Shrinking Encoded DVD with DVD Shrink

[punk_rocker128]

ive always wondered if this way of encoding would work. because DVD shrink can shrink a DVD9 disc to fit onto a DVD5 disc, would it work to encode your avi's with the highest bitrate possible? if the output files were 8 or 9 GB in size, could DVD shrink be able to shrink these files down to 4.7 GB? i just kept wondering why i was worrying about my encoded movies fitting on a DVD5, when i could use DVD shrink to shrink it in size.

have people been doing this and im just now figuring out about this method? or is this a bad way to encode?

[guns1inger]

Transcoding is not the best way to reduce file size. It's major advantage is speed. However, if you are starting with a avi source file, you will get much better quality by encoding to the correct size in the first place, than encoding to a larger size and then using DVD shrink.

[punk_rocker128]

well the reason i am asking is because i have a 1.6 GB avi file (4 hours 20 minutes) that when encoded to the max size that can fit on a disk, it gets pixelated pretty bad. would could i do for this? i was thinking of splitting it in half and using 2 discs, so i could use higher bitrates.

[guns1inger]

4 hours 20 minutes at full D1 will look pretty bad. You could try resizing to half-D1 instead, but splitting it to two disks would be the better option. Any transcoder would produce low quality on a fie of that length.

Your other choice is to author for a dual layer disc instead.

[punk_rocker128]

ok, now im back to my original question. lets say i split it to 2 discs so each disc had around 2 hours of video on it. if i encoded using the max bitrate then used DVD shrink on it, would that be that bad? what would be the disadvantages of doing this? it couldnt be THAT bad, could it?

i dont really want to use DVD9. i have an old burner and the discs are pretty expensive.

[fritzi93]

Sorry, that's a bad idea. Best to get the bitrate right the first time.

Even if you were re-encoding, there'd be a slight quality hit versus getting it right the first time. But you're not, you're transcoding.

For starters, you have motion vector data and DCT coefficient data. DCT data is pixel residual data. "I" frames are all residual or DCT data. DCT data is also used for error correction in the other two frame types, "P" and "B". Each movie is different, but in general "action" movies will have relatively more motion vector data and less DCT data than do "talking head" movies. And transcoders work by discarding DCT data *ONLY*.

Furthermore, that reduction will be mostly in "B" frames, since errors are not thereby accumulated. At low compression, "I" and "P" frames are essentially untouched. When the compression is high enough, it will have to take away from other frames as well, which means there WILL be obvious artifacts.

What you propose is 50% compression, or thereabouts. Say, for the sake of argument, that a particular movie has an average of 50% DCT data, you're removing all error correction essentially. Surely you see that's not gunna be a good idea?

I won't even comment on AVI conversions.

[EDIT] Geez, you guys are quick! Anyway, my response refers to the original post only, without the later clarifications.

[guns1inger]

my point (and fritzi93's more detailed explanation) still stand. It is always - no exceptions better to encode at the correct bitrate in the first place, than to encode too big and reduce later with a transcoder.


Actually, I can think of one exception - if you have really bad cataracts and can't see a damn thing. Then it doesn't really matter what you do to the picture.

You obviously care to some dgree about the quality, or you wouldn't consider splitting it across two discs. If you care that much, then do the job properly. Get a bitrate calculator, crunch the numbers (will take all of 30 seconds), and encode to size.