[FFmpeg-devel] [PATCHv2][GSoC]audio filter-use cellular automata to generate tones

Thilo Borgmann thilo.borgmann at mail.de
Thu Mar 26 13:11:03 EET 2020


Am 26.03.20 um 11:58 schrieb Nicolas George:
> Thilo Borgmann (12020-03-26):
>> Let's look at this. There are two things here:
>>
>> a) Is it a good enough solution of the qualfification task Paul has
>> given to Ashutosh?
>>
>> b) Is it unsuitable to apply? Well that might be but we need to get
>> clear about a) for the sake of GSoC.
>>
>> Also, a) & b) might be a reoccuring problem this year, as IIRC Jai
>> also gave a qualification tasks that might be on its own useless to
>> apply to git HEAD. But AFAIKT this is not on devel by now. So please
>> let's concentrate on these two questions. Any thoughts appreciated.
> 
> The original proposal for a tone audio source was from me, more than a
> year ago. It was for testing purposes: to be able to feed filters and
> codecs realistic input without relying on external files. For that, it
> would have been completely internal and bit-exact.
> 
> Now, there is a new proposal. Without looking into it closer, I supposed
> it was based on the same goal: testing. Apparently not. But I have not
> seen that discussed on the list.
> 
> I have no objection to the task if it serves a purpose. But it has
> nothing to do with my original proposal and this version, using an
> external library, would be worthless for testing.

Ok. So if I understand all of this correctly, the overall goal of the tones project is already redundant to your old proposal (or at least partially, Paul?).

Has there been any reason why it never made it into git HEAD?

So for this patch as-is, AFAICT Paul wanted the external lib as qualification task only and finally get things natively implemented during GSoC?

If so, we can first review this patch as RFC to give Ashutosh feedback to his coding & things. And wether or not he'll pass qualification, gets selected for a probably available slot etc, reads as if he's doing the GSoC project finally, we can have Ashutosh work on an implementation of the audio tones project. Assuming there haven't been any fundamental objections to Nicolas old proposal about such a thing already.

And then, it would of course only make sense to do this project based on what Nicolas old proposal was. Assuming it served the same goal.

Nicolas, can you please give a link here to your proposal for Ashutosh to find it easily?

-Thilo





More information about the ffmpeg-devel mailing list