Can I get uncompressed video via ADVC?

[ViMan]

I'm looking to purchase ADVC-55 or ADVC-110 but still not sure if it is what I need.

As far as I see it provides DV format output.
Does ADVC-xxx allow me to get uncompressed (looseless, about 1 GB per minute) video?
I want to capture my VHS video and process it by filters that is not so good after DV codec compression.

[THoff]

No, the ADVC always outputs DV. You could have your NLE save that in uncompressed format instead of DV, but you would not gain any quality, you would merely increase the storage requirements.

[GrassValley_BH]

To elaborate a little more...

Via Firewire: No, it's always DV via FireWire (or MPEG-2 Transport stream for the HD series ADVCs)

Via SDI: Yes, in the ADVC3000 (currently this is the only unit that does SDI-DV-Analog cross-conversion)


For your intended purpose, I would go for a good analog-capture board like EDIUS NX.

[ViMan]

It is strange that I need to pay more to remove DV codec :)
Moreover I think external device in metal box has less noise then internal card collecting all fields inside PC case.
Is something else better in EDITUS then in ADVC? I don't mean some functionality, just picture quality.

I found another device - Dazzle Video Creator:

http://www.pinnaclesys.com/PublicSite/uk/Products/Consumer+Products/Dazzle/Dazzle+Video+Archiving/Dazzle+Video+Creator.htm

It uses:
Philips 9-bit video input processor SAA7113H chip (http://www.semiconductors.philips.co...AA7113H_2.html)

I have no info about Canopus.

Could you please help me to understand if Pinnacle device is worse in quality and why (vs. ADVC and EDITUS)?

Thanks a lot!

[kbosward]

Could you please help me to understand if Pinnacle device is worse in quality and why (vs. ADVC and EDITUS)?

That's a loaded question to ask on a Canopus forum :) Why don't you do a search for "Dazzle" in these forums to find out why people changed to Canopus?

Ken.

[GrassValley_BH]

The Dazzle is a hardware MPEG encoder. Good if you just want to "dump" stuff to DVD. Bad if you want to edit or filter the video.

The Philips chip is used in many devices - it's a tuner. It goes as far as making the raw video available to something else, whether that's a computer interface or a hardware encoder.

Remember that standard definition video is 720x480x29.97 fps or 720x576x25 fps, translating into about 10,368,000 pixels/second (NTSC is just slightly less at 10,357,632). Now using 4:4:4 sampling, that's now 248,832,000 bits/sec or 31,104,000 bytes/sec. That's 29.7 MB/sec for just the video, not including the audio!

USB 2.0 supposedly can support up to 480 Mbps, but realistically with other USB devices and CPU utilization, that 248,832,000 bits/sec eats up more than half of USB 2.0's bandwidth. That's a lot. Not to say that it can't be done, but it'll be tough to do reliably.

Your run-of-the-mill TV capture card will feed uncompressed data to your CPU, unless it's one of the ones with hardware MPEG encoders like the Hauppauge PVR350. The problem is that the CPU will then need to capture all of that data. A good RAID can sustain 29.7 MB/sec of writes, but you'll be quick to run out of space, as that's 1782 MB/min - almost 2 GB a minute of data.

So then you come to compression. Now you're using the CPU not only to receiving the incoming uncompressed video data, but also to compress that data, then send it to disk. If the CPU takes too long to compress a frame, then it'll "miss" the incoming data for the next frame, and a frame drop occurs. This means you now have a "hole" in your captured data, audio and video go out of sync, and overall it's a bad day.

Now DV, on the other hand, is only 25 Mbps and it's already compressed. Capturing DV is just a matter of taking that 25 Mbps of data and sticking it on the disk. Simple.

The crux of my point here is that a lot of people want to capture uncompressed, but few people need to. I've been through the route of uncompressed TV capture using an ATI card and such, and as long as you use a decent compression method, capturing compressed makes the whole process more reliable.

It's pretty annoying to have to recapture 2 hours of footage because you sneezed and moved the mouse 1.5 hours into the capture and dropped 20 frames.