Buffering frames

[NuPogodi]

Hi!
I've recently read a guide Ripping subtitles from video files using SubRip (http://zuggy.wz.cz/guides/video.htm) where ai4spam wrote
AviSynth seems to have its own buffering, but only between the previous and next keyframes.
Well, the questions are
1. Whether this statement is correct?
2. If yes, how AviSynth knows about the structure of compressed video (i.e. where is i-, p- or b-frames). Decoder knows it, but it seems to send back already uncompressed video.
3. If not, whether AviSynth allocates frames within a value defined by SetMemoryMax()? Are there any references for a proper documentation?
Thanks.

[Fizick]

NuPogodi,
wait :)

1) no, but buffering (cache) exist.
2) only Decoder knows it.
3) Yes. No, only in this forum.
All IMHO :)
And I never used Subrip and do not know what aviynth source function (decoder) it use (avisource, directshowsource, or Mpeg2source via dgdecoge - they are very different for seeking)

May be these questions are related to avisynth developer forum?
What do you really want to do?

[sh0dan]

As Fizik points out, the statement is incorrect, but if a frame out-of-sequence is needed, the decoder will decode from the nearest keyframe.

The cache saves as many frames as possible within the memory limit.

Moved to Development

[NuPogodi]

Thanks, thanks a lot...
What do you really want to do?
Well, i was just a question, nothing more. I've read a manual (it's a stupid idea, i know), I've found some things which sound strange to me and, therefore, I've asked profi's.

Almost forgotten... SubRip usually generates own AVS-scripts opened by DirectShowSource().