[FFmpeg-user] "vibrance" video filter, default luma coefficients swapped?

Jim DeLaHunt list+ffmpeg-user at jdlh.com
Fri Apr 23 04:04:11 EEST 2021

On 2021-04-19 00:55, Jim DeLaHunt wrote:

> On 2021-04-14 22:33, Jim DeLaHunt wrote:
>> ...So, luma = 0.2126 * r + 0.7152 * g + 0.0722 * b, a value in [0.0, 
>> 1.0] for r, g, b in [0.0, 1.0]
>> But the default values I see in the FFmpeg vibrance filter are[4]:
>>   rlum (default 0.072186), glum (default 0.715158), blum (default 
>> 0.212656)
>> Thus it looks like the default values for rlum and blum are swapped....
> I entered ticket #9195, "vibrance (video filter) incorrect defaults 
> for luma", https://trac.ffmpeg.org/ticket/9195 , to track this.

Someone closed this ticket as "wontfix", without comment about why the 
current behaviour is correct, or if incorrect, why the ticket should be 
ignored. I suspect that person is the author of the vibrance filter. If 
that is true, then the author of the filter is declining to respond to a 
good-faith bug report about the filter.

That makes me despair of getting the filter corrected (if incorrect). I 
think that is a sad situation for the project.

More information about the ffmpeg-user mailing list