Re: Grabbing input from CoreAudio driver
Re: Grabbing input from CoreAudio driver
- Subject: Re: Grabbing input from CoreAudio driver
- From: Jeff Moore <email@hidden>
- Date: Wed, 27 Apr 2005 11:44:31 -0700
On Apr 27, 2005, at 7:41 AM, Tommy Schell wrote:
Hi,
I'm developing a CoreAudio (kernel space) driver.
When input begins, does the HAL, via convertInputSamples(), ignore
the initial audio data if it is all zero's?
No.
Also I notice that it appears the HAL doesn't typically start
exactly on the 1st sample in my input buffer, but
somewhere around the 50th sample, when input begins.
So I guess there is an assumption that initial data will not be valid?
There are no assumptions. The HAL starts as soon as it can. However,
this is typically not at the beginning of the ring buffer. The reason
why is that by the time the has bootstrapped itself and the time
stamps and notifications have been received, time has elapsed. Those
samples are outside of the range that the first IO cycle reads for.
This is no different than if a second process started up after the
hardware had been started. The second process starts reading from the
ring buffer from where ever the current time dictates it should.
This is why the architecture time stamps everything and provides
description of the latency. Clients are supposed to use this
information when they are doing synchronization.
--
Jeff Moore
Core Audio
Apple
_______________________________________________
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