mirror of
https://github.com/mpv-player/mpv
synced 2025-02-11 09:27:26 +00:00
libmpeg2 codec ctrl removed
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@405 b3059339-0415-0410-9bf9-f77b7e298cf2
This commit is contained in:
parent
76e5256225
commit
aa9517f18a
@ -106,40 +106,6 @@ Now, go on:
|
||||
And they call the appropriate demuxer, to get the compressed data.
|
||||
(see 2.)
|
||||
|
||||
5.a Codec controller: this is the greatest hack in the whole :)
|
||||
The libmpeg2 is so unstable, that I can't believe it.
|
||||
Of course I don't mean it's bullshit :) rather it only accepts
|
||||
totally perfect, error-free streams. If it founds error, it
|
||||
just segfaults ;) And don't start laughing, this is great this way,
|
||||
from the view of speed it would be 50-100% slower if stuffed full with
|
||||
verifications. That's why I solved it by running it in a separate
|
||||
process, and if it dies, who cares, just start another.
|
||||
However, a few things are needed for this:
|
||||
- codec controller process: a separate process, which sleeps, but if
|
||||
its child (the libmpeg2 process) dies, it quickly starts another.
|
||||
So the MPlayer doesn't have to care about this, it just pumps the
|
||||
compressed stuff into the child, which displays it.
|
||||
- shmem: the compressed data, and the uncompressed frames are both
|
||||
in shared memory, so all 3 processes (mplayer, codeccontrol,
|
||||
libmpeg2 codec) sees 'em, so they can trade data fast.
|
||||
- FIFO is used for the communication between them.
|
||||
- If the child dies while decoding, the succesfully decoded data
|
||||
isn't lost, it's inherited by the new child through the
|
||||
shared mem! So only a little error can be seen in the video,
|
||||
it won't disappear or turn green, as in the older versions.
|
||||
|
||||
The disadvantage of this all is that since the libvo and libmpeg2
|
||||
are closely related, the libvo needs to run in the same process as
|
||||
the libmpeg2, in the one that keeps dying/reborning, and not in the
|
||||
one that has the controlling process, the MPlayer. This causes a
|
||||
lot of problems, mostly at the handling of events in the libvo window
|
||||
(keypresses, etc). So there are miscellaneous workarounds, a lot of
|
||||
FIFO, and trick which exploits that X doesn't care which process
|
||||
queries its events.
|
||||
|
||||
I'd like to solve this in the near future, and use the signal/longjmp
|
||||
(this is a hack, too:)) method, developed on the mpeg2dec-devel list.
|
||||
|
||||
5. libvo: this displays the frame. There are 2 different output routines in it:
|
||||
|
||||
5.a draw_slice(): this displays YV12 pictures (3 frames, a full sized which
|
||||
|
@ -106,39 +106,6 @@ na nezzuk tovabb:
|
||||
kell lejatszani (lasd 3. pont elejen)
|
||||
ezek pedig hivjak a megfelelo demuxert hogy megkapjak a tomoritett
|
||||
adatokat (lasd 2. pont).
|
||||
|
||||
4.a codec controller: na ez a legnagypbb gány az egeszben :)
|
||||
a libmpeg2 ugyanis annyira instabil hogy az mar hatareset.
|
||||
persze ezt nem ugy kell erteni hogy szar :) hanem ugy, hogy csak
|
||||
teljesen szabvanyos, hibatlan mpeg streamet eszik meg. ha hibat
|
||||
talal, egyszeruen segfault ;) es nem rohogni, ez nagyon jo igy,
|
||||
teljesitmeny szempontbol 50-100%-al lasabb lenne ha teleraknak
|
||||
ellenorzesekkel. ezert csinaltam azt a megoldast, hogy kulon
|
||||
processzben futtatom, es ha elszall, hat kit izgat, majd inditok
|
||||
egy masikat. ehhez azert kell par dolog:
|
||||
- codec controller process: egy kulon processz, ami sleep-el, de
|
||||
ha a gyereke (a libmpeg2 processz) meghal, akkor indit gyorsan
|
||||
egy masikat. igy az mplayer-nek nem kell ezzel fogallkoznia, o
|
||||
csak pumpalja a gyerekbe a tomoritett adatot az meg rakja kifele.
|
||||
- shmem: a tomoritett adatok, es a kitomoritett framek is shared
|
||||
memoryban vannak, tehat mind a 3 processz (mplayer, codeccontrol,
|
||||
libmpeg2 codec) is latja. igy tudnak gyorsan adatot cserelni.
|
||||
- a koztuk levo kommunikaciora meg FIFO van.
|
||||
- valamint ha dekodolas kozben meghal a gyerek, az altala sikeresen
|
||||
dekodolt adatok nem vesznek el, hanem a shmem-en keresztul oroklodik
|
||||
az uj gyereknek! ezert max egy pici hiba latszik a kepen, nem tunik
|
||||
el az egesz meg zoldul be, mint a regi verzioban.
|
||||
hatranya ennek az egesznek, hogy a libvo-libmpeg2 szoros kapcsolodasa
|
||||
miatt a libvo is abban a processzben kell fusson, amiben a libmpeg2,
|
||||
tehat abban ami allandoan megdoglik-ujraszuletik, es nem abban amiben
|
||||
a vezerlo processz, az mplayer fut. ez eleg sok gondot okozik, foleg
|
||||
a libvo ablakban tortent esemenyek (billentyunyomas pl) kezelesekor.
|
||||
erre mindenfele workaroundok vannak, FIFO-k minden mennyisegben, meg
|
||||
trukk ami kihasznalja hogy az X-nek mind1 melyik processz kerdezi
|
||||
le az Event-jeit.
|
||||
|
||||
szeretnem a kozeljovoben ezt megszuntetni, es az mpeg2dec-devel
|
||||
listan kidolgozott signal/longjmp (szinten gány :)) modszert hasznalni.
|
||||
|
||||
5. libvo: ez vegzi a kep kirakasat. jelenleg 2 kulonbozo kepkirako
|
||||
van benne:
|
||||
@ -152,5 +119,3 @@ na nezzuk tovabb:
|
||||
es csak packed formatumot (YUY2, RGB/BGR) tud.
|
||||
ezt hasznaljak a win32 codecek (divx,indeo stb).
|
||||
|
||||
|
||||
|
||||
|
Loading…
Reference in New Issue
Block a user