Forum Replies Created

Page 7 of 24
  • Michael Rampe

    July 6, 2011 at 11:37 pm in reply to: Problem with beginning of 2-pass encoded video

    OK. did some more digging and testing and got some solid results and answers.

    To use a football metaphor, this is a case of “it is the ball, not the player”

    Check out this post:
    https://doom10.org/index.php?topic=1343.0

    Dark Shikari (aka Jason Garrett-Glaser – developer of x264) states that youtube has a broken mp4 demuxer.

    I followed his suggestions and rewrapped the file to an mkv wrapper.
    $ffmpeg -i badfile.mp4 -vcodec copy goodfile.mkv

    This worked (no grey mess at the start) with the B-frames intact.
    My other test (without B-frames) also worked but who wants to be using a baseline profile for HD video…. no one;-)

    Michael

  • Michael Rampe

    July 6, 2011 at 9:45 pm in reply to: Problem with beginning of 2-pass encoded video

    [William Lipinski] “Regardless, problem solved. Cheers!”

    Maybe.

    I just replicated your problem and also got the grey frames at the start using the -crf 1 pass method I suggested.

    After some searching I found this post on the youtube support site:
    https://www.google.com/support/forum/p/youtube/thread?tid=33d1cf29da811e45&hl=en

    This seems quite suspect to me but I am currently uploading a baseline profile video (no B-frames) to make sure. If it is true, it is quite a letdown from youtube;-(

    Michael

  • Michael Rampe

    July 5, 2011 at 11:46 pm in reply to: .3GP encoding. SOS!

    [John Hicks] “Thing is, there’s a few videos that need to be encoded to 3GP. Each video has to be encoded into 6 different .3gp’s:”

    Tricky task indeed seeing you have been supplied with patchy information.

    What you need to know is:
    AUDIO: codec, sample rate, channels, bitrate
    VIDEO: codec (and profile for h264), frame size, bitrate, framerate

    The information you have been provided does not give you all of these details and looks like it was copied and pasted from several phone manufacturers tech specs;-)

    Summary:

    1. AUDIO – AAC 24Khz [unknown audio channels][unknown audio bitrate], VIDEO – h263 176X144 [unknown video bitrate][up to 30fps – need specific framerate]
    2. AUDIO – AAC [16khz-32khz – need specific sample rate] Stereo [8kbps-80kbps – need specific audio bitrate], VIDEO – h263 320X240 64kbps [up to 30fps – need specific framerate]
    3. AUDIO – AMR [8khz-16khz – need specific sample rate] Mono [8kbps-80kbps – may need specific sample rate, see note below], VIDEO – h263 320X240 64kbps [up to 30fps – need specific framerate]
    4. AUDIO – AMR 8khz Mono [unknown audio bitrate], VIDEO – h263 128X96 [unknown video bitrate][unknown video framerate]
    5. AUDIO – AMR [8khz-16khz – need specific sample rate] Mono [8kbps-80kbps – may need specific sample rate, see note below], VIDEO – h263 176X144 64kbps [up to 30fps – need specific framerate]
    6. AUDIO – AAC [16khz-32khz – need specific sample rate] Stereo [8kbps-80kbps – need specific audio bitrate], VIDEO – [h264 – need profile restrictions, if any] 176X144 64kbps [up to 15fps – need specific framerate]
    NOTE: “The AMR codec uses eight source codecs with bit-rates of 12.2, 10.2, 7.95, 7.40, 6.70, 5.90, 5.15 and 4.75 kbit/s.” … from wikipedia

    All of the fields in square brackets are either missing or vague.

    Apart from all of that, FFmpeg is the tool that can achieve all of this.

    Basic syntax for your needs is:

    ffmpeg -i [inputfile] -vcodec -b -s [size] -r -acodec -ar -ab -ac output.3gp

    (for H.264, a preset is needed – either using -vpre or -preset depending on FFmpeg version)

    Hope this helps. Maybe ask for a bit more information from your boss and I will see what more I can do.

    Michael

  • [Terry McDanel] “FFmpegX 0.0.9y”

    FFmpegX is quite outdated now and no longer maintained. Have a look at the iFFmpeg GUI for a better OSX alternative.

    Michael

  • Michael Rampe

    July 4, 2011 at 11:18 pm in reply to: Command Line to Convert AVI to MPEG-2 (ts)

    [Francis Lee] “I’ve tried it with dotnet ffmpeg but I keep getting an Mpeg-2(ps) as the output.”

    Can you post what you have tried? (command line)

    Michael

  • Michael Rampe

    July 4, 2011 at 11:17 pm in reply to: using ffmpeg to convert libx264 to MOV DVNTSC

    [Adam Hicks] “Does it even make sense to be using ffmpeg to do what I need to do?”

    Have a look at FFMBC which is a customised version of FFmpeg exactly for your purposes.
    https://code.google.com/p/ffmbc/

    Michael

  • Michael Rampe

    July 4, 2011 at 11:04 pm in reply to: issues converting sequential images to video

    [Richard Jeong] “Any notion as to why it’s not accepting the sequential files?”

    Not sure. I tried to replicate the issue and your line “ffmpeg -f image2 -i /tmp/img%05d.jpg /tmp/a.mpg” worked fine for me.

    Maybe try an update…. Not sure what else to offer as I cannot replicate.

    Michael

  • Michael Rampe

    July 4, 2011 at 10:57 pm in reply to: -vcodec copy -acodec copy

    [richie bee] “is there any reason why I shouldn’t use -vcodec copy?”

    Depends on what you are doing.

    If you are just repacking the stream into a different container or splitting the audio from the video, -vcodec copy is the right choice as it does exactly what you would expect, copy the stream without re-encoding.

    If you are changing bitrate, size, aspect or applying filters, this is not a good choice.

    Michael

  • Michael Rampe

    July 4, 2011 at 10:55 pm in reply to: Adaptive quantization

    [Stan Juznitok] “There are few of tehm that can be reached from x264 itself, as described in:
    https://sites.google.com/site/linuxencoding/x264-ffmpeg-mapping

    This lists all of the options…..
    Which ones specifically are you after?

    Michael

  • [matt kim] “I am aware that positioning the -ss option before the -i option may allow it to run faster. But it would seem that the videos that I am filtering are unable to do this as well. So at the moment the -ss option runs very slowly–about 1 sec per 1 minute.”

    Putting the -ss option before -i skips through the I frames only and is not truly frame accurate. Putting it after the -i decodes every frame to find the exact one you set and is thus much slower.

    What is your source codec?

    Michael

Page 7 of 24

We use anonymous cookies to give you the best experience we can.
Our Privacy policy | GDPR Policy