Another test run with ATI capture

[Lem Motlow]

I decided to run another test using the ATI MMC for capture. I'm trying to make sense out of all this and I know that higher bit rates should make the result better. I have also read several threads discussing the merits of capturing in 352x480 vs. 720x480.

I did 6 captures of the same 5 minutes of 8mm video with different resolutions and bit rate settings. The constant settings were as follows: MP2-DVD, encode interlaced, no 3:2 pulldown, record cropped, 48k stereo, 4-2 closed GOP, motion estimation quality 100%, Audio 256k and no video soap.

Test 1: Res: 352x480 Bite Rates: Max 4M Target 3.5 File size: 145mb
Test 2: Res: 352x480 Bite Rates: Max 6M Target 5.5 File size: 239mb
Test 3: Res: 352x480 Bite Rates: Max 8M Target 7.5 File size: 317mb
Test 4: Res: 720x480 Bite Rates: Max 4M Target 3.5 File size: 143mb
Test 5: Res: 720x480 Bite Rates: Max 6M Target 5.5 File size: 237mb
Test 6: Res: 720x480 Bite Rates: Max 8M Target 7.5 File size: 333mb

Referring to the bell-curve graphs found at http://www.digitalfaq.com/capture/avivsmpeg.htm I really was not expecting to see an increase in quality on the 352x480 captures by raising the bit-rate. I did however, expect to see an increase in quality of the 720x480 captures as the graphs indicate. I authored and burned all of these as different chapters using Nero and viewed them on the TV.

Results
For the most part, all of these captures look the same.

1) I could see no difference in the 352x480 tests 1, 2, and 3. Wasn't really expecting any though.

2) I could maybe tell a very small gain in quality on the 720x480 tests 4, 5 and 6 as I raised the bit-rate, but it was very small. According to the bell-curves, if I am reading them right, there should be a large increase in quality up to a bit-rate of 8M.

3) I couldn't really see any difference when comparing the 352 and the 720 captures.

I would really appreciate any insight, opinions and tips regarding all of this.

Questions
Am I on the right track here? Really thought I would see some big differences in quality on the 720x480 captures with the different bit-rate settings.

Is it possible Nero is doing something to the files? (I think re-encoding is the proper word), but I don't believe this to be the case.

I know this is very subjective, but I still am wondering what folks around here consider to be a good quality capture from analog material. I would estimate that I have about 90% of the source quality-wise.

[lordsmurf]

- VBR will alter this slightly
- active capture is not a good test for this
- ATI MMC lies about bitrate sometimes
- your source can determine some of this

What is Nero doing? Why is this part of your equation?

[Lem Motlow]

- VBR will alter this slightly
- active capture is not a good test for this
- ATI MMC lies about bitrate sometimes
- your source can determine some of this

What is Nero doing? Why is this part of your equation?

I don't know what you mean by "active" capture.

Bitrates are close to capture settings- I just viewed the DVD in PowerDVD and the info shows they are.

So you are saying that with a source like 8mm, raising bit rate to an indicated max of about 8 may not help much on a 720x480 res capture, but in theory it would.

Nero- just wondered if Nero re-encodes the MPEG files before it burns. From reading here, I know that some programs re-encode files, but I think this applies mainly to editors?