Vedran Miletić wrote:> Yeah, ALSA's documentation is actually quite lacking on many fronts> and is mostly scattered around various wikis and mailing lists.> > But I guess it is as it is, complaining about it won't fix it.> > On 2/9/09, Demian Martin <demianm_1@xxxxxxxxx> wrote:>> Thanks, I'll try that. I guess this e-mail will serve as more documentation>> on that feature (bug). I had not seen that factoid anywhere.>>>> Demian Martin>> PDS>>>>>>> -----Original Message----->>> From: Vedran Miletić [mailto:rivanvx@xxxxxxxxx]>>> Sent: Monday, February 09, 2009 1:39 PM>>> To: Demian Martin>>> Cc: Pavel Hofman; alsa-devel@xxxxxxxxxxxxxxxx>>> Subject: Re: Juli@ ICE1724 and 24 bit audio>>>>>> The plughw part is what makese it resample to 16-bit. You should just use>>> hw.>>>>>> On Mon, Feb 9, 2009 at 10:29 PM, Demian Martin <demianm_1@xxxxxxxxx>>>> wrote:>>>> Pavel:>>>> Thanks for all your help and work on this project.>>>>>>>> I'm using the Juli@ card because it supports 176.4KHz. The only other>>>> candidates are much more expensive and bring their own problems.>>>>>>>> I'm using Xine set for dolby/dts passthrough which seems to send the>>> audio>>>> data at its native rate. The resampling is bypassed (Dolby Digital would>>> be>>>> trashed otherwise).>>>>>>>> I will try playing the files from the command line when I am next in>>> front>>>> of the system, tomorrow.>>>>>>>> This is the Asound.conf:>>>> Asound.conf (/etc)>>>>>>>> pcm.asym_spdif {>>>>>>>> type asym>>>>>>>> playback.pcm "plughw:0,1">>>>>>>> capture.pcm "plughw:0">>>>>>>> }>>>>>>>>>>>> pcm.!default asym_spdif>>>>>>>> Demian Martin>>>> PDS>>>>>>>>> -----Original Message----->>>>> From: Pavel Hofman [mailto:pavel.hofman@xxxxxxxxx]>>>>> Sent: Monday, February 09, 2009 12:27 PM>>>>> To: Demian Martin>>>>> Cc: alsa-devel@xxxxxxxxxxxxxxxx; pavel.hofman@xxxxxxxxx>>>>> Subject: Re: [alsa-devel] Juli@ ICE1724 and 24 bit audio>>>>>>>>>> Demian Martin napsal(a):>>>>>> I have the driver working fine for sample rates from 44.1K to 196K>>>>> however>>>>>> it seems to be truncating the data to 16 bits. I wasn't sure until I>>>>> started>>>>>> testing with the HRx 176.4K 24 bit files that have an embedded HDCD>>> flag>>>>> in>>>>>> the LSB. The files work OK on Windoze systems (with a lot of low>>> level>>>>>> settings tweaked) and they play fine but the flag isn't detected by a>>>>> system>>>>>> that can detect them. Further looking at the data stream with a scope>>> it>>>>>> seems the last 8 bits aren't changing. Is there anything I can do to>>>>> control>>>>>> the driver to confirm this problem or change the playback settings to>>>>> make>>>>>> it work?>>>>>>>>>>>>>>>>> How do you play the files? Do you use the hw or plug:hw device? IIRC,>>>>> the standard-setup dmix resampler is 16-bit only.>>>>>>>>>> I did tests with SPDIF OUT/IN and found it bit-perfect, for 24bit too.>>>>>>>>>> Regular ICE1724 cards do not output 176.4kHz SPDIF, only 88.2kHz. But>>>>> Juli is not affected by that bug.>>>>>>>>>> Regards,>>>>>>>>>>>>>>> Pavel.>>>> _______________________________________________>>>> Alsa-devel mailing list>>>> Alsa-devel@xxxxxxxxxxxxxxxx>>>> http://mailman.alsa-project.org/mailman/listinfo/alsa-devel>>>>>>>>>>>>> -->>> Vedran Miletić>> What makes you guys think the plug plugin trims the data down to 16bits? Sure if the rate is changed (the source code suggests only the low-quality linear rate algorithm supports above-16bit format - operation convert, the other resampling algorithms convert using the operation convert_s16). Ice1724 supports all the general rates natively, the rate conversion in the plug plugin should not kick in for common audio formats. The only case I can think of would be switching Juli to external SPDIF-IN clock - the hw.rate_min = hw.rate_max = actual_rate and the automatic rate conversion could happen. I did not use the hw device in my tests since all the tested tracks would have to be 32-bit for the hw device to accept the format when played via my favorite aplay. My 2 cents guess is xine does the conversion. Regards, Pavel. _______________________________________________Alsa-devel mailing listAlsa-devel@xxxxxxxxxxxxxxxxxxxx://mailman.alsa-project.org/mailman/listinfo/alsa-devel
