Re: worth to bother? create a set of test video clips for mobile devices

于 2012年02月01日 01:24, Markus Fischer 写道:
> I'd love to see such information.
Thanks for informing. I'll do this tiny project shortly later.
> One thing I realized: quality is subjective;
True, but -qscale is a bit objective and a starting point:)
> Btw, I wouldn't be much concerned about any Google PR things unless your
> primary intention is to make money off of it.
>
No, not for money. It's an investment of personal time for the public 
benefit. But wouldn't any investment strive for efficiency? But you are 
right to not to concern it much, it's not necessary to start it with 
marketing, and there is always a chance to move the information 
elsewhere after published.

Best regards.

_______________________________________________
ffmpeg-user mailing list
ffmpeg-user <at> ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user

Alter YUV values of AVFrames

I'm trying to apply an effect to a video by altering the YUV values using
FFMpeg programmatically in C.

Let's say I want to increase the luminescence of each pixel of each frame by
100.

I tried just altering the first frame of a video stream. I decoded the
frame, added 100 to each Y value in the AVFrame->data[0], encoded the frame
again and saved it into the video.

However when I play the video back, it is not only the first frame that has
been altered, but the first 30 frames.

Why are the other frames effected when I only change the AVFrame->data of
the first frame?

I tried again by changing all the Y, U and V values in AVFrame->data to 0
for only the first frame.

When I play back the video, it starts off completely green like expected,
after the first frame the video stays green for 30 frames. I can see the
other frames start to come slowly through the green for 30 frames and then
suddenly the green disappears and the video plays as normal. Why is it not
just the first frame that is green?

My method is as follows:

Video Stream

 |

Re: Can H263+ be in .3gp container?

于 2012年01月31日 19:39, Carl Eugen Hoyos 写道:
> Or did you only mean that FFmpeg does contain an encoder for H.263+ with no
> connection to which formats it can be muxed into?
Yes, it is this case. Your words serve as a good footnote of the table.
_______________________________________________
ffmpeg-user mailing list
ffmpeg-user <at> ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user

Choosing between experimental libavcodec aac or faac encoders

Hello,

I use ffmpeg as part of automated and batch jobs processes converting video
files to mp4 with H.264 and AAC.
Quality and processing times are important to me, though I currently do not
wish to use AAC+.

My random tests show that the libavcodec aac encoder that comes with ffmpeg
is more time efficient, and has about the same quality.

Around the 0.7-0.8 releases of ffmpeg (up until 0.8.4, I think) I had
several occasions that on my random tests ffmpeg with libavcodec would get
stuck (which for my automated and batch jobs is even worse than a crash).

I chose to use libfaac back then, and did not get back to it until now I'm
upgrading to ffmpeg 0.10.

I redid these random checks now, and was not able to have it get stuck on
the files, and again it was consistently quicker than libfaac.

I noted that I still must pass "-strict experimental" to use that coded.

I would like to know which codec most of you are using, and whether you
feel that the libavcodec one is stable enough to be usable in production
environments.

Thank you very much!

Liran Zvibel

Re: Choosing between experimental libavcodec aac or faac encoders

Liran Zvibel <liranz <at> audish.com> writes:

> Around the 0.7-0.8 releases of ffmpeg (up until 0.8.4, I think) I had
> several occasions that on my random tests ffmpeg with libavcodec would get
> stuck (which for my automated and batch jobs is even worse than a crash).

Did you report those problems?
If not, a report (of such a serious bug) for an old version that 
clearly explains the problems are not reproducible with current git 
master is very welcome!

Carl Eugen

Re: Alter YUV values of AVFrames

killkenny64 <kennethjjgibson <at> gmail.com> writes:

> I tried just altering the first frame of a video stream. I decoded the
> frame, added 100 to each Y value in the AVFrame->data[0], encoded the frame
> again and saved it into the video.
> 
> However when I play the video back, it is not only the first frame that has
> been altered, but the first 30 frames.

Sounds as if the codec the original video was encoded with used a gop-size 
of 30.

Try rawvideo or an intra-only codec (or try forcing a gop-size of 1).

Carl Eugen

Re: swopping channels in a 5.1 layout

Youri Feenstra <youri.feenstra <at> gmail.com> writes:

> Sorry, but I'm using Windows. When I use "amovie=C:\file.mkv" I get
> several errors refering to "C:file.mkv" (without the backslashes).

Shouldn't you use two backslashes?
c:\\file.mkv
(I am just guessing.)

Carl Eugen

Re: Choosing between experimental libavcodec aac or faac encoders

On 1 Feb 2012, at 12:30, Carl Eugen Hoyos wrote:
> 
>> Around the 0.7-0.8 releases of ffmpeg (up until 0.8.4, I think) I had
>> several occasions that on my random tests ffmpeg with libavcodec would get
>> stuck (which for my automated and batch jobs is even worse than a crash).
> 
> Did you report those problems?

I did not file a bug report. 
I can try to recreate that environment, but it's not going to be easy.

My question is about the current state, though, of 0.10 -- Which AAC codec people use? The libfaac or the
"experimental" libavcodec.

AFAIK, AAC is the current "mainstream" acodec on the internet, and I would like to understand what the
recommended coded is. 
If I did not have to add "-strict -experimental" I would pick the ffmpeg implementation, and this is the
reason I'm asking.

Regards,
Liran.

Re: swopping channels in a 5.1 layout

On Wed, Feb 01, 2012 at 10:37:26AM +0000, Carl Eugen Hoyos wrote:
> Youri Feenstra <youri.feenstra <at> gmail.com> writes:
> 
> > Sorry, but I'm using Windows. When I use "amovie=C:\file.mkv" I get
> > several errors refering to "C:file.mkv" (without the backslashes).
> 
> Shouldn't you use two backslashes?
> c:\\file.mkv
> (I am just guessing.)
> 

Actually, I think ':' has to be escaped; it's an argument separator (you
can amovie=file.mkv:si=2 to select stream index 2 for example).

--

-- 
Clément B.

_______________________________________________
ffmpeg-user mailing list
ffmpeg-user <at> ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-user

Re: adding -ss and -t loses the audio stream on the output

On 24/01/2012 20:05, Carl Eugen Hoyos wrote:
> Tim Nicholson<tim.nicholson<at>  bbc.co.uk>  writes:
>
>> So I have a command line to make two versions of a file that works
>> perfectly until I add in "-ss time -t duration". Then although
>> everything seems to work as normal, and the created files are the right
>> duration they lack the audio stream
>
> Is this also reproducible with a shorter command line (without filters)?
> Are external libraries (x264, faac) necessary to reproduce the problem?
>

I get the same result when changing the ouput to dvvideo/pcm16 and with 
only the lavi filter for doing the audio mapping.

If I remove the lavi filter I get the audio so it seems to be that which 
is causing the problem.

--

-- 
Tim

http://www.bbc.co.uk/
This e-mail (and any attachments) is confidential and may contain personal views which are not the views of
the BBC unless specifically stated.
If you have received it in error, please delete it from your system.
Do not use, copy or disclose the information in any way nor act in reliance on it and notify the sender immediately.
Please note that the BBC monitors e-mails sent or received.
Further communication will signify your consent to this.