Re: phantom MultibandCompressor
Re: phantom MultibandCompressor
- Subject: Re: phantom MultibandCompressor
- From: William Stewart <email@hidden>
- Date: Wed, 5 May 2004 17:21:26 -0700
It maybe that you aren't feeding silence into the mix? The Matrix Mixer
does honour the silence flag, and doesn't do any work with a buffer
that's marked as silence... Are you doing a rate conversion anywhere?
Bill
On 05/05/2004, at 6:17 AM, Robert Grant wrote:
>
Well Rax when it's idling is taking ~ 6% of my CPU and that's with no
>
UI update timer active, no MIDI input and no AUs in the graph other
>
than Apple Mixers and Output units.
>
>
The Render thread is taking 5.1 % of the total CPU time and
>
AUMultibandCompressorEntry (or whatever it is) is taking 70.4% of that
>
5.1 %.
>
>
Perhaps that's the expected amount of overhead? I'm really just
>
wanting a sanity check as I can't do much about the graph composition.
>
>
Thanks,
>
>
Robert.
>
>
On May 4, 2004, at 6:19 PM, Doug Wyatt wrote:
>
>
> On May 4, 2004, at 15:02, Robert Grant wrote:
>
>> Ahh,
>
>>
>
>> But it's a bit worrisome that Shark is telling me that 67% of my
>
>> audio render thread time is being spent in
>
>> AUMultibandCompressorEntry. Any suggestions as to what might be
>
>> happening? I thought the mixers were highly optimized. This is a
>
>> graph with only Apple AUs in it which are either OutputAUs or
>
>> Mixers!
>
>>
>
>> Robert.
>
>>
>
>> On May 4, 2004, at 5:45 PM, Doug Wyatt wrote:
>
>>
>
>>> On May 4, 2004, at 14:27, Robert Grant wrote:
>
>>>> I'm getting an AUMultibandCompressorEntry showing up in my Shark
>
>>>> stack trace but I haven't added one to my AU graph. And actually
>
>>>> I'm getting a lot more SystemOutputAUEntry's than I have Output
>
>>>> AUs (HAL or Generic) in my graph.
>
>>>>
>
>>>> Are there extra AUs being utilized under the covers by other AUs?
>
>>>
>
>>> No. Shark only sees a small handful of symbols from
>
>>> CoreAudio.component, which contains all the system AU's. There are
>
>>> so few exported symbols that the one nearest to any given piece of
>
>>> code is more likely to be misleading than informative.
>
>>>
>
>>> Doug
>
>
>
> OK, Shark's telling you that 67% of your render thread time being
>
> spent somewhere in our components, and that they're doing roughly
>
> twice as much work as your code. But how much work is that? How much
>
> CPU is being consumed on the whole by the render thread? If it's
>
> taking 1% of the CPU, there's no big deal. If the render thread is
>
> consuming 80% of the CPU, *then* that 67% is worrisome.
>
>
>
> --
>
> Doug Wyatt
>
> Core Audio, Apple
>
> _______________________________________________
>
> coreaudio-api mailing list | email@hidden
>
> Help/Unsubscribe/Archives:
>
> http://www.lists.apple.com/mailman/listinfo/coreaudio-api
>
> Do not post admin requests to the list. They will be ignored.
>
_______________________________________________
>
coreaudio-api mailing list | email@hidden
>
Help/Unsubscribe/Archives:
>
http://www.lists.apple.com/mailman/listinfo/coreaudio-api
>
Do not post admin requests to the list. They will be ignored.
>
>
--
mailto:email@hidden
tel: +1 408 974 4056
________________________________________________________________________
__
Culture Ship Names:
Ravished By The Sheer Implausibility Of That Last Statement
I said, I've Got A Big Stick [OU]
Inappropiate Response [OU]
Far Over The Borders Of Insanity And Still Accelerating [Eccentric]
________________________________________________________________________
__
_______________________________________________
coreaudio-api mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/coreaudio-api
Do not post admin requests to the list. They will be ignored.