[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[OpenDivX] OpenDivX Encoder -> OpenDivX Forums Digest

OpenDivX Encoder -> OpenDivX Forums Digest

Topic:		How to use the encore.c
Author:		ciagon
Posted:		 2001-02-21 21:07
I am looking over encore.c and have a few questions about hot to use it.
This is my best guess as to how it works:

My app fills in a reference struct somehow with all the info for the

Then, I call encore(...) for each frame I want to compres.

Once again, this is only a guess.  For all I know, this could be
completely wrong.

Also, how do I get the compressed data stream back?


Topic:		Does anybody want this?
Author:		gruel
Posted:		 2001-02-22 16:16
On 2001-02-21 18:07, czw wrote:
I happen to have some free time right now. Instead of spending it on
something constructive, like drinking beer, I decided to come up with
two ideas for speeding up video encoding: 3DNow assembler code and
distributed encoding.

By converting the critical parts of the encode code to AMD:s 3DNow
instruction set, one would see a speed increase. The same could be done
with SSE(2), but I can't test that since I only have an Athlon and a
K6-3 available.

Distributed encoding sounds even funnier: send unencoded data (in blocks
of about eight seconds) to two or more computers, who perform the video
encoding a and returns the finished DivX  video sequence. This way, you
may almost cut the encoding time in half if you have two equally fast
computers in a network.

Is there some work done on any of these projects, and/or do they sound
good to you? The second project would be great in local area networks
with lots of spare processing power.

[ This message was edited by: czw on 2001-02-21 18:07 ]

I like the process of parallelization, so 
project 2 would certainly be my personal favourite. But the discussion
about thread-safety made quite clear: In the near future, 95%(99%?) of
the users will have no access to parallel machines, and surely not to a
cluster of worktations. 
However there will be in fact many people whose processors support 3DNow
and/or SSE. 
Also interior parallelism in processors is a growing topic these days. 
Now since the main points of critcs at the moment is speed, I would say:

Number 2 is more fun. 
Number 1 is more useful. 


Topic:		Does anybody want this?
Author:		elbows
Posted:		 2001-02-22 18:09
Yep number one is definately more useful to the average user of DivX,
but for hardcore encoders & pro's number 2 would be great.

I'm no programmer but I've been keeping up(ish) with the way the project
is going.  I think they are concentrating on improving features and
fixing bugs at this stage, rather than doing cpu-specific optimisations.
Having a 1Ghz AMD TBird I would love to see some optimisations though
(hint, hint)

Topic:		Does anybody want this?
Author:		czw
Posted:		 2001-02-22 19:24
I have a gigahertz Athlon myself, but I am connected to a network with
at least 200 students. There are several computers in the network that
don't do much (Linux file servers), and that would gladly do some DivX
encoding in their spare time. On the other hand, quite a few of the
computers around here are K6:s, Durons and Athlons.

Since the OpenDivX project doesn't  seem close to the optimize stage
just yet, the distributed encoding might be a better long-term project.
When the source code is mature enough to do some serious optimizations,
I will sit down and do just that.

Anyone with some knowledge about the current code status and general
attitude towards assembler code, who would like to comment on this?

OpenDivX mailing list
[email protected]

Reply To Poster

Local References / HOW-TO / FAQs