2000 VHS to digitalise, what capture device?
 

Hello everyone,
I'm managing the multimedia library of a swiss high-school.

The guy who was working before me started to record TV documentaries, reports, and Super8/16mm on VHS from 1979. We already sent one doc to the national swiss TV who was missing document from the 80's.

I've more than 2000 VHS (and the same amount of DVD) to digitalise (maybe not everything is needed).

Actually, I was using an "AJA Io" and a "Director's Cut" (ANALOGUE TO IEEE-1394 CONVERTOR, TYPE TAKE 2).
Both are quit old stuff, linked to old Mac. I'd like to use actual devices with the new Mac Pro I've.

As a sound engineer, I need advices, please!

I've looked at Black Magic devices (H.264 or Intensity), but is there something more adapted (pro) to the amount of sources I have?

(For the VCR /VHS players, I've many different machine as a Sony SVO 5800p, a Bang & Olufsen 91.2, and many standard Sony/panasonic player, from the one hundred classroom that needed to be equipped. I've also cleaning tapes)

I hope I'll have the pleasure to read your answers.
Jeremy

Trust me the AJA Io setup you have is the best you'll ever have, For that amount of tapes plus the headache of newer hardware and operating systems you will go crazy, There is nothing wrong with an old operating system, VHS is older anyway. Just get a bigger hard drive or multiple small hard drives if the MAC system has a HDD size limit.

Jwillis has covered this piece of hardware in this thread.

Thanks for the quick reply!

I'll try again. I've dropped frame problem, sound sychronisation and at the end a 80 GB file after 4 hours of digitalisation.
Probably more a parameters (user) problem than the devices....! :)

EDIT:
Should I need a Sync generator?

No, unless you are editing from two VCR's, what you need is a decent VCR with line TBC and possibly an external frame synchronizer but I believe the Aja Io is built in some sort of frame synchronizer, I'm not a Mac guy so software wise Jwillis knows better hope he will chime in and give you some advice, By the look of it if you are going to capture lossless and edit and compress 2000 tapes is a labor of several years, So capturing straight to a lossy format may save you a lot of time.

Also there is a write up here:
https://forum.videohelp.com/threads/...e2#post2551466

And try this software:
https://www.aja.com/support/item/2498

The AJA IO was designed for "Broadcast quality" signal, its not for VHS capture.

You will be starting out with a severe disadvantage.

It is no surprise the problems your seeing.. even though the picture will look good.

The AJA IO hardware was designed by Grass Valley and relied on firmware to start the box up, the connected computer downloaded a copy of its firmware operating system every time its powered up.

The last editions of the firmware re-enabled the built-in TBC, but only for a very narrow set of OS X versions.

The AJA IO is limited by the tiny number of years it was sold and the tiny number of OS X versions it would work with.

Its a big, nice to look at box, but there are better pieces of equipment.

A more flexible piece of equipment would be the Matrox LE or Max products, they came after the Aja IO and had a longer run.. but they were also built for broadcast signals.

You will need some rarer equipment to clean up and prepare the signal for capture to solve your problems with the Aja IO or any other Mac gear you use to perform the video capture.

A VCR with a line TBC is a great first step. But I don't think any SONY equipment ever had a TBC.. they were very late to VHS and refused to adopt common sense improvements to their VHS players until the last two models ever made. They hated VHS and desperately tried to convert their VHS customers to Betamax or other products. SONY VCRs do not age well. Their power supplies tend to die with age and they are irreparable.

After obtaining a JVC or Panasonic VCR with line TBC.

The next bit of equipment you need is a frame TBC. These are expensive and hard to find, a person needs a lot of patience to get one. They always started at 500 usd and quickly get bid much higher online. The first three months of the year are usually the worst time to find one, because they are being heavily used. It gets easier to find one after April. They get bought quickly by impulse buyers with experience because they are somewhat sought after. "Feeding frenzy online" comes to mind.

Then you should look for.. but not hold out hope of finding a good analog proc-amp, like from Signvideo Studio 1 Productions.. they start at 300 usd and sell even faster than TBCs. With TBC people have some pause and confusion.. so they sell a fraction of a second slower. Proc-amps are well known and understood.. brands matter a lot.. and they are slightly cheaper than TBCs.. so people snap them up.

A few people will get a Waveform monitor and a Vectorscope to (look) at the signal levels for IRE and Color axis to make sure they are (a) within the capture window of their video capture device and (b) have the correct "Hue" or relative color balance and voltages for their video capture device. They tend to be clunky and heavy and CRT based.. the LCD versions are astronomical in price and no longer made.

The (best) VHS capture device I have seen is the (last) Grass Valley USB capture device; ADVCmini it is extremely rare. Despite having the name ADVC, it is USB based, and it is PAL and NTSC, and can capture as windows compressed, mac compressed or uncompressed video files. Giving you all the options you need.. but its a dream to find one. I found a couple but its unlikely you will ever see these for sale or resale ever again. Its device driver and capture application is completely custom and unique to the product. Some day I may post a youtube review of one.

If you clean up and prepare the signals, video capture even with the Aja IO will improve and you may not need a different video capture product.. but I'd recommend looking around.

All that said you would do best:

1. Finding a better VCR, even with many SONY vcrs on-hand .. you should see how much better it could be.

2. Finding an external frame TBC, like a Datavideo TBC1000 or Cypress 8714 "greenbox" (I think thats the model)

3. Trying other video capture options for the Mac (or Windows.. there are many more options for Windows even today than for Mac)

4. Staying far far away from "modern" HD or SDI video capture products that "say" (works with s-video or composite) they don't work well.. they were made for Progressive video and disable any TBC features to make HD capture better.. and cannot re-enable the SD features that made them work better for SD

The Sony SVO-5800P does does feature a TBC and various other things.

Now, that's a plan of action! Now i'll study it.

In fact, the SVO-5800P has the TBC conncector (TBC Remote, with still the plastic protection on, on mine! :) ), but I didn't understood to what link it cause the Datavideo TBC-1000 don't have any connector of this kind (I'm maybe out of range, but the only reference I have with synchronisation is wordclock...).

I've found a swiss shop that selling parts and gears (including Datavideo). I'll call them to see what they have in stock.


Here are pic of the two machines

Oh wow those two pieces of gear have matching connectors, XLR to XLR and S-Video to S-Video no adapters needed, Or maybe use the component output Y/Pb/Pr to take advantage of the built in CPI and XY filter circuits, I also like the fact that it has a stereo audio linear track:

http://www.digitalfaq.com/forum/atta...1&d=1580157766

Yes, I tried to use the Y/Pb/Pr output, but no way to parameter AJA VTX Xchange (the capture soft) to have the video (sound ok). So I used the s-video. This morning (now, here it's 07:30), I'm trying to digitalise a conference from the 80's to see if it's working.

So:
Sound: 1-2 XLR-> 1-2 XLR
Video: S-video out -> S-video In
Time Code: Time code Out -> Loop Ref
Soft parameter: 10 bits S-video 25Hz (should I choose 8 bits?)

I have even took out the Mac G4 (OS 10.3.9), but I've only found the Final Cut Express in my archive (who's only compatible from 10.4).
I would need Final Cut Pro 4 (if any one can help me to find this version, thanks in advance!), the last version compatible with the AJA IO. (I even called Apple to ask if they would sell it to me, but, surprise, no...!)

EDIT:
After 45 minutes, 580 drop frame and a 70 Gbits file. I had to stop it because the soft freezed. I realised that the compression parameter was uncompressed. So I choosed H.264 and I started again. To continue...

Use 10bit if you are planning on doing restoration such as color correction ... otherwise 8bit should be fine.
Is this the final cut pro you're looking for?

Ok! So 10bits it is, for now.

For the Soft, I've exactly the same package, but the DVD's are missing. I'll dig a little bit more to now if they are still in the school, somewhere... (In fact, this is the B plan, for now, the A plan, using the AJA capture soft, don't seems too bad)!
But thanks anyway!
After 15 minutes, only 2 frames dropped.
EDIT: After 52 minutes, 22'500 frames dropped, file 1 Gb but the computer went to save screen a the capture video freeze for 5-10 minutes... . Let's try again without the screensaver.

After 2 hours (complete), 300 frames dropped, 6Gb file. The video as a delay of 3-4 second with the sound(... :'( )

and here is a lot of blue green horizontal line...

With 2000 tapes, you need to not mess with iffy/budget options, and go directly for an optimum conversion setup. Don't even attempt to get a "good" or "decent" workflow, but literally the best gear. With good/decent, you'll have issues. With budget, more issues. Now multiply that by 2000. You'll literally waste weeks to months simply fighting the tapes, instead of capturing them.

Until you have all the TBCs in place, you're just wasting time, spinning wheels.

I don't understand, my Sony SVO5800P is sending TBC to the AJA...!

Time code out (VCR) -> Ref in (AJA) with a BNC ...!

Compressing H.264 is quite processor-heavy (unless the system has a dedicated hardware H.264 encoder). I would think Apple ProRes may be the best compromise here (if supported at least, don't know when it was added). It's smaller than uncompressed, but lighter than a delivery format like H264. Uncompressed may be a bit too bandwidth-heavy to write to disk fast enough, while with h.264 the processor may not be able to keep up with encoding fast enough (and it's not ideal for editing either.) On windows or linux I would use a lossless video codec like huffyuv or utvideo, but that won't be available here.

TBC = Timebase correction, it's not the same as time code. The reference in on the AJA is for something else than time code, maybe a genlock input to syncronise two or more video sources.

TBC refers to the signal processing that is done to stabilize the analog video signal coming off the tape. As mentioned, there is a firmware update that improves the TBC functionality in the capture devices. How well it works I don't know, maybe jwillis can answer that, he's written about this device earlier. I don't know whether the TBC in the Sony VCR fully buffers frames and output a stable video stream, or if it just acts to correct horizontal wigglyness (like the TBC in consumer VCRs).

The component output requires an add-on board so it may not be available

Time code isn't TBC. Nothing is being sent, that's not how TBC (timebase correction) works.

Correct. And that's why both formats will drop massive frames.

Reading around, it does near-nothing akin to ADVC-300 boxes.

It's always been my understanding that it's just line TBC -- which is nice, of course, but still not frame-level buffering, which is also required. Again, for TBC, for VHS sources, you need both internal line and external frame. This isn't an option, it's a basic requirement to perform the process.

I can appreciate wanting to curtail costs, and find alternatives, but it will not be possible for this project. Period, the end. Even attempting, with 2000 tapes, will result in a lousy video experience that will end in giving up. I see this too often. (And because the person doesn't want to spend funds, doesn't even outsource, just quits, the videos never get converted.)

The only "alternative" is just do terrible quality work, have dropped frames, bad image, etc. Many errors will outright prevent capturing. Many captures will literally be unwatchable. And to me, that's not at all an alternative.

If you want to DIY video, great! But specific tools are required.

Yeah, h.264 and 10bit requires some horse power, It's a good thinking to capture compressed for that many tapes but if performance is an issue try maybe 8bit and see how it goes. If the frames drop is not related to performance issues you may need to get an external TBC it's different from the one in your VCR.

Ok, thank you everybody. I'll try a little bit more, because that's how we learn but...I've contacted the National Television Archive, and if there is something valuable to digitalised, it will be done by them

It's note that I don't want to invest, but as a State service, I'm not allowed to buy outside the contractors (they obviously don't sell gears for this project), and my budget is not dedicated to that project at all. And, I was trying to do it with the materiel that probably recorded most of the VHS.

Thanks for the help and info
About the 2000 tapes, I think only a few hundred will finally be copied.

-- merged --

I hope it won't just make the bosses angry:

Finally, I'm using a "Miglia technology - Director’s cut take"(S-video/cinch/Firewire) and Finaly Cut X - "Creat an Archive" Import system.

The quality is enough and the weight is correct. No frames dropped problem neither synch.

Something that must be remembered is that commercial releases, with high quality distribution, won't be anywhere as bad as home-recorded and home-shot amateur consumer tapes. Those tapes are not recorded, but rather made by contact duplication. The main issue with commercial releases is cheap tape stock, anti-copy, and overuse/abuse. Then all the issues inherent to all VHS tapes, namely chroma and line timing, requiring a line TBC.

So, in terms of tests, it's very partial at best. In my TBC research, playability of retail tapes is only step 1 of about 10. (Literally, the 1st thing I do is see how a TBC chokes on Macrovision tapes. Those that choke tend to fail about half of all subsequent tests.)