connect_icut

I'm thinking of investing in an external sound card, in the hope that it will help my Max patches to run with greater stability.

Ideally, this would be a firewire device that would take some of the audio processing load off my CPU.

Anyone got any opinions on which product would suit my needs best?

I'm thinking of investing in an external sound card, in the hope that it will help my Max patches to run with greater stability. 

Anyone got any opinions on which product would suit my needs best?


best-external-soundcard-for-maxmsp

...but is there something similar that doesn't cost $1500?

I should point out that I don't need this for any kind of input, output functionality, just to improve the audio processing power of my Mac.

>I should point out that I don't need this for any kind of input, 

>output functionality, just to improve the audio processing power of 

i might be wrong, but i don't see how and if an external sound card 

would improve the internal audio processing of your mac....

of course if you take your audio out of it and then in (using some 

external hardware sound processing for exemple) or mix in the analog 

world, it is useful - but for striclty working IN the mac (using 

>Okay, so now we're down to $1000.
>
>I should point out that I don't need this for any kind of input, 
>output functionality, just to improve the audio processing power of 
>my Mac.

i might be wrong, but i don't see how and if an external sound card 
would improve the internal audio processing of your mac....

of course if you take your audio out of it and then in (using some 
external hardware sound processing for exemple) or mix in the analog 
world, it is useful - but for striclty working IN the mac (using 
maxmsp)... i am not sure

kasper
-- 
  Kasper T. Toeplitz
noise, composition, bass, computer
http://www.sleazeArt.com

So am I totally wrong in thinking that there are external sound cards available that that reduce the processing load on the CPU?

External sound cards route audio into and out from your computer, 

they do nothing regarding the actual DSP load of the host itself.

There are, on the other hand, DSP acceleration cards that work with 

specific plugins, for example, that offload the processing from the 

host to the external card, but that is something completely different 

- they are specific to particular families of plugins. There are also 

external VSTi players that offload the processing of VSTi's from the 

host to the external card, but again these are not the same as an 

external sound card. Neither of these work, to my knowledge, as 

general sound cards in addition to what they do.

On Jul 18, 2007, at 6:27 PM, Sam Macklin wrote:

> So am I totally wrong in thinking that there are external sound 

> cards available that that reduce the processing load on the CPU?

External sound cards route audio into and out from your computer,  
they do nothing regarding the actual DSP load of the host itself.

There are, on the other hand, DSP acceleration cards that work with  
specific plugins, for example, that offload the processing from the  
host to the external card, but that is something completely different  
- they are specific to particular families of plugins. There are also  
external VSTi players that offload the processing of VSTi's from the  
host to the external card, but again these are not the same as an  
external sound card. Neither of these work, to my knowledge, as  
general sound cards in addition to what they do.

>
> So am I totally wrong in thinking that there are external sound  
> cards available that that reduce the processing load on the CPU?

----
Steven M. Miller
Professor, Contemporary Music Program
College of Santa Fe

Home 
SFIFEM 
Atrium Sound Space 
OVOS 
CMP 

There seems to be some disagreement on this, doesn't there?

I wonder if there is a particular sound card that Cycling '74 recommends for use with Max/MSP. I'll have to ask.

For what it's worth, I think the instability I've been experience with my computer audio (especially in but not limited to Max) is caused by changing the sampling rate in the Max DSP Status window (especially while Max audio is on). Possibly, this corrupts a couple of the computer's audio library files (which are easy to replace). Anyone else experienced this?

I wonder if there is a particular sound card that Cycling '74 recommends for use with Max/MSP.  I'll have to ask.

For what it's worth, I think the instability I've been experience with my computer audio (especially in but not limited to Max) is caused by changing the sampling rate in the Max DSP Status window (especially while Max audio is on). Possibly, this corrupts a couple of the computer's audio library files (which are easy to replace).  Anyone else experienced this?


>There seems to be some disagreement on this, doesn't there?

but I nevre heard about some of the computation being done on teh 

sound card (ok you could do some DSP on the RME, but it's internat TO 

the card - mostly routing - and when i tried to do it from inside max 

it was with limited effect - no spatialisation, you would have 

>I wonder if there is a particular sound card that Cycling '74 

>recommends for use with Max/MSP. I'll have to ask.

>For what it's worth, I think the instability I've been experience 

>with my computer audio (especially in but not limited to Max) is 

>caused by changing the sampling rate in the Max DSP Status window 

>(especially while Max audio is on). Possibly, this corrupts a couple 

>of the computer's audio library files (which are easy to replace). 

before change anything in the dsp window i always set the audio OFF, 

>Hmmm...
>
>There seems to be some disagreement on this, doesn't there?

but I nevre heard about some of the computation being done on teh 
sound card (ok you could do some DSP on the RME, but it's internat TO 
the card - mostly routing - and when i tried to do it from inside max 
it was with limited effect - no spatialisation, you would have 
cliks...)

>
>I wonder if there is a particular sound card that Cycling '74 
>recommends for use with Max/MSP.  I'll have to ask.
>
>For what it's worth, I think the instability I've been experience 
>with my computer audio (especially in but not limited to Max) is 
>caused by changing the sampling rate in the Max DSP Status window 
>(especially while Max audio is on). Possibly, this corrupts a couple 
>of the computer's audio library files (which are easy to replace). 
>Anyone else experienced this?

before change anything in the dsp window i always set the audio OFF, 
the ON again

Ah ha! I fool around with the DSP status all the time while my patches are running, which may explain a thing or two.

For anyone who is an idiot like me and has corrupted their audio preferences, try deleting these files...

(probably in you computer's main library)

... then restart the computer to create new, uncorrupted files.

com.apple.audio.DeviceSettings.plist 
(probably in you computer's main library)

com.apple.audio.AudioMIDISetup.plist 
(probably in your user account library)

... then restart the computer to create new, uncorrupted files.


Giving more cycles to your CPU would probably come from lowering your sample rate and DSP block size [sample size]. I think too that, if you're using a universal audio ASIO driver you may get lower performance [such as ASIO4ALL].

Giving more cycles to your CPU would probably come from lowering your sample rate and DSP block size [sample size].  I think too that, if you're using a universal audio ASIO driver you may get lower performance [such as ASIO4ALL].


MOTU 828. You can usually find the original 828 for about $200 on ebay.

Off course, changing the DSP status may be what caused all my problems in the first place.

What's the best way to set up the computer's audio/MIDI preferences and the Max DSP status for minimum CPU load, without crashing the audio in Max or corrupting the computer's preference files?

Also, is it possible that some external sound cards use their own third-party audio drivers that deal with audio more efficiently than the Mac core audio driver - thus reducing CPU load?

By the same token, is it better to use Soundflower with Max, instead of Core Audio?

Quote: connect_icut wrote on Thu, 19 July 2007 16:16

----------------------------------------------------

> Also, is it possible that some external sound cards use their own third-party audio drivers that deal with audio more efficiently than the Mac core audio driver - thus reducing CPU load?

> By the same token, is it better to use Soundflower with Max, instead of Core Audio?

Yes, I believe that logic is correct. On windows we have a universal ASIO driver that is inferior to a third party driver. I can run many, many more plugins on an RME card than I can on a built-in Realtek card using ASIO4ALL.

Quote: connect_icut wrote on Thu, 19 July 2007 16:16
----------------------------------------------------
> Also, is it possible that some external sound cards use their own third-party audio drivers that deal with audio more efficiently than the Mac core audio driver - thus reducing CPU load?
>
> By the same token, is it better to use Soundflower with Max, instead of Core Audio?
----------------------------------------------------

Yes, I believe that logic is correct.  On windows we have a universal ASIO driver that is inferior to a third party driver.  I can run many, many more plugins on an RME card than I can on a built-in Realtek card using ASIO4ALL.


So does anyone know what the most efficient third party audio driver for the Mac is?

> So does anyone know what the most efficient third party audio driver 

All audio on mac os x must go through CoreAudio. It's an abstraction 

layer between the thirdparty soundcard drivers and audio software. If 

it didn't exist, you'd have to rewrite your rendering code for every 

Unless you write your own audio subsystem that talks directly to the 

hardware driver, you're going through CoreAudio. Now on windows, I've 

heard that you can get ASIO or Direct-X or other methods of outputting 

sound... this is just not so with OS X (unless of course like I said 

you're talking directly to a piece of hardware via a kernel extension, 

Luckily, CoreAudio has been demonstrated to be one of the most 

performant and stable audio subsystems in any OS. It is usually the 

absolute last place you'll find a bottleneck or instability.

As far as the reduction of CPU code... you're probably SOL for the 

most part there buddy. DSP code written for both the Mac and Windows 

runs on the CPU, not on some pc-card. There are a few exceptions 

where dedicated DSP boards can be used to run individual plugins, but 

there is no way that you'll see any benefit in Max/MSP from those.

What sort of instability are you seeing in Max? What are the specs of 

your machine? Before you go running around looking for hardware to 

solve your problem, perhaps you should take a look at your patches, 

and before you do that you should make sure there is actually a 

problem. If you can't describe the issue in detail, it probably 

Are you saying that Max crashes when you change the sampling rate 

while the DSP is rendering? Don't do that, it's probably not 

supported, and most definitely not necessary. What are you trying to 

As far as hardware goes, save yourself a hassle and go for a MOTU 

Ultralight. Best 500 bucks I've ever spent. Solid as a rock 

> So does anyone know what the most efficient third party audio driver  
> for the Mac is?
First: What are you talking about?
All audio on mac os x must go through CoreAudio.  It's an abstraction  
layer between the thirdparty soundcard drivers and audio software.  If  
it didn't exist, you'd have to rewrite your rendering code for every  
thirdparty driver interface.

Unless you write your own audio subsystem that talks directly to the  
hardware driver, you're going through CoreAudio.  Now on windows, I've  
heard that you can get ASIO or Direct-X or other methods of outputting  
sound... this is just not so with OS X (unless of course like I said  
you're talking directly to a piece of hardware via a kernel extension,  
I've never encountered this).

Luckily, CoreAudio has been demonstrated to be one of the most  
performant and stable audio subsystems in any OS.  It is usually the  
absolute last place you'll find a bottleneck or instability.

As far as the reduction of CPU code... you're probably SOL for the  
most part there buddy.  DSP code written for both the Mac and Windows  
runs on the CPU, not on some pc-card.  There are a few exceptions  
where dedicated DSP boards can be used to run individual plugins, but  
there is no way that you'll see any benefit in Max/MSP from those.

What sort of instability are you seeing in Max?  What are the specs of  
your machine?  Before you go running around looking for hardware to  
solve your problem, perhaps you should take a look at your patches,  
and before you do that you should make sure there is actually a  
problem.  If you can't describe the issue in detail, it probably  
doesn't exist.

Are you saying that Max crashes when you change the sampling rate  
while the DSP is rendering?  Don't do that, it's probably not  
supported, and most definitely not necessary.  What are you trying to  
do?

As far as hardware goes, save yourself a hassle and go for a MOTU  
Ultralight.  Best 500 bucks I've ever spent.  Solid as a rock  
hardware, can take a beating.

> Wait for Max 5 to be 64bits and to support Quad Core

I wince at the task ahead for our heroic developers when they venture 

That will be almost as awe inspiring as their near seamless Windows 

San Francisco, CA    Work: 857-544-3967

I wince at the task ahead for our heroic developers when they venture  
into making Max 64 bit.

That will be almost as awe inspiring as their near seamless Windows  
and then MacTel transition.

San Francisco, CA	Work: 857-544-3967
Email: bthrew@gmail.com
IM: captogreadmore (AIM)
http:/www.barrythrew.com


> What sort of instability are you seeing in Max?

Basically, I've been working on a live performance environment for the last three years or so. At the moment, this consists of a dozen or so patches designed to talk to one another in a variety of combinations.

A couple of months ago, the audio in Max started cutting out when I was using the environment. I've been trying to debug the patches but the problem doesn't seem to be related to anything in the patches in a logical way. The behaviour of the problem isn't consistent but it seems to have deteriorated in the last few days to the point that audio cuts out as soon as I turn the audio on (even though I haven't actually changed anything for a couple of weeks).

This is why I've been searching for OS and hardware solutions - to no avail.

I've had random corruptions in Max patches a couple of times before and managed to solve them by systematically trying to find out which part of the patch is causing the problem, deleting it and replacing it with the a new version of the same thing. But the set up is just too big and gnarly to get into now.

Solution: start programming again from scratch.

> What sort of instability are you seeing in Max?  

Solution: start programming again from scratch.


> Last, why do you use MaxMSP if you don't want any audio input/output ?

All I meant was that I'm not doing any multi track recording/using multi-channel PAs etc... so I don't need 24 ins and outs. I just need one mono audio in and one stereo audio out.

All I meant is that I would guess its going to be REALLY hard to 

update Max to be 64 bit capable (I guess perhaps it could go in 

sections), so hard that I wouldn't expect it by Max 5. It seems like 

a task equally, or more, difficult than the Windows or Intel 

transitions. Everyone send Cycling a case of beer when that's done.

I remember a few hundred posts ago (from the Max 5 at NIME thread) 

someone saying that multi-core support was primarily enabled through 

a poly~ like scheme. While this makes me cry a bit (it would be nice 

if it were application wide), it will be fantastic for any type of 

support, so I don't have to keep opening a separate version of 

runtime when I kid myself by thinking I need that much patch.

I'm not sure if you are asking me this...but if you are a) I do want 

audio input and output, but b) I can think of lots of reasons to use 

All I meant is that I would guess its going to be REALLY hard to  
update Max to be 64 bit capable (I guess perhaps it could go in  
sections), so hard that I wouldn't expect it by Max 5.  It seems like  
a task equally, or more, difficult than the Windows or Intel  
transitions.  Everyone send Cycling a case of beer when that's done.

I remember a few hundred posts ago (from the Max 5 at NIME thread)  
someone saying that multi-core support was primarily enabled through  
a poly~ like scheme.  While this makes me cry a bit (it would be nice  
if it were application wide), it will be fantastic for any type of  
support, so I don't have to keep opening a separate version of  
runtime when I kid myself by thinking I need that much patch.

I'm not sure if you are asking me this...but if you are a) I do want  
audio input and output, but b) I can think of lots of reasons to use  
Max without them.

Quote: Mark Pauley wrote on Thu, 19 July 2007 20:14

"Before you go running around looking for hardware to 

> solve your problem, perhaps you should take a look at your patches,> and before you do that you should make sure there is actually a > problem."

To clarify my position on this - I agree, a shoddy workman blames his tools.

Still, there is a problem - the audio cuts out totally in my Max set-up at random times for no reason I've been able to discern in two months of de-bugging.

Also there does seem to be something wrong with the audio in my computer - occasionally the input receives nothing but feedback and distortion (in all applications). I've done a hardware check plus a massive OS and software re-install and it hasn't helped much.

I have been able to improve the reliability of my Max set-up quite a lot by tidying up my patches but the audio still cuts out from time to time, so it's really no good as a live performance tool (I already had to cancel one show because of this issue).

(i) to keep refining the existing set-up to increase reliability and keep using it for recording at home

(ii) build an entirely new set-up from scratch, taking care to keep it as de-buggable and optimized as possible.

Quote: Mark Pauley wrote on Thu, 19 July 2007 20:14
----------------------------------------------------
"Before you go running around looking for hardware to  
> solve your problem, perhaps you should take a look at your patches,> and before you do that you should make sure there is actually a  > problem."  

To clarify my position on this - I agree, a shoddy workman blames his tools.  

(ii) build an entirely new set-up from scratch, taking care to keep it as de-buggable and optimized as possible.


Just to be clear - certain drivers may be more or less efficient, but 

that's not because of the sound cards, per se. Also, the latency 

issue is not the same as CPU load, which is what the original poster 

> Yes, me too. I was surprised to read someone saying that soundcards 

> don't improve efficiency. With Realtek drivers only, i *can't* even 

> fire up the config i use with my RME (Nuendo with huge hq movie 

> which takes a lot of cpu, MaxMSP in Rewire and a lot of RTCMIX~ 

> inside, reverbs, etc.). Also, Realtek is 200ms latency when RME is 

Just to be clear - certain drivers may be more or less efficient, but  
that's not because of the sound cards, per se. Also, the latency  
issue is not the same as CPU load, which is what the original poster  
referred to.

> Yes, me too. I was surprised to read someone saying that soundcards  
> don't improve efficiency. With Realtek drivers only, i *can't* even  
> fire up the config i use with my RME (Nuendo with huge hq movie  
> which takes a lot of cpu, MaxMSP in Rewire and a lot of RTCMIX~  
> inside, reverbs, etc.). Also, Realtek is 200ms latency when RME is  
> 2ms.
>
> f.e
>

Quote: eanwhite wrote on Fri, 20 July 2007 14:52

I'm afraid that price is most definitely an object, particularly as I only need the most basic functionality and I'm not 100% convinced that it'll help me with my hard-to-diagnose problems anyway.

Is there a super-basic, cheapo sound card out there that doesn't totally suck?

Or should I just bite the bullet and get a MOTU Ultralite?

Quote: eanwhite wrote on Fri, 20 July 2007 14:52
----------------------------------------------------
"> Best? Price no object?"

Or should I just bite the bullet and get a MOTU Ultralite?


I have had no problems with the Motu UltraLite except for one (see 

below). The software and drivers are tight and very reliable. The 

sound is great. The case is sturdy and slick-looking. There is no 

optical in or out but the coaxial is fine.

Problem: the knobs on the front are clumsily close together and 

fragile. Their tight placement can be gotten used to, but if you get 

The knobs fell off of mine regularly when transporting the device in 

my back pack, having been gingerly placed within it and carefully 

remove, and in no way jostled or mistreated in transit. In most 

cases this was quickly repairable by snapping them back in place, but 

most recently the power knob fell off, and I had to ship the 

UltraLite back to Motu (I paid to ship it to them, they paid to ship 

it back), and pay them $90ish USD (their standard repair fee) for 

The whole customer support and replacement process was smooth and 

fast, but IMO it should never have been necessary. It strikes me as 

Best external soundcard for Max/MSP?