premiere + xvid = nuclear disaster

[BlindSword]

im not pretending i was asking a question, im trying to learn about something as I got along, thats not a problem is it? Im trying to figure it out based on your posts. I was ocnfused because your post went from 1 gb to 15 and I thought the codecs also compress the video files.

[Kariudo]

codecs do compress files.

but dvd footage is already compressed (mpeg2 format, lossy)

if you took one episode from a dvd and saved it as uncompressed video...it would be really big (I haven't done this...nor do I plan to...if I had to guess I'd say somewhere around 50GB).

so huffy and lagarith do a really good job of compressing...but the files still seem large. The solution is to only encode the clips that you want to use instead of the whole episode. (in my current project I have ~260 clips at ~1-2 seconds each, encoded with huffy they take up about 5 GB)

[DJ_Izumi]

Let's explain: DivX, XviD, WMV, Mpeg-2, all that bullshit are formats designed for VIDEO PLAYBACK. They are ment to have you view the video from the begining and just play it, they are entirely optomized for this exclusive task. You should notice something when you drag along the time line in the video player, the video will lag or pause for a moment after you stop dragging before it plays back. That is because the video has to go to the nearest I frame and then build the requested frame from that. However, some players are smart and will ONLY drag to the I frames and play from them and then there is no lag. Thing is, I frames are sorta 'rare' in a video file. At 24fps there will be an I frame maybe every 1-10 seconds, depending on the settings during encoding.

These formats are GREAT for getting good quality playback. Once you load them into a video editor, where you are dragging backwards and forwards a hundred times and seeking out very specific frames from every cut, you are doing something this shit is NOT designed to do.

And yes, if you trans-code your file to XviD or DivX or what ever and make EVERY frame an I frame it should actually work pretty damn well. Thing is, in encoding from a lossy codec to another lossy codec you lose quality, especially when you use these kinds if distrobution codecs. Think of it like making a photocopy of a photocopy like your teachers did in high school, with every generation it gets shittier and shittierlooking.

This is why you use uncompressed or lossless codecs, every frame is an I frame, they are fast to decompress (Most of them are at least) and they do not use any quality.

Uncompressed video is the default, it uses a what is basicly a bitmap for every frame. But this is HUGE.

Example, I took a 10second clip from a VOB and saved it as an uncompressed AVI. It is 270mb. This is actually unreasonably large even if it is a good method. It's just too big for most people to make use of without buying extra hard drive.

Using HuffYUV, a fast editing and capture codec, which is lossless, the same 10sec clip is only 97mb. Almost 1/3rd the size of the uncompressed video. And there is ZERO loss of quality. Every frame will be decoded to be mathmaticly identical to the original video. It's just like zipping a bitmap file.

Lagarith, while slower, (My Athlon 2500 XP+ encoded this test at 15fps or so) made the same file only 60mb in size. So while there is no loss in quality it is a hell of a lot more efficent.

[BlindSword]

Thanks for explaining, im actually converting a few files right now through huffyuv and ill be sure to use those from now on.