home *** CD-ROM | disk | FTP | other *** search
- Path: Hermes.grace.irl.cri.nz!maths!peterm
- From: peterm@maths.grace.cri.nz (Peter McGavin)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: Is a real good mpeg player possible on the amiga ?
- Date: 31 Jan 1996 22:22:50 GMT
- Organization: Industrial Research Ltd
- Message-ID: <PETERM.96Feb1112250@tui.maths.irl.cri.nz>
- References: <4e2gqu$5ca@caers3.unicaen.fr> <4e2o57$5a2@serpens.rhein.de>
- <574.6598T297T795@login.eunet.no> <4e7n75$krh@serpens.rhein.de>
- <386.6603T1404T1784@login.eunet.no>
- NNTP-Posting-Host: tui.grace.cri.nz
- In-reply-to: patrick.hanevold@login.eunet.no's message of 31 Jan 1996 18:48:35
- GMT
-
- patrick.hanevold@login.eunet.no (Patrick Hanevold) writes:
- >>>There is FLIC players around that is just as slow as MP.
- >>Does this say anything ?
- >Yes. They happen to be coded very slow. Just like MP.
- >Maby some asm would help a little?
-
- I tried compiling the Unix mpeg_play v1.2 first with optimised SAS/C,
- then later with optimised GNU C, with appropriate changes for CGraphX
- (no c2p). In both cases the result ran about twice as slow as mp103.
- Rewriting the critical parts in asm would probably achieve about the
- same speed as mp103. To get significantly more speed, we need a
- faster algorithm (if one exists), I think.
-
- I haven't followed all the details, but I can see that mpeg_play
- calculates YUV for each frame in one buffer, then converts it to RGB
- in another buffer, then finally renders it to the display device (in
- my case with cybergraphics.library WritePixelArray()). It is
- conceivable that combining these operations could save a lot of memory
- reads/writes. Unfortunately I can't see how to do it because the YUV
- and RGB frames are calculated in non-sequential order and saved in
- buffers until they are needed.
- --
- Peter McGavin. (p.mcgavin@irl.cri.nz)
-