Re: multichannel mixing on iphone with multiple mixers
Re: multichannel mixing on iphone with multiple mixers
- Subject: Re: multichannel mixing on iphone with multiple mixers
- From: Seth Delackner <email@hidden>
- Date: Thu, 29 Oct 2009 16:29:03 +0900
I didn't feel like writing an 8.24 fixed point math function for sin,
mul, etc so I modified my code to just deal with interleaved data.
Out of curiosity I ran shark against the app running on hardware, and
I see it spending fully 7% of its time in
"SRC_convert_table_i16_scalar_stereo", which is the tail end of the
call tree: AudioConverter::FillBuffer ->
AudioConverterChain::RenderOutput ->
BufferedAudioConverter::FillBuffer -> SRCWrapper::RenderOutput
That's a sizeable amount of time for what is supposedly a "canonical"
output format. Does this mean that the 8.24 format is the only
*truly* canonical (no conversion overhead) format?
On Wed, Oct 28, 2009 at 12:37 AM, Doug Wyatt <email@hidden> wrote:
> The iPhoneOS mixer should accept interleaved 16-bit or non-interleaved 8.24
> as output formats.
>
> Doug
>
> On Oct 27, 2009, at 0:42 , Seth Delackner wrote:
>
>> I figured out how to use AudioUnitAddRenderNotify to post-process my
>> down-mixed 6 streams, but I'm surprised that the downmixed buffer is
>> interleaved. Everywhere else it seems like the graph is using
>> non-interleaved (separate buffers for each channel) data. If I try
>> setting the output stream format of the mixer to be explicitly
>> non-interleaved ( adding kAudioFormatFlagIsNonInterleaved) it
>> refuses...
>
>
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden