Audiobus: Use your music apps together.

What is Audiobus?Audiobus is an award-winning music app for iPhone and iPad which lets you use your other music apps together. Chain effects on your favourite synth, run the output of apps or Audio Units into an app like GarageBand or Loopy, or select a different audio interface output for each app. Route MIDI between apps — drive a synth from a MIDI sequencer, or add an arpeggiator to your MIDI keyboard — or sync with your external MIDI gear. And control your entire setup from a MIDI controller.

Download on the App Store

Audiobus is the app that makes the rest of your setup better.

AUM: Bad Signal alarm

I don’t see that anyone else has mentioned this here. I’m surely not the only one who has experienced this.
It happened to me last night after loading Rymdigare . I’ve been using multiple instances of DRC in that template you see for many months. The only other effect was rymdigare, which I had just loaded. So, I suppose let the Rymdigare Dev know about it?
I don’t use it extensively, but have used it before.

«1

Comments

  • @Blipsford_Baubie said:
    I don’t see that anyone else has mentioned this here. I’m surely not the only one who has experienced this.
    It happened to me last night after loading Rymdigare . I’ve been using multiple instances of DRC in that template you see for many months. The only other effect was rymdigare, which I had just loaded. So, I suppose let the Rymdigare Dev know about it?
    I don’t use it extensively, but have used it before.

    >

    Is the problem reproducible?

  • i had the same problem in the last 2 or 3 weeks. with several auv3 plugs from different developers. i don't think it is a bug in the auv3's. i guess it is a ios problem or a bug in AUM ? but unfortunately i could not find a way to reproduce it.

  • I've had it happen to me twice using AudioDamage Replicant...can't reproduce it at will though...

  • I've only had that happen when I tried to push my old iPad Air 2 over it's limits...
    ...audio started to crackle and the error popped up.

    My guess it that something goes wonky and and empty frame slips in when the plug-ins can't deliver data.
    That would explain the Inf/NaN ie. a frame with zero lenght / no content.

    This is more likely to happen when the audio frames are not time-stamped (ie. with older apps).
    That's just a rough guess, but since I can't intentionally and methodically provoke it, it gets tricky...

  • wimwim
    edited July 2022

    Do you think I'd get sued by Nine Inch Nails if I named my new group NaИ?

  • edited July 2022

    @espiegel123 said:

    @Blipsford_Baubie said:
    I don’t see that anyone else has mentioned this here. I’m surely not the only one who has experienced this.
    It happened to me last night after loading Rymdigare . I’ve been using multiple instances of DRC in that template you see for many months. The only other effect was rymdigare, which I had just loaded. So, I suppose let the Rymdigare Dev know about it?
    I don’t use it extensively, but have used it before.

    >

    Is the problem reproducible?

    No. It’s been a few days that I’ve been trying. But I was also on 15.4.1 when it happened, and I upgraded right after it happened, not to imply that was the solution.

  • @wim said:
    Do you think I'd get sued by Nine Inch Nails if I named my new group NaИ?

    Just keep it as NIN, but they’ll them you swapped the 1st letters….like Com Truise or others in that name fad.

  • It came with the last update 1.4.0 from AUM.

    I think this option make the problem.

    Ive an old project, never had problems with it. I use one channel- with the plugins you see in the photo below - in some projects. Since this update i got problems with bad signals…

    Even if i take all plugins out.

    The ios version is the same, the problem came with the AUM update 1.4.0.
    What do think @j_liljedahl?

  • I've had this problem many times. In some cases I found it reproducible. I remember for example a session where I would have no problems until I added a Rum instance after Thermo. Mentioned it to Jonathan, AUM dev, this is what he said:

    It means that a plugin in that channel is producing invalid floating
    point values in the signal, e.g. NaNs or Infinities. Report it to the
    plugin developer so they can fix it.

    I'll update the users guide to mention this.

  • The tricky part is trying to nail down which plug-in in the chain causes it...
    Don't know if AUM can internally track this and put a red-circle on the potential pair of trouble-makers?

  • One case I can think if is when multiple channels bombard the same output...
    ...ie. they are not properly bussed/mixed before being passed on to the main output.

  • @headmademusic said:
    It came with the last update 1.4.0 from AUM.

    I think this option make the problem.

    Ive an old project, never had problems with it. I use one channel- with the plugins you see in the photo below - in some projects. Since this update i got problems with bad signals…



    Even if i take all plugins out.

    The ios version is the same, the problem came with the AUM update 1.4.0.
    What do think @j_liljedahl?

    Your screenshot shows there are more plugins hidden away in the scrollable effect chain. Are you sure all is bypassed?

  • @Samu said:
    The tricky part is trying to nail down which plug-in in the chain causes it...
    Don't know if AUM can internally track this and put a red-circle on the potential pair of trouble-makers?

    In theory I could check the signal for bad values after each plugin, but that would use a bit too much extra CPU power for my taste.

    @Samu said:
    One case I can think if is when multiple channels bombard the same output...
    ...ie. they are not properly bussed/mixed before being passed on to the main output.

    That's not relevant. Any number of good signals won't produce a bad signal when mixed, regardless if they are mixed via a bus or sent to a common output. The maths are the same. However, a bad signal mixed with any number of good signals will still produce a bad signal, which is the core problem with bad signals and why they need to be avoided.

    @Samu said:
    I've only had that happen when I tried to push my old iPad Air 2 over it's limits...
    ...audio started to crackle and the error popped up.

    My guess it that something goes wonky and and empty frame slips in when the plug-ins can't deliver data.
    That would explain the Inf/NaN ie. a frame with zero lenght / no content.

    A zero length buffer means there are no values to check, so it wouldn't trigger the bad signal detection. However in theory some plugin might have an algorithm that breaks when there is a dropout and written in a way that it produces bad values.

    This is more likely to happen when the audio frames are not time-stamped (ie. with older apps).
    That's just a rough guess, but since I can't intentionally and methodically provoke it, it gets tricky...

    timestamps are not relevant here.

    @TomNoise said:
    i had the same problem in the last 2 or 3 weeks. with several auv3 plugs from different developers. i don't think it is a bug in the auv3's. i guess it is a ios problem or a bug in AUM ? but unfortunately i could not find a way to reproduce it.

    Why don't you think it's a bug in the plugins? I have tested this bad signal detection thoroughly and it seems solid to me. Note that if a plugin produces just a single bad value sample, it would probably not be noticed, but it could still blow things up later down the fx chain if it hits a recursive filter or reverb etc (any algo with feedback in it).

  • I think the confusion here is in the reference to the “problem”. From @headmademusic ’s viewpoint there was no problem because it wasn’t visible. Now it is, and it gets in the way, so to them it’s a new problem.

    Personally I think it’s good to report it because it’s likely that previously invisible problem will cause other problems and those will be tough to isolate. But I can see the other perspective as well. That seems like something I’d make able to switch off. It would be a great troubleshooting tool, but would be really frustrating if it interrupted a creative session or performance.

  • @wim said:
    I think the confusion here is in the reference to the “problem”. From @headmademusic ’s viewpoint there was no problem because it wasn’t visible. Now it is, and it gets in the way, so to them it’s a new problem.

    Personally I think it’s good to report it because it’s likely that previously invisible problem will cause other problems and those will be tough to isolate. But I can see the other perspective as well. That seems like something I’d make able to switch off. It would be a great troubleshooting tool, but would be really frustrating if it interrupted a creative session or performance.

    Switching it off would just mean the problem shows up in different ways, which leads to support emails about why their session suddenly got silent, or even the whole device. CoreAudio sometimes ducks the level if there is a sample with a huge value, and it can takes minutes for it to raise up to normal level again.

    NaNs and Infs should not be tolerated, at all.

  • The user and all related content has been deleted.
  • @j_liljedahl said:

    @wim said:
    I think the confusion here is in the reference to the “problem”. From @headmademusic ’s viewpoint there was no problem because it wasn’t visible. Now it is, and it gets in the way, so to them it’s a new problem.

    Personally I think it’s good to report it because it’s likely that previously invisible problem will cause other problems and those will be tough to isolate. But I can see the other perspective as well. That seems like something I’d make able to switch off. It would be a great troubleshooting tool, but would be really frustrating if it interrupted a creative session or performance.

    Switching it off would just mean the problem shows up in different ways, which leads to support emails about why their session suddenly got silent, or even the whole device. CoreAudio sometimes ducks the level if there is a sample with a huge value, and it can takes minutes for it to raise up to normal level again.

    NaNs and Infs should not be tolerated, at all.

    I’m totally with you on that. As you see though, people will still tend to blame AUM event though you make it clear that the problem is with a plug-in.

    Better the enemy you see than one that you don’t though, I guess. 😉

  • edited July 2022

    @j_liljedahl said:

    Your screenshot shows there are more plugins hidden away in the scrollable effect chain. Are you sure all is bypassed?

    Im totally sure that all plugins bypassed, there are 6 plugins in the chain + one synth. How i said i was using this chain in different projects without a dropout or chrash. In the 2nd screenshot you see the last auv3 in the chain pcm2612 in the 3rd screenshot you see the first auv3 koalafx in the chain. So, totally sure. :smile:

    @wim said:

    @j_liljedahl said:

    @wim said:
    I think the confusion here is in the reference to the “problem”. From @headmademusic ’s viewpoint there was no problem because it wasn’t visible. Now it is, and it gets in the way, so to them it’s a new problem.

    Personally I think it’s good to report it because it’s likely that previously invisible problem will cause other problems and those will be tough to isolate. But I can see the other perspective as well. That seems like something I’d make able to switch off. It would be a great troubleshooting tool, but would be really frustrating if it interrupted a creative session or performance.

    Switching it off would just mean the problem shows up in different ways, which leads to support emails about why their session suddenly got silent, or even the whole device. CoreAudio sometimes ducks the level if there is a sample with a huge value, and it can takes minutes for it to raise up to normal level again.

    NaNs and Infs should not be tolerated, at all.

    I’m totally with you on that. As you see though, people will still tend to blame AUM event though you make it clear that the problem is with a plug-in.

    Better the enemy you see than one that you don’t though, I guess. 😉

    But which of the 7 auv3 make the problem? I understand the point and i know that not AUM is the problem but it was not a problem for me before the update.

  • @headmademusic said:

    @j_liljedahl said:

    Your screenshot shows there are more plugins hidden away in the scrollable effect chain. Are you sure all is bypassed?

    Im totally sure that all plugins bypassed, there are 6 plugins in the chain + one synth. How i said i was using this chain in different projects without a dropout or chrash. In the 2nd screenshot you see the last auv3 in the chain pcm2612 in the 3rd screenshot you see the first auv3 koalafx in the chain. So, totally sure. :smile:

    If all plugins are bypassed, there should be no signal at all that could contain bad values. If it happens again, please record a video where you scroll through all nodes showing them bypassed. Also note that the ”bad signal” sign stays lit a couple of seconds before it resets. If it stays forever, with all plugins bypassed, it would be a bug in AUM that I’d like to fix. But I need actual ”evidence” before spending time on hunting that down!

    @wim said:

    @j_liljedahl said:

    @wim said:
    I think the confusion here is in the reference to the “problem”. From @headmademusic ’s viewpoint there was no problem because it wasn’t visible. Now it is, and it gets in the way, so to them it’s a new problem.

    Personally I think it’s good to report it because it’s likely that previously invisible problem will cause other problems and those will be tough to isolate. But I can see the other perspective as well. That seems like something I’d make able to switch off. It would be a great troubleshooting tool, but would be really frustrating if it interrupted a creative session or performance.

    Switching it off would just mean the problem shows up in different ways, which leads to support emails about why their session suddenly got silent, or even the whole device. CoreAudio sometimes ducks the level if there is a sample with a huge value, and it can takes minutes for it to raise up to normal level again.

    NaNs and Infs should not be tolerated, at all.

    I’m totally with you on that. As you see though, people will still tend to blame AUM event though you make it clear that the problem is with a plug-in.

    Better the enemy you see than one that you don’t though, I guess. 😉

    But which of the 7 auv3 make the problem? I understand the point and i know that not AUM is the problem but it was not a problem for me before the update.

    The idea is that you can then just bypass plugins to see which one is producing the bad values.

  • @j_liljedahl said,
    Also note that the ”bad signal” sign stays lit a couple of seconds before it resets. If it stays forever, with all plugins bypassed, it would be a bug in AUM that I’d like to fix. But I need actual ”evidence” before spending time on hunting that down!

    Hi, I’m the OP, and that is precisely what happened to me.
    The “bad signal” sign remained lit not only after I bypassed my plugins, but also after ejecting my plugins. I had to quit and relaunch AUM for the sign to go away. But, I have no proof. Also, I was on an older version of iOS 15. I will remember the screen recording next time, good to know, thanks for the info and for an amazing app.

  • @Blipsford_Baubie said:
    @j_liljedahl said,

    Also note that the ”bad signal” sign stays lit a couple of seconds before it resets. If it stays forever, with all plugins bypassed, it would be a bug in AUM that I’d like to fix. But I need actual ”evidence” before spending time on hunting that down!

    Hi, I’m the OP, and that is precisely what happened to me.
    The “bad signal” sign remained lit not only after I bypassed my plugins, but also after ejecting my plugins. I had to quit and relaunch AUM for the sign to go away. But, I have no proof. Also, I was on an older version of iOS 15. I will remember the screen recording next time, good to know, thanks for the info and for an amazing app.

    Ok, please let me know if you find a way to reproduce that, and I'll have a look.

  • @j_liljedahl said:
    ......why their session suddenly got silent, or even the whole device.

    Well this explains a lot from a past issue I've had. I just stopped using the plugin, but it wasn't making any sense to me since its a pretty popular app here (ravencroft).

  • I got the Bad Signal warning using the xy pad in KB-1 to alter filter resonance and cutoff in Sunrizer today, it stopped everything, froze the screen, and I had to dump the session and reboot my iPad to get sound (or anything) back again. Not great.

  • @AlmostAnonymous said:

    @j_liljedahl said:
    ......why their session suddenly got silent, or even the whole device.

    Well this explains a lot from a past issue I've had. I just stopped using the plugin, but it wasn't making any sense to me since its a pretty popular app here (ravencroft).

    Ravenscroft going silent was likely nothing to do with AUM. Ravenscroft can only handle a tiny amount of polyphony and stops working when fed too many notes. Eg sending sustained chords from Polyphase into Ravenscroft will often result in silence sooner or later. This can then be fixed by switching presets in Ravenscroft and switching back if necessary. A better solution is to use pure piano, which handles polyphony better. If you stopped using AUM for this reason, it was definitely the wrong reason!

  • @Svetlovska said:
    I got the Bad Signal warning using the xy pad in KB-1 to alter filter resonance and cutoff in Sunrizer today, it stopped everything, froze the screen, and I had to dump the session and reboot my iPad to get sound (or anything) back again. Not great.

    Well there we go, yes, there are definitely issues with this new bad signal setup. I guess ppl need to start sending Jonatan their screenshots / vids of this type of thing

  • edited July 2022

    @Gavinski said:

    @AlmostAnonymous said:

    @j_liljedahl said:
    ......why their session suddenly got silent, or even the whole device.

    Well this explains a lot from a past issue I've had. I just stopped using the plugin, but it wasn't making any sense to me since its a pretty popular app here (ravencroft).

    Ravenscroft going silent was likely nothing to do with AUM. Ravenscroft can only handle a tiny amount of polyphony and stops working when fed too many notes. Eg sending sustained chords from Polyphase into Ravenscroft will often result in silence sooner or later. This can then be fixed by switching presets in Ravenscroft and switching back if necessary. A better solution is to use pure piano, which handles polyphony better. If you stopped using AUM for this reason, it was definitely the wrong reason!

    I know it’s an issue with raven. I didn’t stop using AUM. Swap out raven for anything else, and it would never happen. But it was doing it with only itself loaded in a single track with barely any notes going into it.

  • @AlmostAnonymous said:

    @Gavinski said:

    @AlmostAnonymous said:

    @j_liljedahl said:
    ......why their session suddenly got silent, or even the whole device.

    Well this explains a lot from a past issue I've had. I just stopped using the plugin, but it wasn't making any sense to me since its a pretty popular app here (ravencroft).

    Ravenscroft going silent was likely nothing to do with AUM. Ravenscroft can only handle a tiny amount of polyphony and stops working when fed too many notes. Eg sending sustained chords from Polyphase into Ravenscroft will often result in silence sooner or later. This can then be fixed by switching presets in Ravenscroft and switching back if necessary. A better solution is to use pure piano, which handles polyphony better. If you stopped using AUM for this reason, it was definitely the wrong reason!

    I know it’s an issue with raven. I didn’t stop using AUM. Swap out raven for anything else, and it would never happen. But it was doing it with only itself loaded in a single track with barely any notes going into it.

    Ah I get you now! Did u report that to Ravenscroft dev? Mind you, from previous experience, they don't seem responsive

  • edited July 2022

    I hadn’t. Honestly there was a lot of weird things going on with my device then so I chalked it up to that. Then the busy season came up for me….
    But now I’ll revisit it. All of that was before the update tho so never a bad signal warning.

  • @Gavinski said:

    @Svetlovska said:
    I got the Bad Signal warning using the xy pad in KB-1 to alter filter resonance and cutoff in Sunrizer today, it stopped everything, froze the screen, and I had to dump the session and reboot my iPad to get sound (or anything) back again. Not great.

    Well there we go, yes, there are definitely issues with this new bad signal setup. I guess ppl need to start sending Jonatan their screenshots / vids of this type of thing

    My guess is that something regarding modulating the Sunrizer filter made their algo blow up (for example, some common filter algorithms break if their parameters are changed too fast), which in turn made it produce bad signal values. Device freezing is a sign of some process spinning the CPU at 100%, it could be Sunrizer who did this as a side effect of the above mentioned filter blow up. There is nothing in my bad-signal-detection code in itself that would explain such a behavior. What it does is alert the user and silence that channel, so that the rest of the channels can continue sounding, instead of as before: letting the bad signal values through to CoreAudio which would then detect them and silence the whole device. Also remember that bad signals can "contaminate" any dsp algorithm that has feedback in it (like all delays, reverbs, and almost every filter), which means they might break until restarted depending on how they are coded.

    Now, if you see the Bad Signal alert, it means bad signal values was detected in that channel strip (NaN's and Infinities). Then go and contact the developer of the plugin producing the bad values. But if you see any bugs in AUM regarding this, such as the Bad Signal sign staying after ejecting all nodes in the channel and waiting a couple of seconds, then please report it to me.

  • Ditto replicant 2 and DRC…possibly more…?

Sign In or Register to comment.