MPEG2 vs AVI capture quality

[EAO]

While I was thrilled with just being to capture TV in synch, my wife thought the captures looked a little fuzzy on playback.

Not too bad - very watchable - but you could tell it wasn't TV quality. I had been using WinDVR at the highest NTSC DVD settings that it allows, so for archiving this would not be acceptable.

I finally loaded Virtual VCR and since then Virtual Dub. I'm using PicVideo MJPEG at 19, and my resolution size is 720 x 480. (I'm having to use YUY2 - cause when I use RGB 24 it drops frames.) I convert it using TMPGenc plus to MPEG2.

The differences between the two techniques are amazing. The picture using Virtual VCR or V Dub is so much clearer! But the amount of HD you use up is huge.

What I'd like to know is what makes .avi so much better/clearer than MPEG2?

Which gets me wondering. Those TIVO devices or DVD recorders with the built in HD. That HD is not really big. What format are they coding into to fit all that high quality clear data onto it?

[Craig Tucker]

What I'd like to know is what makes .avi so much better/clearer than MPEG2?

You answer your own question here

But the amount of HD you use up is huge.

You can get very good quality mpeg2, after all DVD is mpeg2. But you need a good source and it is better to capture to high quality AVI then convert to mpeg2 after. Rather than capturing to mpeg2 in real time. (at least with current technology)

[EAO]

I realise that the data captured is bigger with .avi, but if what you are saying is true, wouldn't the quality drop from source to avi to MPEG2 be just as dramatic as source to MPEG2?

For some reason MPEG2 codes better from .avi than from the original source. That seems strange.

[clanger]

I might be wrong here, but my thinking is when you encode on the fly using a PC, all the encoding is done in software (unless the capture card has a dedicated encoder chip).
Encoding to AVI results in huge files, but not much work is being done to actually compress the data, you dont have the time otherwise you start losing frames, and it's a lossless compression.
when encoding to Mpeg2 on the fly a similar thing happens, but you get smaller files with less quality, again due to the amount of work your asking the processor to do, and as it's a lossy compression the codec will just chuck away (or ignore) a lot of the data. Hence the drop in quality.
When you encode from AVI to Mpeg2, I'll bet it takes at least twice as long to encode (ie nowhere near real time !), but the quality is a lot better, because you can choose to have a better loss / quality ratio.

Basically if you want a good Mpeg2 image it'll take a fair amount of time and the file will be reasonably large, a small file wont take long to make, but the quality suffers.

It's got something to do with the way the codecs work and how the different file types work, ie lossless vs lossy.

I'm sure that someone here will be able to explain it in greater detail !

[EAO]

So since my capture card is not able to process MPEG2 by hardware, it's making compromises.

I'm using a Winfast 2000XP dlx. Lastest drivers, but my PC did not like any of the stock programs.

Would I be better off - quality wise - to have a capture card that can process MPEG2? If so is the quality good enough on a TV capture to look like TV on playback or will I still be able to tell it's a recording with my poor old eyes?

[n8wheel]

I use a Adaptec Videoh PCI card to capture MPEG2 in realtime. The current Adaptec product is the Media Center; you might want to check the Capture section of this website for some more possibilities. Apparently there are now a lot of happy ATI All-in-Wonder users out there now.

I am very impressed with the quality of the captures with the Videoh PCI. You should not be able to tell that it was a recording.