On Sun, Jan 13, 2013 at 2:37 PM, Ross Bencina <rossb-lists_at_audiomulch.com>wrote:
> On 14/01/2013 9:22 AM, David Barbour wrote:
>
>> I've been contemplating use of speculative evaluation as a way to hide
>> latency for changes in code. Unfortunately, I've not found any sound
>> APIs (ALSA, JACK, etc.)that would effectively support speculative
>>
> > buffering
>
> What exactly do you mean by speculative buffering? write-ahead with
> optional rollback?
>
Write-ahead with optional over-write (relative to any future time) would
enable relatively robust failure modes without blocking very low-latency
edits - i.e. where valid sounds are still played.
>
> The low latency APIs (and others too) are typically designed so that the
> point where audio is passed to the API, and/or the point where control is
> returned to the API is the commit point.
Indeed. I've become quite aware of this. :)
It wasn't an essential feature. JACK be nimble and quick enough for most
use-cases, if used right. But I've been wondering how to reach the
ultra-low latencies, and speculative buffering was a promising possibility.
Regards,
Dave
--
bringing s-words to a pen fight
Received on Sun Jan 13 2013 - 23:03:18 GMT