[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
OpenDivX digest, Vol 1 #25 - 8 msgs
Send OpenDivX mailing list submissions to
[email protected]
To subscribe or unsubscribe via the World Wide Web, visit
http://lists.projectmayo.com/mailman/listinfo/opendivx
or, via email, send a message with subject or body 'help' to
[email protected]
You can reach the person managing the list at
[email protected]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of OpenDivX digest..."
Today's Topics:
1. OpenDivX Encoder -> OpenDivX Daily Forums Digest (Darrius "Junto" Thompson)
2. OpenDivX Decoder ->OpenDivX Daily Forums Digest (Darrius "Junto" Thompson)
3. Re: using mpeg2 decoding hardware for divx decoding (Christian Buchner)
4. DirectShow Filter Bugs (Andrea Graziani)
5. General Development Discussion - All Projects -> OpenDivX Daily Forum Digests (Darrius "Junto" Thompson)
6. DirectShow Filter Bugs (Andrea Graziani)
7. OpenDivX Decoder ->OpenDivX Daily Forum Digest (Darrius "Junto" Thompson)
8. OpenDivX Encoder -> OpenDivX Daily Forum Digests (Darrius "Junto" Thompson)
--__--__--
Message: 1
Date: Fri, 23 Feb 2001 18:13:21 -0800
From: "Darrius \"Junto\" Thompson" <[email protected]>
To: <[email protected]>
Subject: [OpenDivX] OpenDivX Encoder -> OpenDivX Daily Forums Digest
OpenDivX Encoder -> OpenDivX Daily Forums Digest
Topic: Does anybody want this?
Author: gruel
Posted: 2001-02-23 14:23
------------------------------------------------------------------------
--------
Quote:
On 2001-02-22 19:24, czw wrote:
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.
Actually both projects do not interfere with each other at all. All
optimization (AMD/Intel/RISC) will be part of the encode-library. To
create a distributed version would only make sense outside the encoder
itself, say in the application that calls it.
I guess it could be less than a days work to turn an existing mpeg2avi
or similar into a version using MPI or PVM (I like MPI more), running on
a cluster of Linux/NT-PC or Workstations, if documentation is available.
After that a more advanced step would be possible: Two-pass-encoding,
load balancing depending on processor speed and scene complexity etc.=20
The more I think about it, the more I like the idea: Since with a
shared-lib mechanism it would always use the most recent encode-lib
present, it could be a great tool for optimization. Not for speed,
because running a cluster is surely the exception for users, but for
_quality_ improvements.=20
Since you get almost a linear speedup when just splitting a video into n
parts for n processors, you could run several tests of different
parameters over night (when most machines are not used anyway).=20
Hey, I like that idea!=20
Now who has a documented encoder by hand to turn into a cluster-version?
How about the MPlayer? Is A'rpi reading this?=20
chl
Topic: Auto fast/slow motion detection in encoder?
Author: stobo
Posted: 2001-02-23 14:38
------------------------------------------------------------------------
--------
Quote:
On 2001-01-25 10:29, Anonymous wrote:
Microsoft uses a new method in its new WM8 codec to produce better image
quality in high motion scenes: 2-pass . The encoding process is done
twice. In the first run the media is analyzed for high and low motion
scenes. That is to set the best highest possible bitrate for each scene.
The low motion scenes don't require such a high bitrate so more space is
available for the high motion scenes. I've tested it, and I think it is
a really good method for long films that should fit on 1 CD.
So, to come to a conclusion: what about intergrating such a feature in
opendivx?
i thought the parameter on how many frames to averate the bitrate on
does just this.. during this number, the quantization is changed to
slowly react to the differences in encoding efficiency. quantization is
the absolute quality of one frame, so when there is a slow scene, the
codec ends up using less bits per frame 'cause motion prediction already
gets all the quality..
try to set this number to something like 10000 (or more for a full
movie)
am i wrong? (should go look at virtualdub's bits/frame graph to verify
this.. maybe some player also can show bitrate variations over time.)[
This message was edited by: stobo on 2001-02-23 14:42 ]
Topic: Main differences between MPEG2 and DivX/MPEG4 ?
Author: stobo
Posted: 2001-02-23 14:47
------------------------------------------------------------------------
--------
Quote:
On 2001-02-12 15:36, u08194 wrote:
>From what I saw, most of the difference is due to H673 quantizer being
used instead of the MPEG quantizer. =20
MPEG4 is also supposed to support adaptive quantization matrixes and
quarter pell motion prediction. None of these are currently
implimented. May be adaptive quantizer will be implimented as part of
VBR.
=20
i thought that the longer keyframe interval plus wider motion prediction
ranges also have something to do with this..
and decoder accuracy! the encoder can really assume that minor
modifications are still valid if the frame data has been updated 100
times since the last key block. then again accuracy might be part of the
quantizer spec.
Topic: Awesome quality!! but porblems with overlay driver in
fullscreen.
Author: stobo
Posted: 2001-02-23 14:50
------------------------------------------------------------------------
--------
Quote:
On 2001-02-12 10:15, musicvid wrote:
Btw i cant see any codec other than the mp3 codec in properties on
mediaplayer but dont think thats the problem since it plays openDivX
files.
i believe the story is simple as that there is not a directshow filter,
just the old-fashioned avi codec, and windows media player is stupid
enough not to use the overlay driver in this case.. removing the filter
from any divx or other decompressor causes just the same.
bottom line: get a better player software=20
Topic: OpenDivX Variable Bitrate Encoding
Author: stobo
Posted: 2001-02-23 15:05
------------------------------------------------------------------------
--------
Quote:
On 2001-01-25 08:05, Sandman wrote:
Hi,
all around they write about "true Variable Bit Rate" ..what about a
"false Constant Bit Rate"? I mean: you set the overall medium bitrate of
the movie/clip/video and the encoder uses it freely to achieve constant
quality in all scenes..you should say to the encoder "I have a clip 45
min long and I want it to fit EXACTLY a 650Mb CD, doesn't matter if a
frame uses 1 byte and another one 1 Mb" ..obviously to do so the encoder
should have a "preprocessing step" that takes almost the same time of
the real compression, in order to decide how many bytes assign to each
frame...Divx4 has some flexibility in this sense, but it was tought on
the range of few frames, I think...
Anyone has ideas about such a thing?
(I'm not a programmer,unluckly)
Sand
i thought the min/max quantization and the average bitrate over n frames
parameters were just about this.. try setting the frame count very high
and see what happens
--__--__--
Message: 2
Date: Fri, 23 Feb 2001 18:16:58 -0800
From: "Darrius \"Junto\" Thompson" <[email protected]>
To: <[email protected]>
Subject: [OpenDivX] OpenDivX Decoder ->OpenDivX Daily Forums Digest
OpenDivX Decoder ->OpenDivX Daily Forums Digest
Topic: Faster code
Author: mcmab
Posted: 2001-02-23 01:55
------------------------------------------------------------------------
--------
Southampton????
At the University by any chance???
Tell me where to e-mail the whole file and I'll send it.
If you're at the University we will have friends in common (and we're
6000 miles apart)
Topic: Decore Options to optimize for smaller devices
Author: mindpower
Posted: 2001-02-23 02:54
------------------------------------------------------------------------
--------
Feature suggestions:
- flag to make decoded BMP compatible with the CreateBitmap routine in
Windows. (i.e. e.g. lines need to begin at word boundary, rest filled
with zeroes)
- flags to make decoded BMP compatible with 8-bit, 16-bit display
buffers. This way I could let decore decode directly from a memory
mapped file into the display buffer giving RAW PERFORMANCE.
- flag to automatically reduce frame size so that I only get a smaller
frame in the decoding buffer.=20
- flag to automatically dither when size get's reduced or dither anyway
even if it get's not reduced.
cheers
Mindpower
Topic: Divx4windows 4a48 codec crashes in decoding.
Author: wanton
Posted: 2001-02-23 04:50
------------------------------------------------------------------------
--------
I encoded file with default setting and playing it with mediaplayer of
loading it to virtualdub makes it crash virtualdub first show couple of
frame garbage and then crashes. But playing file with theplaya works
fine so problem must be in decoder part of windows codec.
--__--__--
Message: 3
From: Christian Buchner <[email protected]>
To: [email protected]
Subject: Re: [OpenDivX] using mpeg2 decoding hardware for divx decoding
Date: Fri, 23 Feb 2001 07:54:22 +0100
Am Donnerstag, 22. Februar 2001 13:12 schrieb zap twentyfive:
> Hello,
>
> I'm a software engineer working for Philips semiconductors and currently
> I'm looking into porting parts of divx to Philips Trimedia:
> Both platforms have hardware support for MPEG2 decoding as well as video
> coprocessor units which can do things like filtering and colorspace
> conversion.
Well, the main problem that you will have is a compatibility issue.
OpenDivX in its current state is not compatible to the original DivX ;-)
codec. That is because DivX ;-) is a hacked Microsoft MPEG-4 V3 codec.
And Microsoft has decided to diverge slightly from the ISO MPEG-4 standard
in an undocumented way. That is probably why up to date the OpenDivX
codecwill not decode DivX ;-) movies, but instead uses the "official" ISO
MPEG4 format (it is based on ISO reference software, actually. Has anyone
checked whether this is legal in the first place?)
Sincerely,
Christian Buchner
--__--__--
Message: 4
Date: Fri, 23 Feb 2001 18:47:53 -0800
From: "Andrea Graziani" <[email protected]>
To: <[email protected]>
Cc: <[email protected]>, <[email protected]>
Subject: [OpenDivX] DirectShow Filter Bugs
If you haven't already noticed we have released a DirectShow filter.
The DirectShow filter has a couple of problems... I still didn't figure
out how to solve them so I'm posting these known issues on the forum. I
hope some good guy can come out with some suggestion.
1) The filter uses the VIDEOINFOHEADER2 structure. This solves an
incompatibility problem between the way that the GUI and DirectDraw
interpret the field biHeight of the BITMAPINFOHEADER structure...
without this fix sometimes the output bitmap appears reversed until the
decoder reaches a keyframe. By the way... this structure seems to work
only with an Overlay Mixer filter connected to the output pin of the
DivXFilter and some graphic card don't like this. For example, on the
i810 graphic chip the filter can't work.
2) WMP goes in deadlock when I click on stop during the playback. This
fact made me crazy today ;-( Since the filter is overriden from the
CVideoTransform basic class I suppose that all the thread/critical
section management should be done in the base class for me. I'm so
wrong?
Thanks!
Andrea
--__--__--
Message: 5
Date: Mon, 26 Feb 2001 09:46:07 -0800
From: "Darrius \"Junto\" Thompson" <[email protected]>
To: <[email protected]>
Subject: [OpenDivX] General Development Discussion - All Projects -> OpenDivX Daily Forum Digests
General Development Discussion - All Projects -> OpenDivX Daily Forum
Digests=20
Topic: PCTVPRO/Importing DivX
Author: divxrocks
Posted: 2001-02-24 21:42
------------------------------------------------------------------------
--------
I love DivX and Mayo. I was using my PCTVPRO to capture video from VCR
with the Studio PCTV software and it was decent when saving the file as
a real meadia file. My friend told me to try out divx. Well I can't
import codecs with the Studio PCTV program, so I used Adobe Primiere and
that was AWESOME!! But, I don't like having to render the file before I
see the transition that I add. With the PCTV Studio software, it shows
the transitions instantly as soon as I add it without having to render.
I would REALLY like to know if there is some way that I could trick
the Studio PCTVPRO software into using the Divx Codec?
Thanks allot!![ This message was edited by: divxrocks on 2001-02-24
21:44 ]
Topic: Speaking of better codecs...
Author: Brians256
Posted: 2001-02-25 00:17
------------------------------------------------------------------------
--------
I would suggest against getting 0.0.7 just yet. Stick with 0.0.6 until
some important fixes are introduced.
It doesn't compile with microsoft C++ compiler, and the author has fixed
some bugs in main.cpp and intra_dc.cpp. It might work in Linux, but I
don't run Unix anymore (much to my regret).
Oh, I get the same distortions in the encoded frame with 0.0.6. Perhaps
it is fixed in 0.0.7 (or in an unreleased patch?).
I'm still trying to get it to compile with a replacement implementation
of hash_map, and it doesn't quite work yet.[ This message was edited by:
Brians256 on 2001-02-25 00:18 ]
Topic: Not YCbCr420?
Author: Brians256
Posted: 2001-02-25 00:28
------------------------------------------------------------------------
--------
It looks like I jumped the gun in responding to the question. Sorry!
The web pages I read said that YUV was equivalent to YbCbCr. However,
even after reading the page that Eagle recommended, I'm still unclear as
to the difference between the two.
Is YbCbCr a nonlinear conversion from the RGB colorspace? YUV is (as
far as I can tell) a linear conversion.
Topic: Speaking of better codecs...
Author: Isibaar
Posted: 2001-02-25 13:20
------------------------------------------------------------------------
--------
Brians, I compiled version 0.0.7 yesterday with the MS VC6. My initial
thought was right, you need STLport to compile. I downloaded Version 4
of STLport from www.stlport.org
The problem with the hash_map implementation will then be solved.
Moreover I had to make some minor changes to the code and the project
environnement.
Unfortunately my build always crashes during encode with VirtualDub and
I had not the time to look over why.
But I just downloaded Sparky 0.0.8 and hope to get better results.
Topic: Newbie help
Author: maul1616
Posted: 2001-02-25 16:40
------------------------------------------------------------------------
--------
Hi, I have been trying for weeks and weeks now to get divx running on my
system. I have installed the new divx 4 codec and then run "run me
first." When i try to run a short divx clip it serches for a codec and
it can not find it. then it just plays the audio. Im running on
windows 98se with t-bird 800 mhz chip. Any help would be good. this is
killing me!! thanks
Topic: Speaking of better codecs...
Author: Brians256
Posted: 2001-02-25 19:22
------------------------------------------------------------------------
--------
I've downloaded 0.0.8 and am testing it. It compiles (at this point)
and I'll see if the quality issues are gone.
Then, I can start working on helping performance.=20
Topic: Newbie help
Author: Brians256
Posted: 2001-02-25 19:29
------------------------------------------------------------------------
--------
You are probably trying to play a clip created with the older Divx
codec. They are not compatible! You can safely have both codecs on
your computer, but the newer Divx codec will not play stuff made with
the older codec. The older Divx codec will not play movies encoded by
the newer codec.
Apples and oranges.
Topic: Windows : codename Whistler
Author: #AS#IS#
Posted: 2001-02-25 19:52
------------------------------------------------------------------------
--------
Is the current stage of opendivx compatable with Windows WP. Has anyone
bothered to test on this platform.
Im assuming it'll be no different to ME or 2000 but we all know
Microsoft dont we ?
--__--__--
Message: 6
Reply-To: "Andrea Graziani" <[email protected]>
From: "Andrea Graziani" <[email protected]>
To: "[email protected]" <[email protected]>,
"[email protected]" <[email protected]>,
"[email protected]" <[email protected]>
Date: Mon, 26 Feb 2001 18:25:06 -0800
Organization: ProjectMayo
Subject: [OpenDivX] DirectShow Filter Bugs
This is a multi-part message in MIME format.
------=_NextPart_000_001F_01C0A021.7162D000
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
I solved the problem of the STOP deadlock in the DirectShow filter. I =
wait for some input for the behavior of the filter with different =
graphic cards (see point 1 of previous e-mail).
Andrea
------=_NextPart_000_001F_01C0A021.7162D000
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META content=3D"text/html; charset=3Diso-8859-1" =
http-equiv=3DContent-Type>
<META content=3D"MSHTML 5.00.2920.0" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT face=3DArial size=3D2>I solved the problem of the STOP =
deadlock in the=20
DirectShow filter. I wait for some input for the behavior of the filter =
with=20
different graphic cards (see point 1 of previous e-mail).</FONT></DIV>
<DIV><FONT face=3DArial size=3D2>Andrea</FONT></DIV>
<DIV> </DIV></BODY></HTML>
------=_NextPart_000_001F_01C0A021.7162D000--
--__--__--
Message: 7
Date: Mon, 26 Feb 2001 18:36:04 -0800
From: "Darrius \"Junto\" Thompson" <[email protected]>
To: <[email protected]>
Subject: [OpenDivX] OpenDivX Decoder ->OpenDivX Daily Forum Digest
OpenDivX Decoder ->OpenDivX Daily Forum Digest
Topic: Decore Options to optimize for smaller devices
Author: mindpower
Posted: 2001-02-26 04:51
------------------------------------------------------------------------
--------
Ok, I found that the yuv2RGB565 variant is already compatible with iPAQs
display buffer. Only problem: the image is upside down.
So I change my above request to this:=20
- image flipping option to do so while decoding
- image resizing would still be great (at deconding time)
- dithering would be nice, but so far it is not essential, the video
looks great without it and without postprocessing, too.
Topic: Faster code
Author: mindpower
Posted: 2001-02-26 05:01
------------------------------------------------------------------------
--------
I am about to start to assembly-optimize vital parts of decore for
StrongARM (iPAQ)
Can anyone - who profiled the C code - tell me which routines take the
longest and thus would benefit the most from assembler?
cheers
Mindpower
Topic: DirectShow filter and relative problems
Author: vn04697
Posted: 2001-02-26 09:33
------------------------------------------------------------------------
--------
May be it will be interesting to you: VIDEOINFOHEADER2 also doesn't work
on my NVIDIA TNT2 card. I was forced to drop call to
ConvertVideoInfoToVideoInfo2() in GetMediaType().
Quote:
On 2001-02-23 21:50, e7abe7a wrote:
If you haven't already noticed we have released a DirectShow filter.
The DirectShow filter has a couple of problems... I still didn't figure
out how to solve them so I'm posting these known issues on the forum. I
hope some good guy can come out with some suggestion.
1) The filter uses the VIDEOINFOHEADER2 structure. This solves an
incompatibility problem between the way that the GUI and DirectDraw
interpret the field biHeight of the BITMAPINFOHEADER structure...
without this fix sometimes the output bitmap appears reversed until the
decoder reaches a keyframe. By the way... this structure seems to work
only with an Overlay Mixer filter connected to the output pin of the
DivXFilter and some graphic card don't like this. For example, on the
i810 graphic chip the filter can't work.
2) WMP goes in deadlock when I click on stop during the playback. This
fact made me crazy today ;-( Since the filter is overriden from the
CVideoTransform basic class I suppose that all the thread/critical
section management should be done in the base class for me. I'm so
wrong?
Thanks!
Topic: Faster code
Author: eagle
Posted: 2001-02-26 10:44
------------------------------------------------------------------------
--------
mcmab: you can email stuff to me at <a
href=3D"mailto:[email protected]";
target=3D"_new">[email protected]</a> but I can't promise a quick
turnaround as I'm very busy right now. Unless of course you can tell me
that your changes make the code run 300% faster=20
I was at Soton uni six years ago - when were you there?
Mindpower: The iDCT and motion compensation parts usually need the most
horsepower. However, I wouldn't recommend that you start on any
hard-core optimisiation without some means of measuring your changes.
The GNU profiler, gprof, is available for most platforms and is very
useful.
Topic: Faster code
Author: mcmab
Posted: 2001-02-26 12:22
------------------------------------------------------------------------
--------
It will be sent today.
300%? Well if we assume that the speed of the program is determined by
memory reads - which looks likely to be the case - the above code, as
noted does 64 reads. In contrast the original code did 1024 reads to
accomplish the same result. That piece of code could be 1000% faster
IMHO.
I was at Soton 81-86. I lived in a house that stood where the back
carpark of the new electronics bldg is located. I presume you were in
Electronics/CS? My friend managed the project that created the newest
Electronics facility, and I'm proud of him - its a fine job. Note
especially the "porthole" type windows that echo Soton's maritime
history. While there I pioneered data acquisition on small Unix boxes,
Z80 & 68000. Hence my strong background in assembler and the then new
"C" language.
Topic: DirectShow filter and relative problems
Author: e7abe7a
Posted: 2001-02-26 13:19
------------------------------------------------------------------------
--------
Mmmmh... Did you try to download the last driver for your graphic card
before going back to VIDEOINFO? In wich mode is working the renderer
now? RGB?
--__--__--
Message: 8
Date: Mon, 26 Feb 2001 18:36:56 -0800
From: "Darrius \"Junto\" Thompson" <[email protected]>
To: <[email protected]>
Subject: [OpenDivX] OpenDivX Encoder -> OpenDivX Daily Forum Digests
OpenDivX Encoder -> OpenDivX Daily Forum Digests
Topic: Does anybody want this?
Author: gruel
Posted: 2001-02-26 02:46
------------------------------------------------------------------------
--------
Quote:
On 2001-02-25 15:56, czw wrote:
The distributed workload code existed from an experiment at the
university, so the only thing we needed to do was to distribute a
different set of data.
[...]
First stop is support for Linux and Win32, support for other OS:es will
come as we find systems or get patches.=20
Well, I meant details on the implementation:=20
What is going to be parallel afterwards? What protocols do you use for
exchange of data?=20
chl
P.S. spatz is right: Beer in the sauna is soooooo swedisch ;-)
Topic: Awesome quality!! but porblems with overlay driver in
fullscreen.
Author: BetaBoy
Posted: 2001-02-26 03:15
------------------------------------------------------------------------
--------
Quote:
On 2001-02-24 16:02, ananonymoususer wrote:
sorry to go off topic, is there a web page where i can learn about
mpeg-4 specs
Here are a few links for you:
<!-- BBCode Start --><A HREF=3D"http://www.cselt.it/mpeg/standards.htm";
TARGET=3D"_blank">www.cselt.it/mpeg/standards.htm</A><!-- BBCode End -->
<!-- BBCode Start --><A HREF=3D"http://www.m4if.org/index.html";
TARGET=3D"_blank">www.m4if.org/index.html</A><!-- BBCode End -->
<!-- BBCode Start --><A
HREF=3D"http://www.cselt.it/mpeg/documents/koenen/mp4ieee.htm";
TARGET=3D"_blank">www.cselt.it/mpeg/documents/koenen/mp4ieee.htm</A><!--
BBCode End -->
-----------------
Dan Marlin
<!-- BBCode Start --><IMG
SRC=3D"http://www.mydivx.com/images/news/divxvideo.gif";><!-- BBCode End
-->
<!-- BBCode Start --><A HREF=3D"http://www.mydivx.com";
TARGET=3D"_blank">www.mydivx.com</A><!-- BBCode End -->
'We are the Music Makers and the Dreamers of the Dreams.' WW[ This
message was edited by: BetaBoy on 2001-02-26 03:17 ]
Topic: Does anybody want this?
Author: SVB
Posted: 2001-02-26 03:35
------------------------------------------------------------------------
--------
So you are gonna develop some distributed Win32 (and Linux) appllication
for video encoding, do I understand right? Something independent of
OpenDivX and not involving any changes in encoder.
Do you want it to be useable with other codecs too?
(I've never drunk beer in the sauna, sorry ;-)[ This message was edited
by: SVB on 2001-02-26 03:36 ]
Topic: Does anybody want this?
Author: czw
Posted: 2001-02-26 07:29
------------------------------------------------------------------------
--------
Quote:
Well, I meant details on the implementation:=20
What is going to be parallel afterwards? What protocols do you use for
exchange of data?=20
The principle is easy: distribute snippets to one or more encoders (say
eight seconds per block). This block will now be encoded and sent back
to the client who requested a movie encoding. When all eight-seconds
parts have arrived in encoded form, they are glued together.
Protocol? It's a really simple telnet-style protocol where the commands
are cleartext and data is sent in binary with CRC32 checksum and
(optional) data compression. All of this communication code is old code
being reused for the third or fourth time.
Topic: Does anybody want this?
Author: gruel
Posted: 2001-02-26 10:09
------------------------------------------------------------------------
--------
Quote:
On 2001-02-26 07:29, czw wrote:
The principle is easy: distribute snippets to one or more encoders (say
eight seconds per block). This block will now be encoded and sent back
to the client who requested a movie encoding. When all eight-seconds
parts have arrived in encoded form, they are glued together.
Hej igen,
are you sure that it's useful to send away such short sniplet of fixed
length? If you use some sort of MPEG-style input, it would surely be
best to only cut at I-frames. But except for OpenDivx and some patched
Divx-versions there still may be produced too many of those. The
advantage of I-frames only at scene-changes should in no way be lost.=20
Also the motion detection should be able to do some "look ahead",
otherwise the method will be fast, but give poor quality.=20
Anyhow, I would be very much interested in the technicalities. Do you
want to extend the encoding-routine itself for this, or write some kind
of "wrapper", that handels snipping and merging again, or is it going to
be a parallel application, calling the ordinary encoder?
Is it perhaps possible for me to look into some of your code? Email is
<a href=3D"mailto:[email protected]"; target=3D"_new">[email protected]</a>
chl
P.S. If we continue like this, there might soon be to be a "parallel
encoding" forum ;-)
Topic: Alpha 48 Bitrate slow to adjust?
Author: tomnot
Posted: 2001-02-26 10:16
------------------------------------------------------------------------
--------
Guys,
Great job with the new release.
I get blockiness when changing to some scenes, (maximum bitrate @
448x192). Any chance someone can make the codec alter the bitrate
faster? The DivX low motion did not seem to have this problem.
Hope it gets even better!
Tom.
Topic: Does anybody want this?
Author: czw
Posted: 2001-02-26 11:01
------------------------------------------------------------------------
--------
Quote:are you sure that it's useful to send away such short sniplet of
fixed length? If you use some sort of MPEG-style input, it would surely
be best to only cut at I-frames. But except for OpenDivx and some
patched Divx-versions there still may be produced too many of those. The
advantage of I-frames only at scene-changes should in no way be lost.
Also the motion detection should be able to do some "look ahead",
otherwise the method will be fast, but give poor quality.
You are correct - we will do some primitive data analysis to break up at
I-frames only. I don't have extensive knowledge in the encoding process
yet, so I can't give a comment about the motion detection. Is the image
quality reduced if there are more I-frames or less I-frames in a scene?
Is there a big processing time difference? And why the hell is the
source code divided in separate places for Win32, Linux, Mac and Core
instead of just one XP source repository?
Quote:Anyhow, I would be very much interested in the technicalities. Do
you want to extend the encoding-routine itself for this, or write some
kind of "wrapper", that handels snipping and merging again, or is it
going to be a parallel application, calling the ordinary encoder?
This will act as a separate application. The OpenDivX code will never
realize that this is a part of a larger distributed task. By doing this,
the encoder engine is free to do large changes without affecting our
code.
About sending code: we keep the source in a private CVS server until it
can be used in the most basic way.
Topic: Does anybody want this?
Author: gruel
Posted: 2001-02-26 12:46
------------------------------------------------------------------------
--------
Quote:
On 2001-02-26 11:01, czw wrote:
Is the image quality reduced if there are more I-frames or less I-frames
in a scene?
I-frames encode the whole frame, P- and B-frames only the differences to
others. So more I-frames make the file _larger_ in first place. If you
stick to a fixed data rate, that will lead to worse quality, of course.
Quote:
About sending code: we keep the source in a private CVS server until it
can be used in the most basic way.
Yes, that's what I thought, and that's why I asked you for a copy. ;-)=20
I'm just interested in any (even non-compiling) snapshot, to see if the
methods are similar to what I had plans for with MPI.=20
But never mind...
chl[ This message was edited by: gruel on 2001-02-26 12:47 ]
--__--__--
_______________________________________________
OpenDivX mailing list
[email protected]
http://lists.projectmayo.com/mailman/listinfo/opendivx
End of OpenDivX Digest
Reply To Poster
Local References / HOW-TO / FAQs