measuring audio latency?

Reece Dunn msclrhd at googlemail.com
Wed Mar 10 02:25:18 CST 2010


On 9 March 2010 23:48, Ben Klein <shacklein at gmail.com> wrote:
> On 10 March 2010 10:01, Avery Pennarun <apenwarr at gmail.com> wrote:
>> On Tue, Mar 9, 2010 at 9:15 AM, Dan Kegel <dank at kegel.com> wrote:
>>> On Tue, Mar 9, 2010 at 2:58 AM, Roderick Colenbrander
>>> <thunderbird2k at gmail.com> wrote:
>>>> I might be able to measure it using my oscilloscope. Somehow I would
>>>> need to play lets say the left channel 'without' latency and the other
>>>> channel with and compare the two signals.
>>>
>>> Yes, absolutely, but it'd be good to make this measurement easy
>>> to repeat by anybody interested.  To do that, let's just loop the
>>> audio output back into the audio input.  The user will have to
>>> provide a loopback cable (or, worst case, put his mike right up
>>> to the speaker, and allow for a tiny bit of extra latency
>>> from that).
>>
>> If sound travels at 340m/s, then a one-second sample is 340m long.  A
>> 1ms delay would therefore be 340mm, or 34 cm.  Your mic would have to
>> be *quite* far away from the speakers to have a significant impact on
>> the delay, unless I'm missing something.
>
> You are: air pressure ;) hehe. Yeah, I know, it won't make a
> noticeable difference either; just being Devil's Advocate.
>
>> Which is good news, I guess, since it means tests are easier.  Hope it
>> goes well :)

But if all you are interested in is the relative latency changes, then
any additional latency effects should not matter as long as that
latency added is constant.

What is being checked is that the latency of wine+OpenAL is not
noticeably greater than the current wine+ALSA implementation (that is,
OpenAL has a latency that is at worst equal to ALSA as used by wine to
within a certain tolerance for error).

- Reece



More information about the wine-devel mailing list