Re: [Jack-Devel] Can't bring F/P below 128

PrevNext  Index
DateMon, 19 Dec 2011 08:31:47 -0500
From Paul Davis <[hidden] at linuxaudiosystems dot com>
ToNikos Chantziaras <[hidden] at arcor dot de>
Cc[hidden] at jackaudio dot org
In-Reply-ToNikos Chantziaras Re: [Jack-Devel] Can't bring F/P below 128
Follow-UpNikos Chantziaras Re: [Jack-Devel] Can't bring F/P below 128
On Mon, Dec 19, 2011 at 1:41 AM, Nikos Chantziaras <[hidden]> wrote:
> On 12/19/2011 03:29 AM, Paul Davis wrote:
>>
>> On 12/18/11, Nikos Chantziaras<[hidden]>  wrote:
>>  some of what it does doesn't necessarily
>>>
>>> work in a beneficial way for your average Linux Multimedia Desktop.
>>
>>
>> jackd doesn't have much to do with "your average Linux multimedia
>> desktop". neither does f/p below 128.
>
>
> I disagree.  A multimedia desktop should be able to achieve this.  For
> example people using FL-Studio (former Fruity Loops) on Windows expect to be
> able to do that (with something like ASIO4all if their card doesn't have
> proper ASIO drivers.)

1) "desktop multimedia" != "specialized music creation applications"
2) there are plenty of desktop multimedia applications on *every*
platform that are entirely happy with hundreds of msecs of latency,
and so are their users. do you actually think that vlc, itunes,
mplayer etc. need latency in the 10's of msec range?
3) there are plenty of reports on the forums of windows music creation
software about issues getting down to 128 f/p on various different
windows systems, for precisely the same set of reasons that exist on
linux (but *not* on OS X because of apple's tight control over the
hardware.

> People expect to be able to play notes on their MIDI keyboard without audio
> delays on their non-specialized computers.

people expect a lot of different things. that doesn't make their
expectations correct.

> Also, as I pointed out in an older post, I am able to use 64 f/p just fine
> with another sound card and *without* an RT kernel.

sure. look, the standard tail i tell about this is my own switch from
a dual opteron system to a 6 core 3.2Ghz Phenom II system. lots more
power, newer motherboard, same (RME) audio interface. the old system
could do 64 frames/period without an RT kernel with very few xruns;
the new system can't do 2048 frames/period without xruns (non-RT
kernel). one day, i'll investigate why. until i do, it just appears
completely ludicrous - 6 3.2Ghz cores available, and the stupid kernel
can't find one to keep a once-every-hundred msec requirement properly
serviced :(

that said, people (yourself included) seem to have this somewhat
justifiable belief that one intel-ish computer is just like any other,
even when it comes time to do really rather specialized things like
"respond to any external signal within 10msec". unfortunately, the
truth is that between the set of possibilities presented by:

    * MSI's from the motherboard
    * ACPI implementation
    * frequency scaling
    * PCI bridge chipsets
    * video interface chipset and driver behaviour
    * USB controller chipset, wiring and driver behaviour
    * USB device behaviour

(and a few more) ... there's ample possibilities for two systems to
have wildly different abilities to do low latency stuff.

there's a very good reason apple retains such tight control over their
h/w, and part of it is precisely to eliminate variability in the sorts
of things i've just listed.

> I believe the RT people ought to explain this then in:
>
> https://rt.wiki.kernel.org/articles/f/r/e/Frequently_Asked_Questions_7407.html
>
> The word "audio" (or "sound") is not even mentioned *once* on that page :-/

why should they care? why do you think that our little world of audio
is of any actual significance at all? the overwhelming use of audio on
linux these days is the stack within android, not the tiny number of
us doing music creation on desktop systems. the predominant use of the
RT kernel is for embedded systems, where there is real revenue from
companies and end-users, rather than the pro-audio and music niche
where there's almost no revenue for anyone (including companies that
are based on windows and os x).

--p
PrevNext  Index

1324301520.10925_0.ltw:2,a <CAFa_cKkzgf2u=3JRX2sWX8LerJ02j3Psno2YjZ62P3RVh8m=HQ at mail dot gmail dot com>