ATI AIW capture filesize stumper.

[Progenator]

Hi,

I am capturing with an ATI AIW and am wondering why I get such discrepencies in my guestimation of filesizes. For instance, I recently used the same preset (LS movies,VBR) to record two different movies from the same source (directTV). The first was Underwo### and had a recorded time of 2hrs and came out at 2.3gb. The second was Amer$%can W@Tding with a time of 1hr45mins, but a filesize of 2.6gb, larger than the previous movie, which was ofcourse longer.

Both were recorded while I was away in a 8hr period, AMer@(can W%^&ding being the last. Using ATIShed. For reasons of convenience I leave the ATI TV running when I leave.
There are the same number of processes running for both as well (minimal).

My first guess would be that American involved more complexity and thus commanded a higher bit allocation, but I can't see how agianst Under@#$%^, which is high action and mostly dark (dark photos use more developer is where this reasoning comes from).

I have also noticed that when I do sample captures of small time limits, they often climb in size at the same preset as I progress, could it be less resources for compression??

Please correct me if I am wrong, or if I am right, it is a bit of a peeve for me to think I am doing something badly that causes less efficient compression. I thought for sure I could fit both those movies on one DVD, with my desired preset.

ATI MMC 8.8, WINXP pro

[Almost Human]

My guess is that letterboxing is the difference.

I record TCM from digital cable box via S-Video into my AIW 7500. At 2.73mb bitrate VBR: 2hr movie full frame display = 2567mb, 2hr movie letterbox = 2132mb. I capture MPEG2 DVD with video cropped.

[Progenator]

Neither of mine are letterbox, or were broadcast as such (unless they were cropped).

But good to know.

[Progenator]

Also should tell you I do the same thing, pretty much; svideo. I record movies at 3.9mb VBR and cartoons at 3.65. 350X480, 256k audio, no filters, DVDmpeg-2.

[Alan69]

Dark photos have nothing at all to do with information content. They only use more because the developer just happens to develop the dark areas. If the developer made the light areas then light photos would take more. Has nothing at all to do with how much detail and information is in the picture. And a picture with one side black and one side white takes far less info to describe than one with lots of small randomly places black dots, even if they're both exactly 50% black and take exactly the same developer.

Dark makes for large patches with lower detail, much easier to encode than full detail and complex backgrounds. A better lit subject shows much more detail than a dark one.

You can usually tell which movies will take more or less than usual just by looking at the backgrounds. The busier the background the more information in the frame, and the more likely camera motions etc will have to encode new info. Even motion info takes a back seat to the general detail level of the whole film. MPEG2 does understand background shifts etc to lower the required data, but it'll still take more on average.

[Progenator]

Thanks very much for the explanation. Nice to know whats happening.
So thats why 2d animations come out small too, cool.

[johns0]

Progenator,you can always author the 2 movies and then use dvdshrink to make them fit on one dvd.

[lordsmurf]

Progenator,you can always author the 2 movies and then use dvdshrink to make them fit on one dvd.

That's what I would do. Use DVD Shrink 3.1 and deep analysis.

[Progenator]

Sheesh, i did that before with two DVD's.
4+ brains are better than one.