USB Capture for VHS and DV8

[lollo]

What I am struggling to find is a stepo by step guide on editing DV and then exporting for general viewing

Basic AviSynth script:

AviSource("file_name.avi") # or whatever source reader is appropriate
crop(crop_left, crop_top, -crop_right, -crop_bottom)
AssumeBFF()
QTGMC(preset="slow")
addborders(crop_left+crop_right)/2,(crop_top+crop_bottom)/2,(crop_left+crop_right)/2,(crop_top+crop_bottom)/2)

encode the avs file with ffmpeg in a command line:

ffmpeg.exe -i <file_name>.avs -c:v libx264 -crf 17 -preset slow -aspect 4:3 -c:a aac -b:a 128k output.mp4

Once you get familiar with the basic, you can introduce more complex processing.

[Alwyn]

What I am struggling to find is a stepo by step guide on editing DV and then exporting for general viewing

Use exactly the same procedure, as for analogue AVI captures: Deinterlace, crop, edit/adjust, resize. You will need much less cropping for DV, if any at all. Also, if you've shot 16:9 DV, you'll need to resize it to 1024x576.

This confused me as I only just found that cropping is accessible from within the filters list

Spotto!

Lollo, you're contradicting yourself:

Cropping should be before deinterlacing

The cropping before deinterlacing can introduce a problem if you do not respect the appropriate mod according to the source.

[lollo]

Cropping should be before deinterlacing

The cropping before deinterlacing can introduce a problem if you do not respect the appropriate mod according to the source.

Lollo, you're contradicting yourself

Only if you do not understand the basics of cropping and deinterlacing.

[Alwyn]

Would you mind enlightening us?

[lordsmurf]

Lollo, you're contradicting yourself:

Cropping should be before deinterlacing

The cropping before deinterlacing can introduce a problem if you do not respect the appropriate mod according to the source.

I don't see contradiction in his statement.***
- Crop before deinterlace.
- Not cropping before deinterlace introduces problems.

*** I didn't read the whole thread. But it needs to be mentioned that SOMETIMES you only crop after, and SOMETIMES you only crop before. It really depends on factors. It's one of those blanket statements that, read wrong, read out of context (like not reading every post in a thread), leads to massive quality loss. And let's face it, everybody skims, few read every word. (I'm not sure if lollo is correct or not here, but I generally trust his Avisynth scripting, so benefit of the doubt here, not having read it myself.)

[Alwyn]

But it needs to be mentioned that SOMETIMES you only crop after, and SOMETIMES you only crop before. It really depends on factors.

We're here to learn. Is it possible to even roughly detail when? I have always cropped after deinterlacing and I have never had "massive" quality loss.

[lollo]

The crop before deinterlacing helps to avoid flickering which may happen sometimes on top and bottom scanlines (especially if you work with material with head swithing noise, introducing bad motion vectors at the border for deinterlacer processing routines). It also reduces computation time, but that's marginal.

When cropping before deinterlacing YV12 material (which is generally not a standard for an Analog Capture), be careful to crop using mod4 on the vertical axis to avoid blurring of chroma.

It was just that, I had no wish to be pedantic

edit: for YUY2 crop mod2 vertically, for the same reason

[JonnyAlpha]

What I am struggling to find is a stepo by step guide on editing DV and then exporting for general viewing

Use exactly the same procedure, as for analogue AVI captures: Deinterlace, crop, edit/adjust, resize. You will need much less cropping for DV, if any at all. Also, if you've shot 16:9 DV, you'll need to resize it to 1024x576.

This confused me as I only just found that cropping is accessible from within the filters list

Spotto!

Lollo, you're contradicting yourself:

Cropping should be before deinterlacing

The cropping before deinterlacing can introduce a problem if you do not respect the appropriate mod according to the source.

How can I find out if Iv'e shot 16:9 aspect ratio?

What I meant above was that I search the menus for cropping but could not find it, then re read your guide and realised cropping appeared in the filter list when you click on a filter

[lollo]

What I meant above was that I search the menus for cropping but could not find it, then re read your guide and realised cropping appeared in the filter list when you click on a filter

If you want crop only in VirtualDub, you can add the filter "null transform", which does nothing and enables cropping feature.

[Alwyn]

If you want crop only in VirtualDub, you can add the filter "null transform", which does nothing and enables cropping feature.

That's only relevant for VDub 1.x. As stated, VDub 2 has a dedicated Cropping filter.

[lordsmurf]

If you want crop only in VirtualDub, you can add the filter "null transform", which does nothing and enables cropping feature.

That's only relevant for VDub 1.x. As stated, VDub 2 has a dedicated Cropping filter.

The problem is that VirtualDub2 cropping filters can have nasty side effects. I reported it several times, with details and samples, and nothing was ever done about it. It's not to be trusted.

[JonnyAlpha]

So feeling brave, I added the filters, (Deinterlace, Denoise, Resize to 768x576), save video options Huffyuv v21.1. Everything else left as standard - Pixel Format RGB24, No Compression. Video set to Full Processing Mode.
I saved the first 7 as a job que. When I hit Job Control / Start, everything apart from the last one immediately went to status Error. The only one that ran was the last one.

Any ideas?


[Attachment 75494 - Click to enlarge]

UPDATED
Sorry, before I ran the job, I moved the capture files to a different folder DOH!!

[Alwyn]

So feeling brave

That's the spirit!

Bummer on the errors.

Hit F8 (Go>Log). I've never read it but there may be something in there.

[dellsam34]

Don't use RGB24, use YUV2 instead.

[DB83]

You started with a DV capture ? I fail to see the benefit of editing to huffyuv. The lossless result comes from a lossy source so there is little benefit. And much larger video files.

[dellsam34]

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.

[JonnyAlpha]

So feeling brave

That's the spirit!

Bummer on the errors.

Hit F8 (Go>Log). I've never read it but there may be something in there.

Alwyn - read the update - I had moved the source files into a different folder - hence the error - all sorted now.

[JonnyAlpha]

Don't use RGB24, use YUV2 instead.

Iv'e already done a load using RGB24 - what would the benefits be in doing them again using YUV2?
Happy to do them all again if needed.

[JonnyAlpha]

You started with a DV capture ? I fail to see the benefit of editing to huffyuv. The lossless result comes from a lossy source so there is little benefit. And much larger video files.

I was following advice from Dellsam34 here and Alwyn here

[JonnyAlpha]

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 stick with Huvyuff?

[DB83]

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.

Fairy nuff. But the way I read it is that ALL the processing is done at DV level with only the final encode to lossless. Surely you would require two encodes. The first a bare conversion. Then the post-processing before a second encode.

[Bwaak]

But the way I read it is that ALL the processing is done at DV level with only the final encode to lossless. Surely you would require two encodes. The first a bare conversion. Then the post-processing before a second encode.

Maybe some NLEs can do some operations on compressed footage, but normally they uncompress it first. Some NLEs can do it on the fly like Vegas, other convert it to an intermediate format first (old FCP and iMovie, not sure whether they still do it). IDK whether they can apply effects to the intermediate format directly, or do they still need to uncompress fully.

In Vegas, you can switch from 8-bit integer to 32-bit floating point. As the Help file explains, "The 32-bit floating point settings allow greater precision for processing video, but require significantly more processing power than working with 8-bit video". I suppose when there are tons of effects, the consecutive rounding errors when dealing with integers can become noticeable compared to processing as float and then converting back to integer. Hmm, I probably should switch to float, especially for 10-bit Cineform files.

[Alwyn]

How can I find out if Iv'e shot 16:9 aspect ratio?

Sorry I missed this. There are a couple of ways: play it in VLC player and the "wide" 16:9 shape will be obvious. Or you can open the file in Mediainfo View>Tree or Text (a great file info program) and the shot ratio will be shown in the video data:


[Attachment 75509 - Click to enlarge]

To clarify my post-filter workflow: I export to lossless eg HUFF, LAGS, Magic only so that I can import into my "proper" video editor in the best quality for further editing (titles, sectioning, transitions). I then export out as MP4 for distribution/consumption.

If I have no need for further external editing, I export out from Virtual Dub as MP4, as per my guide.

[dellsam34]

Iv'e already done a load using RGB24 - what would the benefits be in doing them again using YUV2?
Happy to do them all again if needed.

YUV2 is compatible with most script software, Outputing HuffYUV is just to reduce the overall size of the lossless files, It can also be uploaded directly to youtube without issues.

[JonnyAlpha]

How can I find out if Iv'e shot 16:9 aspect ratio?

Sorry I missed this. There are a couple of ways: play it in VLC player and the "wide" 16:9 shape will be obvious. Or you can open the file in Mediainfo View>Tree or Text (a great file info program) and the shot ratio will be shown in the video data:


[Attachment 75509 - Click to enlarge]

To clarify my post-filter workflow: I export to lossless eg HUFF, LAGS, Magic only so that I can import into my "proper" video editor in the best quality for further editing (titles, sectioning, transitions). I then export out as MP4 for distribution/consumption.

If I have no need for further external editing, I export out from Virtual Dub as MP4, as per my guide.

OK cool, I do need to do extra editing as I need to join all the clips together, before I sporting to mp4. What would be a recommended free editor to do this?

[JonnyAlpha]

Iv'e already done a load using RGB24 - what would the benefits be in doing them again using YUV2?
Happy to do them all again if needed.

YUV2 is compatible with most script software, Outputing HuffYUV is just to reduce the overall size of the lossless files, It can also be uploaded directly to youtube without issues.

Ok thanks, I had a look back in the save video window where RGB24 is set as default, whilst there are lots of other options YUV2 is not one of them?

[Alwyn]

Don't use RGB24, use YUV2 instead.

Did you mean YUY2?


[Attachment 75517 - Click to enlarge]

[JonnyAlpha]

Don't use RGB24, use YUV2 instead.

Did you mean YUY2?


[Attachment 75517 - Click to enlarge]

Ah - didn't spot that

Would it be ok to re-open all of the clips that I have already clipped and filtered and re-save them without any filters again as YUY2? Then carry on with the remainder filtering and using YUY2 output format?

[Alwyn]

Would it be ok to re-open all of the clips that I have already clipped and filtered and re-save them without any filters again as YUY2?

I'm not a whiz at this colour stuff; better wait until Dellsam43 clarifies this.

[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?