MainConcept MPEG-2 plug-in

fongaboo wrote on 2/11/2005, 9:07 AM
Is there a way with the MainConcept MPEG-2 plug-in to reduce the overall bitrate without increasing inter-frame compression? That is, to allow increase compression of each frame but keep the same compression rate between frames? I want to make really compressed MPEG-2 files, but want them to have a pure 29.97fps cadence when viewing on an NTSC monitor.

Comments

Chienworks wrote on 2/11/2005, 9:25 AM
Doesn't lowering the bitrate do exactly this? Bit rate and frame rate are relatively indepedent of each other.
fongaboo wrote on 2/11/2005, 1:43 PM
not exactly.. when you talk about a hold-and-modify type compression like MPEG-2, the actual format framerate can be 29.97.., But MPEG will make I-frames at regular intervals, which are periodic mile-markers of sorts, that it uses to compare subsequent frames to in order to create those subsequent frames. If the inter-frame compression is high, then the frames tend to blend or blur into each other.. creating an illusion of a somewhat slower or more jittery framerate.

If the intra-frame compression is high, but there is not much inter-frame compression, then you will have lots of blocky artifacts in each frame, but the framerate will still seem completely fluid. This is what I am looking to do. Whether the inter-frame and intra-frame compression can be adjustted independently of each other in the MainConcept plug-in specifically is an open question..
Chienworks wrote on 2/11/2005, 1:48 PM
The MainConcept encoder has a setting for I-frames. Presumably you could set this to 1 and get all I-frames with no interframe compression at all.
B_JM wrote on 2/11/2005, 1:51 PM
not a very good idea :(

fongaboo wrote on 2/11/2005, 6:12 PM
Care to elaborate? I was looking for somebody who had more experience with this than I..
Chienworks wrote on 2/12/2005, 2:54 AM
True, i didn't think it would be a good idea. Then again, asking for smooth motion at reduced bit rates is a pretty dodgy idea to begin with.

If you have all I frames, then each frame must be complete, and the few bits it has available must encode the entire image rather than just the few changes from the previous frame. This means that it must be much more compressed than B or P frames would be, and you'll have correspondingly greater compression artifacts.

All I frames is a good idea when using very high bitrates. It is a bad idea at low bit rates. But then, that's what fongaboo was asking for.