Recovering VHS/Beta PCM Video DATA

[SF01]

If your VCR is a low end you may need a robust capture device, DV or non DV don't matter since chroma is not involved, However make sure the software you are using to recover PCM video accepts DV as an input format, As far as I know most of them work with lossless AVI, If they do work with DV then just get a camcorder that is capable of passthrough, Meaning it takes composite in and outputs DV out via iLink (a.k.a firewire), Or use a DV box that takes composite and spits out DV via firewire, You may need firewire 400 4pin to firewire 800 adapter and another firewire 800 to TB adapter if your computer lacks firewire port.

The Thinderbolt adaptor I have uses 9-pin FireWire port.
DV usually comes as AVI, so might work, at least on PC.

[dellsam34]

The 9pin called firewire 800? There is no camcorder or DV box comes with firewire 800 to my knowledge, so you would need to adapt a 4pin or 6pin to 9pin first.

[SF01]

The 9pin called firewire 800? There is no camcorder or DV box comes with firewire 800 to my knowledge, so you would need to adapt a 4pin or 6pin to 9pin first.

The Thunderbolt adaptor has 9-pin FireWire end.

[Cornucopia]

Strongly, strongly recommend against using compression in storage of pcm-on-video streams. Store them losslessly for short-term then decode them and store them as std pcm audio (eg. Wav).

Using dv or worse one of the mpegs will corrupt the feed. To us and to the encoder, the pcm stream looks and acts like video noise, which is notorious for being the hardest thing to encode unless you give it way high bitrate and/or leave it lossless, but there is actual signal in that "noise", and it will get lost, etc if not treated properly. That may be why some are experiencing stuttering, etc.


Scott

[SF01]

Strongly, strongly recommend against using compression in storage of pcm-on-video streams. Store them losslessly for short-term then decode them and store them as std pcm audio (eg. Wav).

Using dv or worse one of the mpegs will corrupt the feed. To us and to the encoder, the pcm stream looks and acts like video noise, which is notorious for being the hardest thing to encode unless you give it way high bitrate and/or leave it lossless, but there is actual signal in that "noise", and it will get lost, etc if not treated properly. That may be why some are experiencing stuttering, etc.


Scott

DV uses intraframe compression and 4:2:0 chroma subsampling, since there is no colour it will more or less only put tape noise in the chroma, the bitrate is 25, but experiments would need to be done to determine if DV is sufficient to record to/from.

[dellsam34]

The 9pin called firewire 800? There is no camcorder or DV box comes with firewire 800 to my knowledge, so you would need to adapt a 4pin or 6pin to 9pin first.

The Thunderbolt adaptor has 9-pin FireWire end.

But camcorders and DV boxes don't, hence needing two adapters.

[dellsam34]

Using dv or worse one of the mpegs will corrupt the feed. To us and to the encoder, the pcm stream looks and acts like video noise, which is notorious for being the hardest thing to encode unless you give it way high bitrate and/or leave it lossless, but there is actual signal in that "noise", and it will get lost, etc if not treated properly. That may be why some are experiencing stuttering, etc.


Scott

DV luma is supposedly lossless, But I find it hard to believe that any PCM software would have the capability to decode DV, and yes mpeg-2 would not work due to compression.

Decoding PCM can be done without saving the video at all, I've done it, but you have to make sure all the parameters are correct and the WAV audio saved is glitch free, otherwise you would have to play the tape again to get another file, But if storage is not an issue saving the AVI video is much safer so that way if something goes wrong you just need to run the file again instead of playing back the actual video tape.

[SF01]

The 9pin called firewire 800? There is no camcorder or DV box comes with firewire 800 to my knowledge, so you would need to adapt a 4pin or 6pin to 9pin first.

The Thunderbolt adaptor has 9-pin FireWire end.

But camcorders and DV boxes don't, hence needing two adapters.

There are hybrid connector cables out there, specifically 4-pin on one and and 9-pin on the other, I have one myself, no need for an adaptor, but there surely are 4-pon to 9-pin adaptors, I have one myself for 6-pin to 4-pin cable.

[SF01]

Using dv or worse one of the mpegs will corrupt the feed. To us and to the encoder, the pcm stream looks and acts like video noise, which is notorious for being the hardest thing to encode unless you give it way high bitrate and/or leave it lossless, but there is actual signal in that "noise", and it will get lost, etc if not treated properly. That may be why some are experiencing stuttering, etc.


Scott

DV luma is supposedly lossless, But I find it hard to believe that any PCM software would have the capability to decode DV, and yes mpeg-2 would not work due to compression.

Decoding PCM can be done without saving the video at all, I've done it, but you have to make sure all the parameters are correct and the WAV audio saved is glitch free, otherwise you would have to play the tape again to get another file, But if storage is not an issue saving the AVI video is much safer so that way if something goes wrong you just need to run the file again instead of playing back the actual video tape.

DV is stored as AVI file, from what I remember DV is either MPEG2, or JPEG2000, or something.

[Cornucopia]

@SF01 You are getting some things mixed up.

A DV stream coming over a dv cable is just a raw dv stream without a container, roughly equivalent to type1 DIFF.
A dv stream when captured and saved in a computer/device from a cable's stream can be:
1. Raw dv like mentioned above. No container. Usually with a .dv or .diff extension.
2. DV-AVI Type1
3. DV-AVI Type2
4. DV-MOV Type 1
5. DV-MOV Type 2
6. DV-MXF (differs on how the essence is encapsulated, but all should be equally compatible)

If you want to understand the difference between types 1 and 2, read my past posts.

DV uses the DV codec, a type of dct-based encoding. It is related to jpeg, jpeg2000, mpeg family, etc but only in some of the basic common algorithms. In its standard SD consumer variant, it has ~5.5:1 compression, as compared to something like jpeg2000 which is about 3:1, or mpeg2 which can be ~2:1 up to ~300:1.

Firewire 400 uses 4pin (no power) mini and 6pin (power optional) full connectors. 4/6 adapters are common, but forego the power, obviously.
Firewire 800 uses 9pin (power optional) full connectors.
There are 400 <--> 800 adapter cables, though all of them end up running at 400 speeds.

Dv runs on firewire isochronously at 25Mbps, expecting the 400a or 400b bandwidth.
Thus, when using some 400/800 adapters, there can be issues getting dv thru, but it SHOULD be possible.

@dellsam34, AFAIK DV luma is NOT lossless. Not sure where you got that.


Scott

[dellsam34]

There are hybrid connector cables out there, specifically 4-pin on one and and 9-pin on the other, I have one myself, no need for an adaptor, but there surely are 4-pon to 9-pin adaptors, I have one myself for 6-pin to 4-pin cable.

That's the second adapter I'm talking about, the other adapter is firewire to thunderbolt, Jees.

[dellsam34]

@dellsam34, AFAIK DV luma is NOT lossless. Not sure where you got that.


Scott

Nowhere, I mean at 25Mbps and 13GB/hr files had to be visually lossless enough to carry the PCM data with no problem, DCT compression is not that bad compared to say mpeg-2, The question is if the PCM software can understand DV codec or not.

[pandy]

@dellsam34, AFAIK DV luma is NOT lossless. Not sure where you got that.


Scott

Nowhere, I mean at 25Mbps and 13GB/hr files had to be visually lossless enough to carry the PCM data with no problem, DCT compression is not that bad compared to say mpeg-2, The question is if the PCM software can understand DV codec or not.

I know you hate me... but MPEG-2 use same DCT transformation as DV - SD MPEG-2 can go up to 15Mbps but also it can use motion estimation, P and B frames... so quality will be at least comparable to DV but it can be better than DV (by simply using semi-intra where GOP=2 and it is made from I,P frame) not mentioning other factors making DV inferior to MPEG-2.

[SF01]

@SF01 You are getting some things mixed up.

A DV stream coming over a dv cable is just a raw dv stream without a container, roughly equivalent to type1 DIFF.
A dv stream when captured and saved in a computer/device from a cable's stream can be:
1. Raw dv like mentioned above. No container. Usually with a .dv or .diff extension.
2. DV-AVI Type1
3. DV-AVI Type2
4. DV-MOV Type 1
5. DV-MOV Type 2
6. DV-MXF (differs on how the essence is encapsulated, but all should be equally compatible)

If you want to understand the difference between types 1 and 2, read my past posts.

DV uses the DV codec, a type of dct-based encoding. It is related to jpeg, jpeg2000, mpeg family, etc but only in some of the basic common algorithms. In its standard SD consumer variant, it has ~5.5:1 compression, as compared to something like jpeg2000 which is about 3:1, or mpeg2 which can be ~2:1 up to ~300:1.

Firewire 400 uses 4pin (no power) mini and 6pin (power optional) full connectors. 4/6 adapters are common, but forego the power, obviously.
Firewire 800 uses 9pin (power optional) full connectors.
There are 400 <--> 800 adapter cables, though all of them end up running at 400 speeds.

Dv runs on firewire isochronously at 25Mbps, expecting the 400a or 400b bandwidth.
Thus, when using some 400/800 adapters, there can be issues getting dv thru, but it SHOULD be possible.

@dellsam34, AFAIK DV luma is NOT lossless. Not sure where you got that.


Scott

Yes, I forgot there are Type1 and 2. I will check what the difference is, I use Scenalyzer, so I end up with AVI file, Type2 for easier editing in Vegas.

The sables yes, they can have any connector with pinout limited by the lower generation end, I have 4-4, 6-4, 4-9 cables, as well 4-9 adaptor that plugs to FW/TB adapor when I'm using 6-4 cable for my DSR-2000AP.

Whether DV is enough and can be encoded/decoded needas to be checked. I have a PCM tape, but it's Betacam, currently nothing to play it back on.

[dellsam34]

I know you hate me... but MPEG-2 use same DCT transformation as DV - SD MPEG-2 can go up to 15Mbps but also it can use motion estimation, P and B frames... so quality will be at least comparable to DV but it can be better than DV (by simply using semi-intra where GOP=2 and it is made from I,P frame) not mentioning other factors making DV inferior to MPEG-2.

I don't hate anyone, The mpeg-2 we know used in DVD and SD files has always been inferior to the DV codec used in camcorders, We are not talking hypothetical. Name one capture device that captures into mpeg-2 from VCR that is better than DV.

[SF01]

I know you hate me... but MPEG-2 use same DCT transformation as DV - SD MPEG-2 can go up to 15Mbps but also it can use motion estimation, P and B frames... so quality will be at least comparable to DV but it can be better than DV (by simply using semi-intra where GOP=2 and it is made from I,P frame) not mentioning other factors making DV inferior to MPEG-2.

I don't hate anyone, The mpeg-2 we know used in DVD and SD files has always been inferior to the DV codec used in camcorders, We are not talking hypothetical. Name one capture device that captures into mpeg-2 from VCR that is better than DV.

And yet all the FB groups pepple still praise ElCrapo.
At least NTSC DVD is 4:2:0, unlike DV. Though a variant of DVCPRO, or something in PAL is also 4:1:1.

[lollo]

Though a variant of DVCPRO, or something in PAL is also 4:1:1.

Which should tell you something.

4:1:1 is not that bad for SD sources, we just started a discussion here: https://forum.videohelp.com/threads/416930-Mini-DV-to-Digital-Transfer-Dissatisfaction#post2760878; will conclude when experimental data will be available.

[SF01]

Though a variant of DVCPRO, or something in PAL is also 4:1:1.

Which should tell you something.

4:1:1 is not that bad for SD sources, we just started a discussion here: https://forum.videohelp.com/threads/416930-Mini-DV-to-Digital-Transfer-Dissatisfaction#post2760878; will conclude when experimental data will be available.

It's worse than 4:2:0, at least 4:2:0 is more common in newer delivery systems, while 4:1:1 only exists in NTSC DV, when converting it to 4:2:0 DVD it effectively becomes 4:1:0. Then again, 4:2:0 loses an entire line of colour from the analog conversion. So it's a tradeoff.

"MiniDV to Digital", MiniDV already is digital. I use DV PAL to transfer analog video, it get's the job done the best ut of available 4:2:0 methods, to upgrade to 4:2:2 I would need more expensive equipment and a lot more digital storage for a client that ultimately would want a portable mp4 file that would be 4:2:0 anyway. With BW video such as PCM video it shouldn't matter.

[lollo]

It's worse than 4:2:0

questionable for SD low resolution like VHS and similar

at least 4:2:0 is more common in newer delivery systems

reasonable, DVD starts (in general) from nice SD sources.

Then again, 4:2:0 loses an entire line of colour from the analog conversion.

That's the point.

I use DV PAL to transfer analog video, it get's the job done the best ut of available 4:2:0 methods, to upgrade to 4:2:2 I would need more expensive equipment and a lot more digital storage for a client that ultimately would want a portable mp4 file that would be 4:2:0 anyway

That's not an excuse. If you start from better material, the final h264 will be higher quality, if that's targeted since the beginning.

[SF01]

Then again, 4:2:0 loses an entire line of colour from the analog conversion.

That's the point.[/QUOTE]

But you can only remain in 4:1:1 workflow to take advantage of that one advantage, granted DVDs are mostly obsolete now, ad nobody is doing BD SD anyore, but I wonder if YouTube and others support 4:1:1, or they convert to 4:2:0 to their internal storage.

I use DV PAL to transfer analog video, it get's the job done the best ut of available 4:2:0 methods, to upgrade to 4:2:2 I would need more expensive equipment and a lot more digital storage for a client that ultimately would want a portable mp4 file that would be 4:2:0 anyway

That's not an excuse. If you start from better material, the final h264 will be higher quality, if that's targeted since the beginning.[/QUOTE]

If I start with 4:2:0 and the final deliverable file is lower bitrate also 4:2:0, albeit with square pixels and trimmed analogue blanking from the sides, what is the benefit of an intermediate 4:2:2 file other than OCD satisfaction? I'm not saying my methid is perfect, or archival-grade, but Colin from video99.co.uk also uses PAL DV for clients saying it's sufficient enough.

I could invest in 4:2:2 equipment, but I doubt clients will be convinced enough to pay more.

[lollo]

But you can only remain in 4:1:1 workflow to take advantage

Yes, an additional 4:1:1 to 4:2:0 conversion is required if you want to deliver in that format.

If I start with 4:2:0 and the final deliverable file is lower bitrate also 4:2:0

I meant starting from YUV 4:2:2 lossless.

[SF01]

But you can only remain in 4:1:1 workflow to take advantage

Yes, an additional 4:1:1 to 4:2:0 conversion is required if you want to deliver in that format.

If I start with 4:2:0 and the final deliverable file is lower bitrate also 4:2:0

I meant starting from YUV 4:2:2 lossless.

Additional 4:1:1 to 4:2:0 conversion will result in 4:1:0, which is not the most desirable outcome now, is it? Since you completely discard every second line and using the top line subsampling to make it 4:2:0, while the 2 is simply that 1 from the first line made into two repeated samples.

It's not technically lossless if it's already 4:2:2, but we all know that.

But why settle on half-measures, how's that RF modded capturing "Domesday" project coming along for capturing and decoding VHS RF signal straight from the head? The demos are very promising, I've seen someone archives an entire Rocky mockbuster using this method, though in NTSC and into 23,976 instead of 24, or using PAL tape and then slow down to 24.

[lollo]

But why settle on half-measures, how's that RF modded capturing "Domesday" project coming along for capturing and decoding VHS RF signal straight from the head?

Great expectation, but so far it did not appear superior to the classic quality flow of high-end S-VHS with TBC, recommended capture card, YUV 4:2:2 lossless.

[SF01]

But why settle on half-measures, how's that RF modded capturing "Domesday" project coming along for capturing and decoding VHS RF signal straight from the head?

Great expectation, but so far it did not appear superior to the classic quality flow of high-end S-VHS with TBC, recommended capture card, YUV 4:2:2 lossless.

On the contrary. The demo I saw and that film restoration appeared quite better. But it might boil down to a matter of opinion.
Personally, I use a D-VHS deck for transfering and duplication.

[lollo]

On the contrary. The demo I saw and that film restoration appeared quite better. But it might boil down to a matter of opinion.
Personally, I use a D-VHS deck for transfering and duplication.

We need real data from a fair comparison, not versus a Funai VCR and an Easycap device. If you have any source, just post it

[SF01]

These are the videos I had in mind:
https://www.youtube.com/watch?v=pEzmbw_Y-Tw
https://www.youtube.com/watch?v=PbFJeBwWVEY
https://www.youtube.com/watch?v=4RcpMx6CRzU
The first one is "Ricky", a lesser known brother of Rocky, the other demo is from EP tape, the demo has sharper image, minimum HSN, the anime demo has less flattening, the letters are more readable and the smear is also less pronounced. Though I see how the less flattening might be considered by some as noise, rather than detail. The aliasing would need to be further developed however, though it could just be badly deinterlaced. But there is definitely less blur. The equipment could also be chosed better. But it's still in development phase, I suspect it may get updates.

[lollo]

I went quicly through the videos (I already know them), but:

the first I do not see a comparison

the second is not a fair comparison (no S-VHS VCR as pointed that out in the comments, for example)

the third is contradictory

There are long discussions here and on digitalfaq forum, involving many of us, but, once more, there is still no evidence nowdays about the superiority of vhs-decode to our recommended workflow.

[SF01]

I went quicly through the videos (I already know them), but:

the first I do not see a comparison

the second is not a fair comparison (no S-VHS VCR as pointed that out in the comments, for example)

the third is contradictory

There are long discussions here and on digitalfaq forum, involving many of us, but, once more, there is still no evidence nowdays about the superiority of vhs-decode to our recommended workflow.

The first is not a comparison however.

We will have to see how it unfolds. One thing the Decode might help solve is offloading PAL D-VHS without using the analog conversion.

[pandy]

I don't hate anyone, The mpeg-2 we know used in DVD and SD files has always been inferior to the DV codec used in camcorders, We are not talking hypothetical. Name one capture device that captures into mpeg-2 from VCR that is better than DV.

Once again intraframe MPEG-2 is not substantially different than DV especially when P frames can be used to reduce quantization factor in overall..
HW MPEG-2 encoders sucks in ME and B frames - for same bitrate there is no difference between MPEG-2 and DV but i think that MPEG-2 may deliver same or better quality with lower bitrate than DV.
Btw you still can use MP@HL and 25Mbps in MPEG-2 so it will be same as DV.

Finally i'm against capturing video with lossy coding solutions - they are OK for storing and distribution but not for workflow.

[lollo]

One thing the Decode might help solve is offloading PAL D-VHS without using the analog conversion.

That's would be a nice feature, assuming we can extract the raw digital data recorded! However, D-VHS was/is not that common.