[FFmpeg-user] Please can someone tell me why this ffmpeg command stopped working

[Paul Gardiner]

On 29/01/2012 00:00, dE . wrote:
> On 01/29/12 00:29, Paul Gardiner wrote:
>> On 28/01/2012 12:20, dE . wrote:
>>> On 01/28/12 22:41, Paul Gardiner wrote:
>>>> On 28/01/2012 16:49, Reindl Harald wrote:
>>>>>
>>>>>
>>>>> Am 28.01.2012 17:33, schrieb Paul Gardiner:
>>>>>> On 28/01/2012 16:24, Paul Gardiner wrote:
>>>>>>> On 28/01/2012 16:10, Reindl Harald wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>> Am 28.01.2012 16:53, schrieb Paul Gardiner:
>>>>>>>>> Can anybody see what is wrong with this:
>>>>>>>>>
>>>>>>>>> /usr/bin/ffmpeg -f ac3 -ar 48000 -i /tmp/mythar4336/audout -f
>>>>>>>>> rawvideo -top 1 -pix_fmt yuv420p -s 1440x1088 -r 25
>>>>>>>>> -i /tmp/mythar4336/vidout -threads 2 -vf
>>>>>>>>> crop=1440:1080:0:0,scale=1280:720 -vcodec libx264 -preset medium
>>>>>>>>> -crf
>>>>>>>>> 20.0
>>>>>>>>> -bufsize 25000k -maxrate 20000k -aspect 16:9 -acodec copy -f mp4
>>>>>>>>> /home/public/video/Archive/Horizon - Playing God.mp4
>>>>>>>>>
>>>>>>>>> Until a recent ffmpeg update, it used to do exactly what I
>>>>>>>>> wanted, but
>>>>>>>>> now it produces the error message: "Option sample_rate not
>>>>>>>>> found." I've
>>>>>>>>> googled and found various references to people having to make sure
>>>>>>>>> arguments are in the right order with later versions of ffmpeg,
>>>>>>>>> but
>>>>>>>>> my command seems to obey the rules, plus attempts I have made to
>>>>>>>>> reorder
>>>>>>>>> have not helped.
>>>>>>>>
>>>>>>>> "-f ac3 -ar 48000 -i"
>>>>>>>>
>>>>>>>> why is "-ar" BEFORE the input file?
>>>>>>>> in this case you set the parameter for the input instead the output
>>>>>>>
>>>>>>> My thinking was that the input was 48kHz and I needed to tell ffmpeg
>>>>>>> that information about the input. I guess that's showing my
>>>>>>> misunderstanding about how it works.
>>>>>
>>>>> i would say ffmpeg is knowing more about the input file as you and i
>>>>> will
>>>>> ever be able and there may be only very few cases where it needs help
>>>>>
>>>>> the output should specifuied so exactly as possible if you nedd videos
>>>>> palying in iPhone as example to make sure they are H264 BASELINE but
>>>>> not the input (use the ffmpeg output to see what is really is *gg*)
>>>>>
>>>>>>> Off to try the alteration now.
>>>>>>
>>>>>> Brilliant! It worked. I am again able to transcode my TV recordings.
>>>>>> I guess I should also move the "-r 25" from before the second
>>>>>> input-file argument. Is that right?
>>>>>
>>>>> it is quite simple, all parameters before any file are taken as
>>>>> params for exactly this one
>>>>
>>>> Yes that's completely plain from the documentation, but did you get
>>>> my point that I didn't know whether -ar 48000 should be used to
>>>> tell ffmpeg the sample rate of the input file I was supplying or
>>>> whether it should be used to tell ffmpeg the sample rate I wanted
>>>> in the output file. I know now of course with your help.
>>>>
>>>> Cheers,
>>>> Paul.
>>>> _______________________________________________
>>>> ffmpeg-user mailing list
>>>> ffmpeg-user at ffmpeg.org
>>>> http://ffmpeg.org/mailman/listinfo/ffmpeg-user
>>>
>>> Also, I don't think you should use maxrate, last time I checked it
>>> didn't work with x264. Use -x264opts qp=<a value from 0 (lossless) to 60
>>> (worst quality)> instead.
>>
>> Yes. I suspected that wasn't ideal. Isn't there another option like
>> -preset that I should be using which sets a whole lot of limits in
>> one go?
>>
>> P.
>
> Limits?
>
> -preset is sort of the lossless compression level.
>
> It includes -
>
> ultrafast, superfast, veryfast, faster, fast, medium, slow, slower,
> veryslow, placebo

-profile was the one I was thinking of. I think -profile implicitly sets
maxrate and bufsize.

Paul.