[alsa-devel] [DAPM question] DAPM support for single input multiple outputs
Hi Mark,
I have one question and need your suggestion. Thanks in advance.
In our DSP, one stream path is like this: There is one bit controlling whether a signal A goes to A_out or B_out. That's the reverse case of DAPM_MUX(multiple inputs, single output).
One way is to use two switches but it seems not good since two switches control the same bit. It's strange. Do you have any opinion about such requirement support?
On 03/09/2015 09:38 AM, Qiao Zhou wrote:
Hi Mark,
I have one question and need your suggestion. Thanks in advance.
In our DSP, one stream path is like this: There is one bit controlling whether a signal A goes to A_out or B_out. That's the reverse case of DAPM_MUX(multiple inputs, single output).
One way is to use two switches but it seems not good since two switches control the same bit. It's strange. Do you have any opinion about such requirement support?
There is currently no support for a demux. But it shouldn't be too hard to add one in the DAPM core, if you want to take a try at it.
A demux would behave very similar to a mux, except that you'd reverse sources and sinks. The only thing that is a bit tricky is that we only have one control element per path and currently it always belongs to the sink. For a demux it needs to be long to the source. This means you can't have a mixer directly connected to the same path as a demux and the core needs to make sure that that doesn't happen.
- Lars
On 03/09/2015 04:53 PM, Lars-Peter Clausen wrote:
On 03/09/2015 09:38 AM, Qiao Zhou wrote:
Hi Mark,
I have one question and need your suggestion. Thanks in advance.
In our DSP, one stream path is like this: There is one bit controlling whether a signal A goes to A_out or B_out. That's the reverse case of DAPM_MUX(multiple inputs, single output).
One way is to use two switches but it seems not good since two switches control the same bit. It's strange. Do you have any opinion about such requirement support?
There is currently no support for a demux. But it shouldn't be too hard to add one in the DAPM core, if you want to take a try at it.
A demux would behave very similar to a mux, except that you'd reverse sources and sinks. The only thing that is a bit tricky is that we only have one control element per path and currently it always belongs to the sink. For a demux it needs to be long to the source. This means you can't have a mixer directly connected to the same path as a demux and the core needs to make sure that that doesn't happen.
I'll have a try to specially handle demux type. Thanks for the suggestions. BTW, is there plan to add demux support?
- Lars
On Mon, Mar 09, 2015 at 07:16:18PM +0800, Qiao Zhou wrote:
I'll have a try to specially handle demux type. Thanks for the suggestions. BTW, is there plan to add demux support?
Not really - it's not that there's any reason not to do it, it's just that nobody is doing the work as far as I know. It's quite unusual to see it in hardware (usually you see mixers in the places where you might see a demux) so few people have the motivation to do the work.
On 03/09/2015 08:20 PM, Mark Brown wrote:
Not really - it's not that there's any reason not to do it, it's just that nobody is doing the work as far as I know. It's quite unusual to see it in hardware (usually you see mixers in the places where you might see a demux) so few people have the motivation to do the work.
Understood and I agree that it is rare. Thanks for clarifying.
participants (3)
-
Lars-Peter Clausen -
Mark Brown -
Qiao Zhou