Forums > MaxMSP

nasty rewire crashes

January 15, 2007 | 4:32 pm

hello

i’m using maxmsp rewired to live. very often when i change the rewire connection, by changing driver to or from ad_rewire, by breaking it or loading a new patch, or quitting a patch, maxmsp will crash, in such a way that i cant force quit it. it just stays in the dock, and i cant start it up again.

I then have to try and restart or logout, which takes several attempts, sometimes i have to turn off the computer manually. ive tried reinstalling maxmsp, rewire and live.

has anyone else experienced this? i waste about an hour a day restarting my computer repeatedly to get things working….

any help appreciated


January 15, 2007 | 5:17 pm

I totally know what you’re talking about (max< ->logic in my case). This is one of the reasons why I stopped using rewire and started using soundflower + aggregate device. I am very happy with my new setup.

Mattijs


January 15, 2007 | 5:38 pm

how do you send midi to maxmsp?


January 15, 2007 | 5:57 pm

>how do you send midi to maxmsp?
>–

they did not gave you the manual????

wow!!

kasper


January 15, 2007 | 6:21 pm

haha. i meant from to maxmsp.


January 15, 2007 | 6:21 pm

i meant from logic to maxmsp



FP
January 15, 2007 | 6:23 pm


January 15, 2007 | 6:32 pm

i’d be interested if there is a something else, IAC timing is not good, and rewire is so unstable… making all my patches into VSTis is a pain in the ass , plus i cant tweak them. i like tweaking.


January 15, 2007 | 6:53 pm

Quote: bin wrote on Mon, 15 January 2007 19:32
—————————————————-
> i’d be interested if there is a something else, IAC timing is not good, and rewire is so unstable… making all my patches into VSTis is a pain in the ass , plus i cant tweak them. i like tweaking.
—————————————————-

We have max separate from logic as well, for the same reasons and the fact that we combine the output of multiple channels in logic into one audiovisual output.

The most common way to send midi from logic to max is to make a midi channel in logic and set its output to ‘to MaxMSP 1′ (or ‘to Max/MSP 1′ in 4.5). Max’ midi device works perfectly here. No delay and accurate timing.

But our setup is different. I have several plugs in Logic that send midi -and- automation data to max. The vst instruments have pp’s but no plugmidiin. The pp’s output their data to max via udp. To send the midi data via midi instead of udp: in logic’s environment, connect the output of the audio track object to a GM multi-instrument. Set the output of this multi-instrument to ‘to MaxMSP 1′

Several different channels of audio generated by max are sent back to logic via the aggregate device that consists of the real output (M-Audio FW410) and soundflower 16ch.

Right now I’m working out a trick to send automation data from logic to max via midi as well, since for a strange reason udp can be quite inaccurate (30 ms off) if used from logic to max (not in other situations).

Hth,
Mattijs


January 15, 2007 | 7:19 pm

When you get a nasty crash, instead of rebooting, just open up the terminal,
get a list of the processes, and kill the bothersome one.

So, lets say Firefox crashed and isn’t disappearing. Open up
Terminal.appand type: ps -aux | grep Firefox

You should get something back like this:

someuser 367 13.6 13.4 348088 105260 ?? R Sun05AM
128:27.58/Applications/Firefox.app/Contents/MacOS/firefox-bin
-psn_0_2097153
someuser 1076 0.0 0.0 18644 92 std R+ 2:14PM 0:00.00 grep
Firefox

see the 367? that’s the PID for Firefox.

Type: kill -9 367

It’ll disappear from the dock immediately.

William

On 1/15/07, bin ray wrote:
>
>
> i’d be interested if there is a something else, IAC timing is not good,
> and rewire is so unstable… making all my patches into VSTis is a pain in
> the ass , plus i cant tweak them. i like tweaking.
> –
> http://www.myspace.com/binray
>


January 15, 2007 | 7:57 pm

Unluckily, this doesn’t always work when audio drivers are involved. I have had occasions when restarting max didn’t help (audio drivers still not working) and the computer even refused to shut down.

Mattijs

Quote: williamcotton wrote on Mon, 15 January 2007 20:19
—————————————————-
> When you get a nasty crash, instead of rebooting, just open up the terminal,
> get a list of the processes, and kill the bothersome one.
>
> So, lets say Firefox crashed and isn’t disappearing. Open up
> Terminal.appand type: ps -aux | grep Firefox
>
> You should get something back like this:
>
> someuser 367 13.6 13.4 348088 105260 ?? R Sun05AM
> 128:27.58/Applications/Firefox.app/Contents/MacOS/firefox-bin
> -psn_0_2097153
> someuser 1076 0.0 0.0 18644 92 std R+ 2:14PM 0:00.00 grep
> Firefox
>
> see the 367? that’s the PID for Firefox.
>
> Type: kill -9 367
>
> It’ll disappear from the dock immediately.
>
> William
>
>
>
> On 1/15/07, bin ray wrote:
> >
> >
> > i’d be interested if there is a something else, IAC timing is not good,
> > and rewire is so unstable… making all my patches into VSTis is a pain in
> > the ass , plus i cant tweak them. i like tweaking.
> > –
> > http://www.myspace.com/binray
> >
>
>
>
—————————————————-


January 15, 2007 | 8:25 pm

well, i tried using MIDI out from Live to max/msp 1, (i guess thats the IAC bus?) the timing is ok, but not as good as rewire.

(as i’ve said before the timing seems better if the bpm is a multiple of the signal vector size)

i tried getting my MSP audio into Live using 2 channel soundflower, and it doesnt sound very nice, it seems to drop out/break up alot.

any tips for optimizing it? i tried adjusting the soundflower buffer size, no change

thanks for the terminal tips, but why is it crashing so bad anyway? is it rewire is going wrong? if anything it was worse with Logic than Live


January 15, 2007 | 9:48 pm

Quote: bin wrote on Mon, 15 January 2007 21:25
—————————————————-
> well, i tried using MIDI out from Live to max/msp 1, (i guess thats the IAC bus?) the timing is ok, but not as good as rewire.

How good is that? When I use [timer] to measure the differences I get a maximum of 0.03 ms variation.

>
> (as i’ve said before the timing seems better if the bpm is a multiple of the signal vector size)

I never experienced that..

>
> i tried getting my MSP audio into Live using 2 channel soundflower, and it doesnt sound very nice, it seems to drop out/break up alot.

Hum.. could this be a Live issue? I don’t have those problems here. Sounds like your buffer size is too small, but if you tried to adjust that I don’t know what else could be wrong. 512 should be majorly enough. I use 128 samples and it works perfectly with at least 4 stereo channels and a lot of dsp processing in max (on a mac pro 4×2,66 ghz that is.. ahum).

>
> any tips for optimizing it? i tried adjusting the soundflower buffer size, no change

I had problems with soundflowerbed though. Are you using that?

>
> thanks for the terminal tips, but why is it crashing so bad anyway? is it rewire is going wrong? if anything it was worse with Logic than Live

This is not the only thing that goes wrong with rewire. You could check this thread:

http://www.cycling74.com/forums/index.php?t=msg&rid=3579&S=8890842524145bfd80f3d495102836d1&th=22199&goto=82211#msg_82277

Mattijs


January 15, 2007 | 9:52 pm

Quote: bin wrote on Mon, 15 January 2007 21:25
—————————————————-
> but why is it crashing so bad anyway?

yea..sometimes it looks like we are stuck in this stone age of computer programming forever.


January 16, 2007 | 12:19 pm

Mattijs Kneppers wrote:
> How good is that? When I use [timer] to measure the differences I get
> a maximum of 0.03 ms variation.

With timer your accuracy is not better than 1ms or the actual scheduler
interval. How do you measure 0.03 ms???

>> (as i’ve said before the timing seems better if the bpm is a
>> multiple of the signal vector size)

I am not sure what would be audible, but assuming scheduler in
audiointerupt is set, a multiple of the vector size would fit exactly,
whereas anything else would sometimes need a tick more or a tick less.

That sounds absolutely reasonable, but what is the vector size? If you
belong to those who can here a few samples of jitter (I am not!) then
you’d have to go always the pure audio route, Midi is then not for
timing critical applications, only for slow controler changes etc…

Stefan


Stefan Tiedje————x——-
–_____———–|————–
–(_|_ —-|—–|—–()——-
– _|_)—-|—–()————–
———-()——–www.ccmix.com


January 16, 2007 | 4:40 pm

Quote: Stefan Tiedje wrote on Tue, 16 January 2007 13:19
—————————————————-
> Mattijs Kneppers wrote:
> > How good is that? When I use [timer] to measure the differences I get
> > a maximum of 0.03 ms variation.
>
> With timer your accuracy is not better than 1ms or the actual scheduler
> interval. How do you measure 0.03 ms???

Good question.. but it’s what I get. I receive midi events in max, put on every 2 beats at 120 bpm and time them. I get 999.9723489, 1000.02345, .. etc

>
> >> (as i’ve said before the timing seems better if the bpm is a
> >> multiple of the signal vector size)
>
> I am not sure what would be audible, but assuming scheduler in
> audiointerupt is set, a multiple of the vector size would fit exactly,
> whereas anything else would sometimes need a tick more or a tick less.
>
> That sounds absolutely reasonable, but what is the vector size? If you
> belong to those who can here a few samples of jitter (I am not!) then
> you’d have to go always the pure audio route, Midi is then not for
> timing critical applications, only for slow controler changes etc…

For me, in this case, an accuracy of 5 ms is really tight enough. So a vector size of 128 samples would do fine.

>
> Stefan
>
> —
> Stefan Tiedje————x——-
> –_____———–|————–
> –(_|_ —-|—–|—–()——-
> — _|_)—-|—–()————–
> ———-()——–www.ccmix.com
>
>
—————————————————-


January 16, 2007 | 5:10 pm

it depends alot what youre trying to do. i’m making sounds that are very fast with alot of chopped noise, basically i want as near to sample accuracy as possible. i test my synching by writing a lot of hits on sixteenths of a note , say, and running it at about 200bpm. if it sounds like a tone, then its good. if its wandering ,then its not. the only way i can get this sounding like a tone is using rewire and multiples of the vector size, such as 215.332031 bpm.
IAC bus just doesnt seem stable, its affected by other processes , sounds sloppy to me.


January 16, 2007 | 5:22 pm

Quote: bin wrote on Tue, 16 January 2007 18:10
—————————————————-
> it depends alot what youre trying to do. i’m making sounds that are very fast with alot of chopped noise, basically i want as near to sample accuracy as possible. i test my synching by writing a lot of hits on sixteenths of a note , say, and running it at about 200bpm. if it sounds like a tone, then its good. if its wandering ,then its not. the only way i can get this sounding like a tone is using rewire and multiples of the vector size, such as 215.332031 bpm.
> IAC bus just doesnt seem stable, its affected by other processes , sounds sloppy to me.
—————————————————-

Aaa! HAHAHA! Yeah, um, sorry. Ok, I understand. That’s a really high standard :) Cool test though.

Mattijs


January 16, 2007 | 5:37 pm

it depends alot what youre trying to do. i’m making sounds that are very fast with alot of chopped noise, basically i want as near to sample accuracy as possible. i test my synching by writing a lot of hits on sixteenths of a note , say, and running it at about 200bpm. if it sounds like a tone, then its good. if its wandering ,then its not. the only way i can get this sounding like a tone is using rewire and multiples of the vector size, such as 215.332031 bpm.
IAC bus just doesnt seem stable, its affected by other processes , sounds sloppy to me.


January 16, 2007 | 5:43 pm

oops…

anyway…it sounds anal, but since i’m doing signal accurate sequencing in max, it seems a shame to lose that timing in sequencing the chunks of audio in live or whatever….ive been struggling with synching for years!!


January 16, 2007 | 6:16 pm

i think the IAC bus is an inbuilt mac osx virtual port – which can be configured from the audiomidi configuration application in /utilities. and to / from maxmsp 1 / 2, are virtual midi ports created by max.

i havent done anything with rewire for a while, but i used hostphasor~ and hostsync~ to create audio rate sequencers in msp which could then be sent back to ableton via rewire~.

perhaps u could explain abit more how u are using rewire…
are u sending audio / midi thru rewire, if so how many channels?
which application is the rewire host?
which combination of rewire objects are u using?
is this happening with live 5/6?

j


January 17, 2007 | 1:36 am

>perhaps u could explain a bit more how u are using rewire…
>are u sending audio / midi thru rewire, if so how many channels?
>which application is the rewire host?
>which combination of rewire objects are u using?
>is this happening with live 5/6?

3 midi channels
3 stereo audio channels
Live 6 is the host, although ive used Logic 6.3.1, which is very unstable. i cant get rewire to start up at all with logic express 7.2 (though that was a crack)

rewire objects in max? just hostsync for bpm info, midiin and dac~

i was thinking….a nice way to get round rewire might be to have soundcard where you can route an output back to an input. i tried short circuiting my SPDIF connection on my firewire audiophile, but that doesnt work at all (it cant synch to itself apparently)…


January 17, 2007 | 9:33 am

Quote: bin wrote on Wed, 17 January 2007 02:36
—————————————————-

> i was thinking….a nice way to get round rewire might be to have soundcard where you can route an output back to an input. i tried short circuiting my SPDIF connection on my firewire audiophile, but that doesnt work at all (it cant synch to itself apparently)…

but, brother… that’s what soundflower is for! If the sound stutters, there is probably a buffer problem on your setup, but if you solve it, your problems will be history!

Amen!


January 17, 2007 | 9:58 am

Mattijs Kneppers wrote:
> Quote: bin wrote on Wed, 17 January 2007 02:36
>
>> i was thinking….a nice way to get round rewire might be to have
>> soundcard where you can route an output back to an input. i tried
>
> but, brother… that’s what soundflower is for! If the sound
> stutters, there is probably a buffer problem on your setup, but if
> you solve it, your problems will be history!

The Totalmix-enabled RME interfaces can do internal loopback (no extra
load anywhere) and it’s a very handy feature, though you do need
otherwise unused inputs and outputs available. (I use a couple of spare
ADAT channels.) I think there are a few other products around that can
do similar tricks. I find it works much better than soundflower or any
other software solution.


January 17, 2007 | 1:44 pm

bin ray wrote:
> sounds sloppy to me.

of course. If you do sound with it, you have to stay always in the
signal domain. No way around. "Sound problems" should not being
considered "timing problems". As soon as you leave the signal domain you
get a jitter of about the scheduler interval. Which is fine for timing,
but not enough for sound.
And if on equal divisions of the signal vector size its sounding good,
that means you have an excelent timing!!! Can’t get any better than that…

Stefan


Stefan Tiedje————x——-
–_____———–|————–
–(_|_ —-|—–|—–()——-
– _|_)—-|—–()————–
———-()——–www.ccmix.com


January 17, 2007 | 3:32 pm

oh, yeah reiwre sounds good, but its just so unstable! i just logged in/out 5 times in the last 20 minutes while it froze my apps up….


January 17, 2007 | 4:04 pm

Mattijs Kneppers wrote:
> How good is that? When I use [timer] to measure the differences I get
> a maximum of 0.03 ms variation.

on my setup, using IAC bus from Live6 to maxmsp and soundflower back to Live6, i am getting hits which wander by up to 8ms from where they should be. some hits seem to be missed altogether.


January 17, 2007 | 4:09 pm

bin ray wrote:
> i was thinking….a nice way to get round rewire might be to have
> soundcard where you can route an output back to an input. i tried
> short circuiting my SPDIF connection on my firewire audiophile, but
> that doesnt work at all (it cant synch to itself apparently)…

So it has to be set to internal clock apparently…

Stefan


Stefan Tiedje————x——-
–_____———–|————–
–(_|_ —-|—–|—–()——-
– _|_)—-|—–()————–
———-()——–www.ccmix.com


January 17, 2007 | 4:32 pm

sending midi via the IAC busfrom Live to maxmsp seems to give fairly good timing, if i set drivers in maxmsp to my audiophile soundcards outs. however when i select soundflower 2in/out as driver in maxmsp, the timing of the hits which come into Live on a soundflower return channel is erratic, some hits are missing altogether. also there is a latency of around 230ms (rewire was 17ms)

i’m not clear which buffer setting i should alter for soundflower to work efficiently, tried altering Live’s buffer and maxmsps signal and vector buffer sizes, seems to be a trade off between timing and sound breaking up, with no real middle ground where it sounds good..


January 17, 2007 | 4:35 pm

bin ray wrote:
> i was thinking….a nice way to get round rewire might be to have
> soundcard where you can route an output back to an input. i tried
> short circuiting my SPDIF connection on my firewire audiophile, but
> that doesnt work at all (it cant synch to itself apparently)…

So it has to be set to internal clock apparently…

no joy.

from Rightmark forum…

"With many interfaces, if you have to check something in the driver control in order for the s/pdif input to work (External clock), the interface’s own sample clock stops until it locks on the s/pdif input clock. For however long the sync lock takes, there is no clock transmitted from the s/pdif output. This would explain why s/pdif loopback can’t work with your FW410."


January 18, 2007 | 2:53 pm

bin ray wrote:
> This would explain why s/pdif loopback can’t work with your FW410."

There seem to be very stupid designs in the world, just avoid those
companies in the future… ;-)

I remeber I had some difficulties once to connect a reverb digital with
ins and outs as send/return via spdif to a Digi 002. But in the end it
did work out fine, can’t remember what was needed to get ProTools to
accept it.
This is also always a loopback and a standard need for any studio, if an
interface would not allow to do it, its just plain useless, or should
better ship without digital connections to not confuse people…

Of course the real professionals sync the whole studio with a wordclock,
and anything without BNC connector is refused and justifies for a 10
times higher price… ;-)

Stefan


Stefan Tiedje————x——-
–_____———–|————–
–(_|_ —-|—–|—–()——-
– _|_)—-|—–()————–
———-()——–www.ccmix.com


Viewing 31 posts - 1 through 31 (of 31 total)