[FFmpeg-devel] [PATCH]lavf/latmenc: Error out for invalid codecs

[Paul B Mahol]

On 11/1/17, Carl Eugen Hoyos <ceffmpeg at gmail.com> wrote:
> 2017-11-01 15:26 GMT+01:00 James Almer <jamrial at gmail.com>:
>> On 11/1/2017 11:23 AM, Carl Eugen Hoyos wrote:
>>> 2017-11-01 14:39 GMT+01:00 Paul B Mahol <onemda at gmail.com>:
>>>> On 11/1/17, Carl Eugen Hoyos <ceffmpeg at gmail.com> wrote:
>>>>> 2017-11-01 13:32 GMT+01:00 Paul B Mahol <onemda at gmail.com>:
>>>>>> On 11/1/17, Hendrik Leppkes <h.leppkes at gmail.com> wrote:
>>>>>>> On Wed, Nov 1, 2017 at 11:09 AM, Paul B Mahol <onemda at gmail.com>
>>>>>>> wrote:
>>>>>>>> On 11/1/17, Hendrik Leppkes <h.leppkes at gmail.com> wrote:
>>>>>>>>> On Wed, Nov 1, 2017 at 10:44 AM, Paul B Mahol <onemda at gmail.com>
>>>>>>>>> wrote:
>>>>>>>>>> On 10/31/17, Michael Niedermayer <michael at niedermayer.cc> wrote:
>>>>>>>>>>> On Mon, Oct 30, 2017 at 11:51:30PM +0100, Carl Eugen Hoyos wrote:
>>>>>>>>>>>> Hi!
>>>>>>>>>>>>
>>>>>>>>>>>> Attached patch makes sure the loas muxer does not try to write
>>>>>>>>>>>> anything but aac and latm.
>>>>>>>>>>>>
>>>>>>>>>>>> Please comment, Carl Eugen
>>>>>>>>>>>
>>>>>>>>>>>>  latmenc.c |    4 ++++
>>>>>>>>>>>>  1 file changed, 4 insertions(+)
>>>>>>>>>>>> 2b64f3d5ecb189e77b85dbab7a6cbfe9657701f2
>>>>>>>>>>>> 0001-lavf-latmenc-Error-out-for-invalid-codecs.patch
>>>>>>>>>>>> From 9f8f39b402f77b53613a395129f96feee5e873ba Mon Sep 17
>>>>>>>>>>>> 00:00:00
>>>>>>>>>>>> 2001
>>>>>>>>>>>> From: Carl Eugen Hoyos <ceffmpeg at gmail.com>
>>>>>>>>>>>> Date: Mon, 30 Oct 2017 23:49:29 +0100
>>>>>>>>>>>> Subject: [PATCH] lavf/latmenc: Error out for invalid codecs.
>>>>>>>>>>>
>>>>>>>>>>> isnt AV_CODEC_ID_MP4ALS supported too ? (i see ALS related code
>>>>>>>>>>> in
>>>>>>>>>>> latmenc.c)
>>>>>>>>>>
>>>>>>>>>> How, when demuxer doesn't support it?
>>>>>>>>>
>>>>>>>>> Isn't ALS basically an extension of AAC? It would make sense that
>>>>>>>>> it
>>>>>>>>> can also go into LATM/LAOS.
>>>>>>>>
>>>>>>>> But do we actually supports it? There is no ALS encoder, and only
>>>>>>>> way to
>>>>>>>> trigger
>>>>>>>> that is via -c copy.
>>>>>>>
>>>>>>> And thats a problem why?
>>>>>>
>>>>>> Untested path.
>>>>>
>>>>> But that seems unrelated to this patch, no?
>>>>
>>>> Its related because you are adding codec which is not going to work.
>>>
>>> Sorry, I don't understand:
>>> Some developers claim that als-in-loas muxing will work, you doubt it.
>>> (I don't know, it's not easy to test and I don't find the spec atm.)
>>> Now you are saying that if I prevent anything but aac, latm and als
>>> from being muxed into loas, I am adding non-working code?
>>>
>>> Please elaborate, Carl Eugen
>>
>> I guess he's saying we'd be muxing files we wouldn't be able to demux,
>> which is a bad practice as we can't really test them.
>
> But how is this related to this patch?
> (Apart from the fact that three developers disagree.)
>
>> Sort of like how we create animated Webp files we can't demux/decode.
>
> You are suggesting we should not create webp files?

Not creating broken webp files, see difference.

>
> Iirc, the practice used to be to first implement encoders / muxers,
> then decoders / demuxers...

No, it is other way around.