After some test there are some issue:
ramdom crash at start up or without sound.
But the real probleme i think is with 5.0 sound.
In the sample the 2.0 audio out is good
but the 5.0 has low volume and metallic sound
just try xine
fra piste 5.0
ffmpeg_audio_dec: unknown header with buf type 0x3000000
qaa 2.0
is good
ffmpeg -i 00001.ts
...
Input #0, mpegts, from '00001.ts':
Duration: 00:02:10.51, start: 10461.634989, bitrate: 6905 kb/s
Program 132
Stream #0.0[0x78]: Video: h264, yuv420p, 1440x1080 [PAR 4:3 DAR
16:9], 50 fps, 50 tbr, 90k tbn, 50 tbc
Stream #0.1[0x82](fra): Audio: eac3, 48000 Hz, 5.0, s16, 256 kb/s
Stream #0.2[0x83](qaa): Audio: eac3, 48000 Hz, stereo, s16, 128 kb/s
Stream #0.3[0x8c](fra): Subtitle: dvbsub
Stream #0.4[0x8d](fra): Subtitle: dvbsub
Le 25/03/2010 16:08, Jose Alberto Reguero a écrit :
> Patch for xine-lib that don't need to patch remux.c to work.
>
> Jose Alberto
>
> El Jueves, 25 de Marzo de 2010, Jose Alberto Reguero escribió:
>
>> Patch against latest xine-lib-1.2. New patch for xineliboutput. Now must
>> work if you patch remux.c.
>> You must have the latest xine-lib and xineliboutput. Yesterday both have
>> changes about eac3.
>>
>> Jose Alberto
>>
>> El Jueves, 25 de Marzo de 2010, zaverel escribió:
>>
>>> The typo was on remux.c => lost : and ;
>>>
>>> Anyway xine-lib has been updated and your patch don't apply.
>>>
>>> Update is for eac3 with mkv
>>>
>>> i don't test it yet
>>>
>>>
>>>
>>> corrected remux.c
>>>
>>> line 533
>>> case SI::AC3DescriptorTag:
>>> +case SI::EnhancedAC3DescriptorTag:
>>>
>>>
>>> and in line 191:
>>>
>>> -Target[i++] = SI::AC3DescriptorTag;
>>> +Target[i++] = SI::EnhancedAC3DescriptorTag;
>>>
>>> Le 24/03/2010 22:58, Jose Alberto Reguero a écrit :
>>>
>>>> I attached a second version of the first patch.
>>>> I make the same changes that in the second patch, but maintaining the
>>>> logic of the first patch. Also I commented the line:
>>>> +// this->context->request_channels = 2;
>>>> because your example has 5 channels. If you have trouble with that you
>>>> can comment the line again.
>>>> Which typo error has the second patch?
>>>>
>>>> Jose Alberto
>>>>
>>>> El Miércoles, 24 de Marzo de 2010, zaverel escribió:
>>>>
>>>>> i 've patched pat.c and now remux.c
>>>>> and with use xine-lib-1.2 with your second patch (who has typo error)
>>>>> and that doesn't work.
>>>>>
>>>>> corrected remux.c
>>>>>
>>>>> line 533
>>>>> case SI::AC3DescriptorTag:
>>>>> +case SI::EnhancedAC3DescriptorTag:
>>>>>
>>>>>
>>>>> and in line 191:
>>>>>
>>>>> -Target[i++] = SI::AC3DescriptorTag;
>>>>> +Target[i++] = SI::EnhancedAC3DescriptorTag;
>>>>>
>>>>>
>>>>>
>>>>> with your previously patch and just pat.c patched with
>>>>> line 402
>>>>>
>>>>> case SI::AC3DescriptorTag:
>>>>> +case SI:EnhancedAC3DescriptorTag:
>>>>>
>>>>> that worked but not stable.
>>>>>
>>>>> Is your sample eac3 has |"Spectral extension" ?
>>>>> because in france dvb-t with eac3 has it
>>>>> and need a ffmpeg patched for that.
>>>>> And i test with that sound.
>>>>>
>>>>> Le 24/03/2010 19:41, Jose Alberto Reguero a écrit :
>>>>>
>>>>>> It works here with a old sample of tdt with eac3. Have you patch also
>>>>>> remux.c? You need to change in line 533:
>>>>>>
>>>>>> case SI::AC3DescriptorTag:
>>>>>> +case SI:EnhancedAC3DescriptorTag:
>>>>>>
>>>>>> and in line 191:
>>>>>>
>>>>>> -Target[i++] = SI::AC3DescriptorTag;
>>>>>> +Target[i++] = SI::EnhancedAC3DescriptorTag
>>>>>>
>>>>>> Jose Alberto
>>>>>>
>>>>>> El Miércoles, 24 de Marzo de 2010, dplu escribió:
>>>>>>
>>>>>>> Hi
>>>>>>>
>>>>>>> I have made previous test with the version release by Petri
>>>>>>> Hintukainen
>>>>>>>
>>>>>>> And I notice this part is not working like it should
>>>>>>>
>>>>>>> - if((m->descriptor_tag == STREAM_AUDIO_AC3) || /* ac3 - raw
>>>>>>> */ + if(m->descriptor_tag == HDMV_AUDIO_84_EAC3) {
>>>>>>> + m->content = p;
>>>>>>> + m->size = packet_len;
>>>>>>> + m->type |= BUF_AUDIO_EAC3;
>>>>>>> + return 1;
>>>>>>> +
>>>>>>> + } else if((m->descriptor_tag == STREAM_AUDIO_AC3) || /* ac3
>>>>>>> - raw */
>>>>>>>
>>>>>>> unfortunaletly, in AC3 or E-AC3 , the descriptor tag is
>>>>>>> STREAM_AUDIO_AC3, so the program never run the first if (installed a
>>>>>>> debug printf here)
>>>>>>>
>>>>>>> It seems that your first approach (at least what I understood) by
>>>>>>> forcing the decoding of all AC-3 stream by ffmepg instead of
>>>>>>> internal lib was nice but generate violent segfault on libavcodec
>>>>>>>
>>>>>>> Hope this help you
>>>>>>>
>>>>>>> Best regards
>>>>>>>
>>>>>>> PS : Sorry to pollute the vdr mailing list (not subscribed to ffmpeg
>>>>>>> or xine-lib....)
>>>>>>>
>>>>>>> Le Wednesday 24 March 2010 18:41:14 zaverel, vous avez écrit :
>>>>>>>
>>>>>>>> hello
>>>>>>>>
>>>>>>>> your second patch doesn't work : no sound on eac3 channel
>>>>>>>>
>>>>>>>> if you want there is a sample of vdr hd-eac3 recording here:
>>>>>>>>
>>>>>>>> http://dl.free.fr/ew4rJddM8
>>>>>>>>
>>>>>>>> 103mo
>>>>>>>>
>>>>>>>> else , i don't know what mailing-list is the more indicate for
>>>>>>>> debuging
>>>>>>>>
>>>>>>>> the vdr or xine-dev mailing-list ?
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> vdr mailing list
>>>>>>>> vdr@linuxtv.org
>>>>>>>> http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
>>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> vdr mailing list
>>>>>>> vdr@linuxtv.org
>>>>>>> http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
>>>>>>>
>>>>>> _______________________________________________
>>>>>> vdr mailing list
>>>>>> vdr@linuxtv.org
>>>>>> http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
>>>>>>
>>>> _______________________________________________
>>>> vdr mailing list
>>>> vdr@linuxtv.org
>>>> http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
>>>>
>
>
>
> _______________________________________________
> vdr mailing list
> vdr@linuxtv.org
> http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
>
_______________________________________________
vdr mailing list
vdr@linuxtv.org
http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
