTW, should I also remove the dv codec before using programs like dvdstyler?

No idea. DVD is MPEG2, so an encoding from DV to MPEG2 is needed. I ignore if that program does that. If so, it is probably able to "read" the DV encoded file and convert it to MPEG2 DVD standard.
If the input to dvdstyler is the post-processed file generated by VirtualDub2 you are maybe introducing too many lossy steps (lossless capture -> mpeg4 (???) -> mpeg2). Detail you workflow if you need assistance

[Leanoric]

I do keep all original captures though, on 2 different hard drives.

So why did you talk about re-capturing? The loss you introduced with the RGB conversion was in post processing. If you have the original YUV 4:2:2 lossless interlaced capture restart from that.

Because after I capture, I tidy up the video by trimming the first and last few seconds of it in vdub2. When I save that file, I have to select lagarith again and it re-encodes it.

Is there a way of trimming the start and end without encoding? Ive tried to do it with direct stream copy but couldnt get it to work.

I know that avidemux can do it but it doesnt have lagarith as an option and thats what I capture with.

[lollo]

Is there a way of trimming the start and end without encoding?

Yes. https://www.weethet.nl/english/video_cutavis.php

What did not work for you?

[JonnyAlpha]

TBH (and my last post in this thread unless...) I am still struggling with the OP's work flow (if I follow other comments directly)

So did the OP do the de-interlacing, cropping, resizing blah.. having merely loading the original DV BEFORE the ouput to huffyuv (whatever the other settings)


And, if this is so, are there no visual issues ? (which I read could be - in the day I would transfer as DV yet use software to 'convert' to DVD without any obvious problem and had no feedback from my 'customers' that there were problems -)


Apologies for the 'pedantic' nature of this post yet I believe we all should be true for the benefit of those that follow and, hopefully, read this rather than 'reinvent' the wheel again.

Here's my workflow:
1. Capture DV using Win DV (Aspect Ratio is 4:3, Resolution is 720 x 576, 25 FPS, Interlaced)
2. Open in VirtualDub2
3. Apply the following filters:

• DeInterlace
• De-Noise - Using MSU De-Noiser (Set scale to around 80)
• Resize to 768x576

4.Video set to Full Processing Mode
5. Save Video Compression Huffyuv Pixel Format YUYV, YUY2

Look and sound ok in Windows Media Player. If you open them in Virtual Dub the sound starts and stops?

[JonnyAlpha]

What I need now is a recommendation for a free video editor that I can join the clips together, maybe add a Title, export to mp4.

[Leanoric]

Is there a way of trimming the start and end without encoding?

Yes. https://www.weethet.nl/english/video_cutavis.php

What did not work for you?

The direct stream copy function does not work for me, I cant remember the exact message but I would get an error message everytime I tried to use it, so I just ended up having to encode again after each capture.

Thats using vdub2. Ill have another go using vdub as thats what theyre using in the guide you linked.

[lollo]

Thats using vdub2. Ill have another go using vdub as thats what theyre using in the guide you linked.

VirtualDub2 (version 44282) works well for me while doing trimming with direct stream copy option on a HuffYUV capture.

[Sharc]

Thats using vdub2. Ill have another go using vdub as thats what theyre using in the guide you linked.

VirtualDub2 (version 44282) works well for me while doing trimming with direct stream copy option on a HuffYUV capture.

I tried one of the earlier lagarith samples of the OP (recovered from the trash can) and trimming with direct stream copy works here as well.
@Leanoric: An issue with the audio? What is the audio format of your captures?

[Alwyn]

@Dellsam43, again (post 84), can you confirm that when you mention YUV2 you actually mean YUY2?

I think we're off on a tangent. Capturing with HUFF or LAGS always results in YUV colour space, not RGB (at least here it does). In capture mode, you can't choose RGB in HUFF. In LAGS you can, but it is ignored by AmarecTV and VDub, which captures YUV.

The dialog box in post 87 (with the various colour spaces and Rec 601/709) is not a capture screen; it's an output/save as screen.

As for needing to convert DV into lossless for editing in an NLE, if an NLE needs to internally convert it to work with it, so be it. One can import DV straight into AVISynth; why is there a need to convert it to lossless? Surely AVISynth can deal with DV just as well as with Lossless?

BTW, if you use an AviSynth approach and a mp4 encoding with a command line as in one of my previous post, you can't introduce any error/mis-processing as with the VirtualDub2 (useless) GUI, where you should have been checked/understood all options prior to use it.

What, AVISynth and FFMPEG don't have any options to check/understand? Of course they do and they are way more complex and difficult than a GUI. What I've learn't about AVISynth shows me that there are fatal traps and gotchas at every turn with it, far more than a GUI. You don't like GUIs. Fair enough. But don't belittle the hundreds of thousands of people who do.

@Leanoric, in that guide:

Then I start moving the slider to right again, using the "Key next" button

is not necessary because the file you're editing is lossless (HUFF, LAGS); every frame is a keyframe and you can cut on the precise frame.

The "keep" bar is also blue, not grey. You can also use the arrow keys to position the Play marker, and delete small sections by highlighting them eg the front noise; highlight the area and use the Edit menu to "crop to selection" or "delete".

[dellsam34]

The direct stream copy function does not work for me, I cant remember the exact message but I would get an error message everytime I tried to use it, so I just ended up having to encode again after each capture.

Thats using vdub2. Ill have another go using vdub as thats what theyre using in the guide you linked.

Full processing mode does not always have to be lossy as long as you are not selecting a lossy codec at the output, You can trim the captured files and use full processing mode if the parameters are the same at the input and output and the entire operation will be lossless. Say you captured uncompressed lossless AVI with and put it in to trim and want to output as HuffYUV, you have to select full processing mode in this case and choose HuffYUV as output and it will still be lossless.

[lollo]

What, AVISynth and FFMPEG don't have any options to check/understand? Of course they do and they are way more complex and difficult than a GUI. What I've learn't about AVISynth shows me that there are fatal traps and gotchas at every turn with it, far more than a GUI

What you have not learned about AviSynth is that it does not perform any colorspace conversion unless you explicitely insert it, it does not hide any manipulation and, in general, it is more "transparent" than a GUI.
The fatal traps and gotchas you refer to are user misunderstanding, and the optimization of the parameters are related to the filters in usage to improve the outcomes.

But don't belittle the hundreds of thousands of people who do.

I do not belittle anyone, but when I see a restoration with VirtualDub deinterlacing + MSU De-Noiser + Resize to 768x576 (nothing against you JonnyAlpha, just my thoughts), I wonder why some of us lose time to promote a quality processing rather that your minimalist approach.

My bad, as you said hundreds of thousands of people prefer the second, so you're absolutely right!

[Leanoric]

Well Ive just tried using direct stream copy, and it works! I definitely remember getting error messages when trying it before. That was about a year ago though when I first started capturing and using vdub, so I was almost certainly just doing something wrong.

I really wish Id tried it again before now. Not only did it take a lot of time re-encoding all of my captures when I didnt need to but I also did them all with the wrong pixel format

[Alwyn]

Nobody's right or wrong, Lollo. Everybody has their own needs and standards.

I thought you might have had an apoplexy at DB 83 capturing in DV then going straight to DVD with a GUI! Oh my goodness, the humanity of it all!

[Sharc]

VirtualDub(2) is still very popular and presents for many a significantly lower entry threshold for basic video processing tasks, compared to AviSynth or VapourSynth.
The problem I see is that it hasn't been updated since long (well, one could call it "stable") and possibly will never be. Currently it has 338 open tickets, 66 open bugs and 102 open requests.
Read the conversation from 2022/2023 with the author starting here:
https://sourceforge.net/p/vdfiltermod/tickets/269/#1856

[DB83]

Nobody's right or wrong, Lollo. Everybody has their own needs and standards.

I thought you might have had an apoplexy at DB 83 capturing in DV then going straight to DVD with a GUI! Oh my goodness, the humanity of it all!

Putting words in to my mouth now. Unless you regard EVERY piece of 'software' - the actual word I used - as a GUI.


The actual process was more involved than that and has no relevance to this topic.


But as for the OP's workflow it does seems to fly in contradiction to what has been recommended by others. He appears not to be using a script, which would be processed by vdub/2 but direct vdub/2 filters with the source being DV. I have not a clue if vdub/2 actually uncompresses on the fly as is done in more standard NLEs

[Alwyn]

@Sharc, for what is being done here, there are no showstoppers in that bug list. And I squashed one.

as for the OP's workflow it does seems to fly in contradiction to what has been recommended by others.

Some others. One does not have to use a script; did you for your DVDs? I have always said that AVISynth will produce, in most cases, a better result, but at a cost. Is the result worth the effort? Perhaps not. For home videos that will be watched probably twice then pulled out again maybe after the person's died, then I would question the effort involved.

If you're a pro or have a bent for this sort of thing, by all means, get into that code and make that video sizzle. What some do here with AVISynth is mind-bogglingly good, but totally out of my league and I could never hope to come close, even with weeks of study. I have got QTGMC going and use it extensively, but it was hard work, not helped by lack of guidance and the complexity of the setup.

So I don't think there's an issue with the OP's workflow: they've got a nicely deinterlaced, correctly-sized file that they can save to MP4 and distribute immediately or save to Lossless and continue editing in an NLE.

As for taking DV files, converting them to lossless, then working on them, I would like to see an example comparing the difference in results.

[JonnyAlpha]

Well, following advice from here, I am currently waiting for my .mp4 export to finish in Shocut!

I had to watch a couple of videos to get the idea on how it all works and managed to import all of my edited .avi clips into the playlist, add them to the timeline ready for export.

The export format threw me a little, at first I selected MPEG-4 and exported successfully, however this was about 6GB. It have crashed with about 10 minutes to go so I had to redo it. The message I received was that qmelt.exe had stopped working? Think it was caused by the virus checker, i disabled everything from starting like AVG, Magic ISO, Adobe Checker etc and rebooted and tried again.

When I went to export it again I saw that the last export was actually a .avi, so selecting MPEG-4 export exports to .avi, not .mp4. This was confirmed when I hit the advance settings which shows you the frame rate and resolution etc.
After poking around in a few other export modes I selected YouTube, which results in a .mp4 file at 50FPS and the same resolution as I have imported 720 x 576 IIRC.

I am running all this on an old AMD Athlon 64x2 Bit machine that I rebuilt running Windows 7 32Bit. I had to install an older version of Shocut as they don't do a 32 bit version anymore.

[Bwaak]

"selecting MPEG-4 export exports to .avi, not .mp4" - codec and container are different things. Apples and apple crates.

[JonnyAlpha]

Well looks like there's a problem. Tried exporting to .mp4 using the export for YouTube and it got to 70% and Shocut stopped again.
Qmelt.exe has stopped working.

Seems either it's reached a limit for saving or there's another problem.
I may have to split the clips and do 2 x .mp4 files

[Alwyn]

I'd only be using the latest 64bit version if I were you. You've got an old machine, on an old OS, using an outdated version; Sep 2020 is a long time ago.