[FFmpeg-user] Use of "-color_range"

[Thomas Worth]

On Tue, Jul 5, 2011 at 12:50 AM, Tim Nicholson <tim.nicholson at bbc.co.uk> wrote:
> On 05/07/11 08:30, Thomas Worth wrote:
>>
>> On Mon, Jul 4, 2011 at 11:33 PM, Tim Nicholson<tim.nicholson at bbc.co.uk>
>>  wrote:
>>>
>>> On 04/07/11 18:13, Thomas Worth wrote:
>>>>
>>>> On Mon, Jul 4, 2011 at 7:44 AM, Tim Nicholson<tim.nicholson at bbc.co.uk>
>>>>  wrote:
>>>>>
>>>>> I am struggling to find any documentation on the use of this and
>>>>> searching
>>>>> the source was not very revealing, to my eye at least.
>>>>>
>>>>> Apart from the fact that it expects an<int>    I am none the wiser.
>>>>>
>>>>> I since it is a DEV function I was hoping I could use it to force
>>>>> ffmpeg
>>>>> to
>>>>> handle source material of a non standard range.
>>>>
>>>> Regarding this and x264 encoding: I had some full range 4:2:2 material
>>>> I was trying to encode with x264 a while back and found that if I
>>>> forced the "yuvj420p" pix format, FFmpeg would pass full range values
>>>> to x264. This was using x264 in pipe mode, but I don't see why it
>>>> wouldn't work with libx264. I would get very different results when
>>>> changing yuv420p to yuvj420p.
>>>
>>> Quite! Except that if you look at pixfmt.h it says (line 74)
>>>
>>> "PIX_FMT_YUVJ420P,  ///<  planar YUV 4:2:0, 12bpp, full scale (JPEG),
>>> deprecated in favor of PIX_FMT_YUV420P and setting color_range"
>>>
>>> Which begs the original question.
>>
>> Does this help? From libavcodec/h264.c:
>> [...]
>
> I think Marks extract was a little more straightforward..... :)

Quite. ;-)