Re: Re: Revo 5.1

[Trevor Feeney <trev@feens.org>]

Takashi Iwai wrote:

>At Fri, 09 Dec 2005 10:31:34 -0500,
>Trevor Feeney wrote:
>  
>
>>Takashi Iwai wrote:
>>
>>    
>>
>>>At Fri, 09 Dec 2005 09:20:51 -0500,
>>>Trevor Feeney wrote:
>>> 
>>>
>>>      
>>>
>>>>Takashi Iwai wrote:
>>>>
>>>>   
>>>>
>>>>        
>>>>
>>>>>At Thu, 08 Dec 2005 21:42:48 +0100,
>>>>>David Peroutka wrote:
>>>>>
>>>>>
>>>>>     
>>>>>
>>>>>          
>>>>>
>>>>>>Takashi Iwai wrote:
>>>>>>  
>>>>>>
>>>>>>       
>>>>>>
>>>>>>            
>>>>>>
>>>>>>>At Tue, 06 Dec 2005 09:34:48 -0500,
>>>>>>>Trevor Feeney wrote:
>>>>>>>
>>>>>>>    
>>>>>>>
>>>>>>>         
>>>>>>>
>>>>>>>              
>>>>>>>
>>>>>>>>Was taking a look at the source for the Revo 5.1, noticed a few things.  
>>>>>>>>You set it up as using only 6 DACs, going by the 5.1 I figure, but when 
>>>>>>>>you consider that it also has the headpone out, it would have the same 
>>>>>>>>number of DACs as the 7.1 (8 DACs in all). 
>>>>>>>>
>>>>>>>>Takashi - that answers your question in the last thread...from what I've 
>>>>>>>>read, the 5.1 only uses one codec, and the headphones would just be a 
>>>>>>>>separate channel.  I tried adjusting the number of DACs to 8 in the 
>>>>>>>>driver, and did manage to get two
>>>>>>>>DAC mixer sliders.  I'm not really sure if that was the way to go or 
>>>>>>>>not.  I tried adjusting the asound plugin to output to 8 channels, pcm 
>>>>>>>>device surround71, but that didn't seem to work.  I wondering if it 
>>>>>>>>needs to be setup separately thant the other 6 channels so that it would 
>>>>>>>>be considered it's own PCM device.  That goes a bit beyond my realm of 
>>>>>>>>knowledge for driver development though...still trying to learn this.  
>>>>>>>>Been reading a lot of the documents and such.
>>>>>>>>      
>>>>>>>>
>>>>>>>>           
>>>>>>>>
>>>>>>>>                
>>>>>>>>
>>>>>>>[Added alsa-devel ML again]
>>>>>>>
>>>>>>>In the case of ice1724, you can open a stereo stream separately.
>>>>>>>The first PCM is assigned to the multi-channel, i.e. 2-8 channels
>>>>>>>interleaved format.  The secondary PCM is for SPDIF.  The third PCM is
>>>>>>>the "independent" streams.  This may have up to 3 substreams, each of
>>>>>>>which is a 2-channel one.  If you change num_dacs = 8, you'll be able
>>>>>>>to access plughw:0,2,3 (suppose the card# is 0) for the last two
>>>>>>>channels individually:
>>>>>>>
>>>>>>>	% aplay -Dplughw:0,1,3 foo.wav
>>>>>>>
>>>>>>>And you have to raise the corresponding DAC mixer values, of course.
>>>>>>>
>>>>>>>
>>>>>>>Takashi
>>>>>>>    
>>>>>>>
>>>>>>>         
>>>>>>>
>>>>>>>              
>>>>>>>
>>>>>>I've enabled all 8-channels of AK4358 in the driver, but I was unable to 
>>>>>>access the 4-th substream, because (as you noted) there are max 3 
>>>>>>substreams per PCM, so "plughw:0,2,3" is actually not available (the 
>>>>>>other substreams work fine).
>>>>>>  
>>>>>>
>>>>>>       
>>>>>>
>>>>>>            
>>>>>>
>>>>>My bad, I mean 'plughw:0,2,2'.
>>>>>
>>>>>But Trevor reported it doesn't work, too, and mentioned that the chip
>>>>>itself (Envy24GT) shouldn't support 4th stream...
>>>>>
>>>>>
>>>>>Takashi
>>>>>
>>>>>
>>>>>     
>>>>>
>>>>>          
>>>>>
>>>>Takashi,
>>>>
>>>>I've been poking around in the source for the driver and have a few 
>>>>questions.  Where is it in the code that the front PCM stream would 
>>>>conncect to output to the DACs associated with that, and is there a way 
>>>>to duplicate that to the extra 2 DACs?
>>>>   
>>>>
>>>>        
>>>>
>>>The front is handled with the multi-channel DMA, corresponding to
>>>snd_vt1724_playback_pro_*() callbacks.  This supports 2-8 interleaved
>>>channels.  The other 3 DMAs are only for 2-channel streams,
>>>snd_vt1724_playback_indep_*().  The 4th DMA (possibly a headphone
>>>output) is, thus, the 3rd substream of this type.
>>>
>>>There is no way to duplicate the signal for both front and another
>>>stream _on hardware_, AFAIK.  It might work with some trick, though
>>>(e.g. sharing the same ring buffer with different DMAs).
>>>
>>>The possible solution is the addition in alsa-lib to duplicate signals
>>>via multi+route plugins.
>>>
>>> 
>>>
>>>      
>>>
>>>> I know that laptop cards tend to 
>>>>have a similar setup, outputting to headphones and speakers, so I figure 
>>>>it can be done.  I'm also thinking that's the most logical solution in 
>>>>this case as well.
>>>>   
>>>>
>>>>        
>>>>
>>>The headphone output of laptops is usually the same to speaker or
>>>line-out.
>>>
>>>
>>>Takashi
>>> 
>>>
>>>      
>>>
>>Just to make sure I'm thinking properly...my assumption is that each PCM 
>>channel on the controller (in this case Envy24GT) has a DAC associated 
>>with it.  The driver also sets up the mixer controls for each of those 
>>DACs.  Now I'm also looking at your message and thinking that a DMA is 
>>associated with the DAC (as you'd have to write to the the DAC at the 
>>proper memory address to output).
>>    
>>
>
>At least, envy24HT has 4 DMAs.  The first one is for multi-channels.
>It can send from 2 to 8 channels as a single stream.
>The rest 3 DMAs are for 2-channels only, rear, clfe and side.  You
>can control them separately unless the channels conflict with the
>mutli-channel setup of the first DMA.
>
>  
>
>> I did some looking over of the AK 
>>4358 Datasheet to try and get a better grasp of what happens on that 
>>end.  Can't find a Datasheet for the Envy24GT, so I have to examine the 
>>driver to figure that out.  Where I get a bit confused is substreams.  
>>If you output to a substream, is that like outputting to a DMA/DAC, or 
>>to a PCM stream on the controller (or maybe I'm way off)? 
>>    
>>
>
>I have no datasheet for GT, too.  Only the datasheet for Envy24HT is
>available (you can find it on alsa's ftp).
>
>  
>
>>As I mentioned before,t he card does list 3 substreams, I can output to 
>>pluginhw:0,2,0 0,2,1 and 0,2,2. The first substream outputs to the 
>>rears, the second to the center/sub out and the third outputs nothing 
>>(there are no errors either).  If those substreams correpsonded directly 
>>with the pair of DACs, I would expect the third to work properly then.  
>>If they corresponded to a PCM stream through the controller, then I 
>>would epxect it to go nowhere.
>>    
>>
>
>Hmm, interesting.  Although GT seems supporting only 6 channels, the
>DMA looks as if working.  Possibly 6-channel is just for the
>multi-channel DMA?
>
>
>Takashi
>  
>
I'll take a look at the HT datasheet so I can get a better grasp how 
things work.  The GT specs on VIA's site say that it's a 6 channel 
controller.  Is there a way that one could acquire the datasheet?  I 
figure the HT can't be too far off the GT, as the card is function using 
the HT specs.   If it is 6 channels for the multi-channel portion, can 
you think of any reason there would still be no output to the last 
substream (which I'm assuming is the headphone stream)?


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click