On Sun, 2009-10-18 at 10:22 +0000, babsyco babsyco wrote:

Thanks guys, but the latency is still around 20ms when I don't use the soundcard (I just tried it with the built in mic-that's the same as jack for all intensive purposes, right?). Any other suggestions as to how I can get it lower? I really wanna use it for live performance.

well, we still don't know about your OS. are you really on OS X? is this problem directly related to pd or do you have the same latency with other software as well? if it's a general issue, i would ask in other - probably more dedicated - channels as well. it's an annoying proposal to make, but if your only goal really is only getting the audio latency low, i'd suggest considering another OS as well. it's a very general advice, i know, but probably you can be a bit more specific about what your're trying to do. there are many reasons not to change the OS frivolously, but if that is the main goal, i'd probably change to one that gives you enough control to change settings as you need them.

if there are, try different drivers (or use jack, as others already suggested, assuming your not on windows). then try to give the pd audio process higher priority.

roman

---

From: babsyco@hotmail.com To: pd-list@iem.at Date: Fri, 16 Oct 2009 05:22:50 +0000 Subject: [PD] latency issue

Hi guys-I just have another fairly simple question. I'm running PD on a macbook pro with a 2.26 Intel Core 2 duo processor, 2 gig of memory 1067 MHz DDR3, and I'm using a terratec Phase 24FW external sound card. On my other computer (which was far inferior) using this soundcard with Cubase I had a latency of around 6ms (I couldn't get PD to run on it so I have no basis for comparison there), but running PD on this new setup I can't get the latency down below 20ms with it turning to mush. What do I need to change to get the latency down below 10ms-the soundcard, more ram, or is the processor just too slow?

Thanks so much. Babsyco.

---

Let us find your next place for you! Need a place to rent, buy or share?

---

Click Here View photos of singles in your area _______________________________________________ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list

Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de

babsyco babsyco wrote:

Thanks guys, but the latency is still around 20ms when I don't use the soundcard (I just tried it with the built in mic-that's the same as jack for all intensive purposes, right?). Any other suggestions as to how I can get it lower? I really wanna use it for live performance.

No. That is not the same as using it with jack, at least it would not be on a Linux system, I could be wrong about mac, but from what I know nothing compares to using jack, performance wise. Another issue could be externals, try measuring the performance without any externals loaded and see if one of the ones you are using is causing latency.

Also, many commercial audio apps misreport latency; for example a large number, if not most of them, report only the latency introduced by the app itself, which is in their case actually cumulative with the latency introduced with the OS drivers and the latency inherant in the DAC / ADC of the card itself. With jack, at least, you are getting a much more accurate picture of the latency.

This is based on reports of users who have measured their latency and compared it with latencies reported by jack and by mainstream nonfree software; there are actually applications to help you measure the latency you are getting by looping back a coded signal between the DAC/ADC and measuring the phase difference, this could actually be coded in PD and would be an interesting exercise come to think of it.

On Oct 18, 2009, at 8:47 PM, Justin Glenn Smith wrote:

Bjørn Nielsen wrote: ...

...

Fast Track Pro soundcards. (I must also add that I have no latency in any other music software I use: pro tools, digital performer, audacity, garageband etc.)

I find this very hard to believe, is this just because you are
auditioning your realtime input directly through the sound card? Because no
software can process audio input and output a modified version without
latency. Try playing through a plugin in realtime in protools or dp, I suspect
you will notice a latency. If not, maybe you should be translating your pd
patches into audio unit plugins, but I strongly suspect you will notice a
latency.

There is no audio software without latency. The lowest that is
feasible on Mac OS X is probably 4ms. Once you factor in distance
and speed of sound, then you might see that 20ms of latency isn't
really all that much.

20ms * 343m/s = 7.546m

So if you are 7.5m away from one of your band mates, there will 20ms
of latency just from the sound travelling thru the air.

.hc

...

...

@Justin Glenn Smith: Do you have links or names of some of that measuring software you referered to? That, I really like to try out.

jack_delay, AKA jdelay: author's homepage is http://www.kokkinizita.net/linuxaudio/

---

Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list

I hate it when they say, "He gave his life for his country." Nobody
gives their life for anything. We steal the lives of these kids. - Admiral Gene LeRocque

Hans-Christoph Steiner wrote: ...

Also, I forgot to mention, there was a thorough discussion on latency a few years ago:

http://lists.puredata.info/pipermail/pd-ot/2006-12/001629.html

Of particular interest from that thread is a simple way to measure latency: http://lists.puredata.info/pipermail/pd-ot/2006-12/001642.html

There is no audio software without latency.  The lowest that is feasible on Mac OS X is probably 4ms.   Once you factor in distance and speed of sound, then you might see that 20ms of latency isn't really all that much.

Yeah, sorry that I said there was no latency, I do know that every software needs to process. What I meant to say was that the latency while recording was so low in fx Pro Tools, that I don't hear it. That is not the case with PD.

20ms * 343m/s = 7.546m

So if you are 7.5m away from one of your band mates, there will 20ms of latency just from the sound travelling thru the air.

This is of couse a correct calculation, but never the reality. I never stand more than maximum a couple of meters from my own amplifier, or else I will lose the timing if I hear my own sound with latency. And if you play at stages so large, that you are 7-8 meters from the drummer or other musicians, then you will almost always have their sound in your personal monitors, just a few meters from you. I actually had a problem some few times, when I tried to play together with the band, while making sound from the mixingdesk 10 meters from the stage. That is not recommendable (though you can do it with headphones).

So my point is here, that a realtime latency (including processing, ad/da conversion and traveling time thru air) above ca. 15-20ms is a real problem for the timing while playing together with a band. It can seems like in some of the comments on this list, that is neglected that the ultralow latency is making a difference and that you are on the wrong track if you wan't to use PD for that purpose. But I see very big potential and fun in doing this. I have skipped must of my analog effects, and instead use some real messy and funny delayeffects in pd.

I would still find it very interesting if anybody on list knows if it you can lower pd latency with better hardware to the unhearable or if 11ms (in the pd settings) is the lower barrier?

And this is possible because ProTools only allow you the use of very specific hardware for which they have created closed, proprietary drivers. Expecting *any* other app to behave in a similar way to ProTools is very unrealistic unless the same kind of software/hardware lock-in relationship exists.

D.

chris clepper wrote:

On Mon, Oct 19, 2009 at 7:47 AM, Bjørn Nielsen <bjoarn@gmail.com mailto:bjoarn@gmail.com> wrote:

Yeah, sorry that I said there was no latency, I do know that every
software needs to process. What I meant to say was that the latency
while recording was so low in fx Pro Tools, that I don't hear it. That
is not the case with PD.

You don't specify which type of PT, but Pro Tools HD has only 5 samples (not ms) of latency. The LE systems have very high latency compared to what Logic and other apps can do though.

...

Yeah, sorry that I said there was no latency, I do know that every software needs to process. What I meant to say was that the latency while recording was so low in fx Pro Tools, that I don't hear it. That is not the case with PD.

You don't specify which type of PT, but Pro Tools HD has only 5 samples (not ms) of latency.  The LE systems have very high latency compared to what Logic and other apps can do though.

I use both Pro Tools LE and M-powered. You probably right, I have never made a comparison. I really didn't meant the discussion to be about latency in pro tools or any other software besides PD (though spinoffs of couse also can be very interesting). It was just that Roman Haefeli asked whether Babsycos experienced a similar latency in other software, and thereby indirectly saying that it could simply be a hardware setup problem and that PD not nessecarily has more latency than other software.

In my experience (and don't claim to be that experienced at all) I've been troubled with the latency in PD and not in any other software. But I would be very pleased if I'm wrong in the underlying assumption here and that is just a question of the right hardware/software setup, to decrease the PD latency to the unhearable.

So Babsycos original question still goes unanswered (besides the really good advice about jack): can you decrease the latency with improving you hardware?

Take 3.25m and 10ms, then its more scale, I was just going with the
20ms number. Then you can look at the latency in many physical,
traditional instruments like the violin and there is probably more
than 20ms between action and sound.

.hc

On Oct 20, 2009, at 4:12 AM, hard off wrote:

...

...

So if you are 7.5m away from one of your band mates, there will 20ms
of latency just from the sound travelling thru the air.<<

if you are big enough to be playing on a stage that has 7.5m
distance between band mates, then you don't need to worry anyway, as
someone else is paid to sort out your technical issues while you do
lines of coke with groupies backstage.

---

Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list

Looking at things from a more basic level, you can come up with a more
direct solution... It may sound small in theory, but it in practice,
it can change entire economies. - Amy Smith

On Mon, 19 Oct 2009, Bjørn Nielsen wrote:

Yeah, sorry that I said there was no latency, I do know that every software needs to process. What I meant to say was that the latency while recording was so low in fx Pro Tools, that I don't hear it. That is not the case with PD.

Btw, in the audio preferences, what do you choose as "Delay (msec)" ? How low can you get this to be?

(This is also known as -audiobuf.)

Probably this is a really silly question and is the first setting you played with, but I don't know.

_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard, Montréal, Québec. téléphone: +1.514.383.3801

On Tue, 20 Oct 2009, Bjørn Nielsen wrote:

The lowest delay in ms I have reached before the sound gets quicky is 11ms, but it depends on whether I use Jackpilot and what audiointerface I use (the macbook internal soundcard can't do 11ms, but my mbox2 can). It was my impression that this delay is all the latency PD have,

I don't know, I'd risk saying that the delay in question is simply added to the base delay of the in/out system, but I wouldn't risk guaranteeing it without rereading all of the s_audio*.c code again... I forgot a fair bit of it.

In theory, the soundcard needs to fill a block before it can pass it to pd, then pd needs to process a whole block before a block of output is ready, and then the time it takes to play a block is a block duration, but if you look at the minimum time difference between in and out, you don't need to count the normal duration of the output block in there. So you only count blocksize once, and then, if the processing time is really smooth, then your minimum audiobuf size is the blocksize delay times the % of CPU that pd is using.

So if your patch takes 75% CPU and has 64-byte blocks at 48000 kHz, then you compute (1+75/100)*64/48000 = 0.002333... s = 2.333... ms; and you could set pd's audiobuf to (75/100)*64/48000 = 0.001 s = 1 ms exactly.

Obviously, this is not what you are getting, because the processing isn't that much smooth. This may depend on the speed of the multitasker in the OS itself, the quality of the realtime priority system, and various little hiccups that happen in Pd itself. Many Pd internals and externals make it hard, by expecting a file-open to take 0 ms of logical time, thus if it really takes 50 ms, then you need at least 50 ms - 0 ms = 50 ms of audiobuf on top of the minimum required audiobuf.

but maybe I'm wrong? Is it even possible to measure where in the system the latency occur?

The best you can do is make a laundry list of things be checking for, but Pd can't possibly have all the information.

Btw, can you get doc/7.stuff/tools/latency.pd to work? and how would one interpret that figure?

_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard, Montréal, Québec. téléphone: +1.514.383.3801