VCR backup: did anyone of you really used it?

[Hewitson]

Originally Posted by TroyWilkins

I could be wrong here, but I suspect the data density on the physical medium has something to do with it. I mean, 1.2MB on a high density 5.25" PC disk, or 1.44MB on a high density 3.5" PC disk. The physically smaller disk has more data packed into a smaller physical space. This even counts for the double density disks that most Amigas used, 880KB on a 3.5" Amiga disk requires more dense storage than 1.2MB stored on a 5.25" disk.

This is of course true but you'd still think that with a fully exposed media the 5.25 would be far more unreliable than a 3.5, which obviously has a shutter protecting the media when not in use.

Originally Posted by Gorf

At least in Europe and a SCART enabled VRC it would not be a problem to use the chroma signal for data (or redundancy / error correction) as well.

Why wouldn't it? The video information is still stored on the tape as composite video. I just don't think the chroma signal would have the accuracy needed for something like this.

[Gorf]

Originally Posted by Hewitson

Why wouldn't it? The video information is still stored on the tape as composite video. I just don't think the chroma signal would have the accuracy needed for something like this.

because SCART* or s-video have a separated Y/C signal lines.
(*SCART in this case is actually RGB that gets translated into Y/C by the VRC)
While it is still stored on tape as a modulated signal, it has a much better signal to noise ratio and less artifacts.
So I do not really see a big technical problem here to e.g. use two different Y signal-states in addition to the two C values (black and white), which would double the density.

For comparison:
Some SVHS VRCs were able to store 48kHz 16-bit stereo digital audio on the audio-track next to the video information. This means 4 bytes x 48k x 3600 / 1024 equals 675 MB per hour! Only on the audio-track, which only occupies a small fraction of the tape.


DTRS and DA-88 use Hi8-video-tapes to store multiple digital audio tracks (24bit and up to 192 kHz) - typically 4,8 GB per tape

[idrougge]

Originally Posted by Gorf

because SCART* or s-video have a separated Y/C signal lines.

Not when connected to a VHS it doesn't. Scart can carry all kinds of video, but that doesn't mean that it does in every case. When connected to a VHS recorder, all you get is composite video.

[Gorf]

Originally Posted by idrougge

Not when connected to a VHS it doesn't. Scart can carry all kinds of video, but that doesn't mean that it does in every case. When connected to a VHS recorder, all you get is composite video.

That depends on your VRC or more general on your setup. When you connect an Amiga via a D-sub to SCART cable, then it is definitively RGB.
(Not all VRCs will be able to record with this setup.)

[Sandancer]

can you still get the cable for to do it as I know of 27-30 tapes have been giving away a few months ago. I'd love to get them and archive for us all. I still have a VHS recorder

[Jope]

Hardly anyone had the original interface, but the schematics were floating around on BBSes back then.

They are probably still available in some BBS dump.

I can also dump the VBS tapes if you can get them to me, I have the tools for this.

[BarryB]

Here's what's inside my VBS hardware:



Second pic shows where the traces go!

Looks to be 6 components that can't be identified, one has the print scratched off, 3 look to be soldered with the values on the other side of the component so they are hidden, the other 2 don't seem to have any values?

[Jope]

Wow, that's advanced. All of the ones I've seen (all were clones) were through hole on vero board..

[Turrican_3]

I still have my VBS interface somewhere, IIRC was even smaller than that... should see if I can disassemble without destroying it to see how it looks inside.

[fryguy]

I found this http://leszczamiga.ppa.pl/elektronika/vbs.gif

[nogginthenog]

I almost bought one. I was living in Eindhoven (NL) as a student (1993?). Got chatting with the author (who also lived in Eindhoven) probably on Usenet.

Found the name: Lyppens Software Productions

[AmigaHope]

Might be worth it just to reverse-engineer the VBS backup format in the software, then decode the videos in software. The actual data stream in VBS seems to be fairly straightforward (4 bytes per line separated by pulses of black, plus a period sync mark) so you'd just have to figure out how the VBS software encodes its compression/error-corection, and that was almost certainly all done in software with the physical device just laying out the bits into the video stream.

[sTe]

I had one of those VBS devices ... only recently I discovered it had a feature I believe for backing up floppy disks to tapes. Used it primarily back in the day for backing up my Amiga HD.

I did reach out to the author to show interest to see if he would give access to source code / github or Amiga Preservation site. He showed interest, but then I got the impression he wasnt interested, after a few further emails.

Great little grey device.

[Gardhul]

Never used, I barely remembered its existance

[pandy71]

Originally Posted by BarryB

Here's what's inside my VBS hardware:



Second pic shows where the traces go!

Looks to be 6 components that can't be identified, one has the print scratched off, 3 look to be soldered with the values on the other side of the component so they are hidden, the other 2 don't seem to have any values?

IC is fast comparator - probably LM311 - i made functionality identical interface with µA710 - it was faster than 311 but required bipolar supply (not a problem in Amiga as it is used anyway in serial port) - format is approx direct UART so Paula can properly detect and "receive" single byte per line (so theoretical performance is somewhere around 15Kbytes per second). Resistors are flipped to hide markings and capacitors are unmarked. Probably instead comparator, modern, fast OPAMP can be used also.

Originally Posted by sTe

I did reach out to the author to show interest to see if he would give access to source code / github or Amiga Preservation site. He showed interest, but then I got the impression he wasnt interested, after a few further emails.

Great little grey device.

There is many software implementations of this but all probably are somehow compatible - video backup was also present on PC. Weirdly H. Lyppens pursuit people creating open source software implementations - strange as he for sure didn't invented this idea - PCM processors to store PCM Audio (digital tape deck) was commercially available since end of 70's there is few projects dedicated this technology: https://en.wikipedia.org/wiki/PCM_ad...useskin=vector .

Originally Posted by Gorf

because SCART* or s-video have a separated Y/C signal lines.
(*SCART in this case is actually RGB that gets translated into Y/C by the VRC)
While it is still stored on tape as a modulated signal, it has a much better signal to noise ratio and less artifacts.
So I do not really see a big technical problem here to e.g. use two different Y signal-states in addition to the two C values (black and white), which would double the density.

Very late reply - VCR by definition store chrominance separately from luminance signal but chrominance has very low bandwidth - approx +-500kHz - this gives approx 40 lines resolution so approx 6 times less than luminance resolution so chrominance will be quite poor way to store data - also data decoding will be much more complicated. You could use luminance channel way more efficiently but this is beyond Amiga capability

[Toni Wilen]

"receive" single byte per line

VBS used 4x9 bit data words per line. (1 start, 9 data, 1 stop)

[pandy71]

Originally Posted by Toni Wilen

VBS used 4x9 bit data words per line. (1 start, 9 data, 1 stop)

Oh, i should have 2 or 3 tapes with backup data from Amiga but they was recorded like 30 years ago and have no working VCR so probably you are right on this - my memory was like single byte per line (counting static stripes for start/stop) - so if you say 4 bytes this mean approx 409kbps UART Rx speed so my assumption is CPU works in pooling mode during reception to make this possible... Also H. Lyppens wrote that VBS on A500 is approx 40MB per hour so this is more on less inline with my estimation - 256 lines (1 byte per line), 50 frames per second - this gives 12800 bytes per second i.e. approx 40 MB per hr. For your 4 bytes per line this gives us 180MB per hr.

[Toni Wilen]

Originally Posted by pandy71

Oh, i should have 2 or 3 tapes with backup data from Amiga but they was recorded like 30 years ago and have no working VCR so probably you are right on this - my memory was like single byte per line (counting static stripes for start/stop) - so if you say 4 bytes this mean approx 409kbps UART Rx speed so my assumption is CPU works in pooling mode during reception to make this possible... Also H. Lyppens wrote that VBS on A500 is approx 40MB per hour so this is more on less inline with my estimation - 256 lines (1 byte per line), 50 frames per second - this gives 12800 bytes per second i.e. approx 40 MB per hr. For your 4 bytes per line this gives us 180MB per hr.

It sets SERPER to about 500000 bits per second (don't remember exact value) and video had clearly visible 4 static vertical bars (START bits).

It probably has error correction or data is repeated or something.

[pandy71]

Originally Posted by Toni Wilen

It sets SERPER to about 500000 bits per second (don't remember exact value) and video had clearly visible 4 static vertical bars (START bits).

It probably has error correction or data is repeated or something.

You are right - i see it here [ Show youtube player ]

[modrobert]

By modernizing the idea, higher resolutions, piping raw data pixelated through 'ffmpeg' or similar, it should be possible to use YouTube as free public cloud storage then?

-"What are you watching?"
-"TOSEC..."