[Bug] Audible glitch when using MIDI tempo linked to the delay time

P

psimon666

Member
Hello

Here is the trick:
1) The concerned effect is the MultiDelay,but i think other blocks are concerned. I chose a delay time linked to the tempo (1/4 or 1/2 for instance).
2) The MIDI input of the AxeFx receives the MIDI clock of the DAW (i'm not using the USB at all in this story)
3) In my DAW (130 bpm) : I play a recorded dry guitar track, and send it to the AxeFxIII (to make it 'wet') using the Analog audio inputs.
4) I hear errratic glitches

After seeking the problem, i discovered that this is because the tempo varies a little bit. This is visible in AxeEdit : a tempo of 130 bpm (in the DAW) is interpretated as a 129 bpm or a 130 bpm. This changes all the time.
Each time this tempo changes in AxeEdit (so inside the AxeFxIII), the glitch is audible

You can solve the problem if you unlink the delay to the tempo, and use a fix value (in ms)
But this is not smart at all

Thanks
 
Changing the tempo on a delay will cause glitches... This is not limited to Fractal.

The problem here (based on my experience with the same) is that the Axe Fx processing of the beat clock clock seems a bit "jumpy".

I reported issues quite a while back and Cliff improved it, bit it is still not super stable.

I have not used midi clock with any other devices so I don't know if the behavior is typical or not.

In my case I was using BandHelper over USB and I was able to program it to just send tempo for 4 bars and then stop. That would set my tempo for a given song...
 
Yes, that's a question of 'jump', that's what i think.
I can't solve the problem the way you do, because i have song with tempo changes (for instance, one starts at 100 bpm and ends at 250 bpm).
Nevertheless, this comes from the MIDI protocol that (as i remember) does not include a 'tempo' command. It has to be interpreted by the receiver (AxeFxIII here) all song long
 
GlennO said:
I beg to differ :). Most delay effects these days, including all of mine, adapt to tempo changes without glitching. A bpm change of 1 shouldn't cause any artifacts.
Click to expand...
Well maybe they don't have to... Especially digital ones where I'm sure the time can be programmatically slewed to account for the changes.

But I was referring to pedals and most (many?) of the Fractal delays model analog pedals (including the glitches). Some users intentionally take advantage of those behaviors ;)

Curious - have you used a beat clock with the Axe Fx III? I haven't recently, but when I was it was jumpy by more than 1 BPM... Occasionally, there were brief but large jumps.

It might be cool if the Axe Fx had an option in the delays to select between "slew" and "jump" modes for tempo changes.
 
Remember that my tempo has not changed. It is fixed to 130 bpm.
The problem here, is that the glitch/pop, appears all the time (at fixed tempo)
 
unix-guy said:
Curious - have you used a beat clock with the Axe Fx III? I haven't recently, but when I was it was jumpy by more than 1 BPM... Occasionally, there were brief but large jumps.
Click to expand...

Since I only use my AxeFX for recording, and not live, I don't use any tempo-based effects in the AxeFX. I'll apply those in the DAW if I need them. I find MIDI clock to be too imprecise for tempo-based effects. It can lead to problems like you're describing.
 
Last edited:
psimon666 said:
Remember that my tempo has not changed. It is fixed to 130 bpm.
The problem here, is that the glitch/pop, appears all the time (at fixed tempo)
Click to expand...
But it is changing on the Axe Fx due to how the Axe Fx interprets the midi beat clock or in how the clock is sent (I'm not sure which).
 
here’s the thing about computers and time:

it’s possible to be really precise by temperature controlling the crystals in the clock chips, but none of us are doing that. at best we’re synchronizing our system clocks to NTP servers, which means lots of tiny shifts.

using a constantly shifting, imprecise clock to control another in an attempt to be more precise is a recipe for frustration.

tldr: does it sound bad when you don’t sync the midi clock? plenty of fantastic music has been made over the years without precise synchronization
 
Last edited:
No problem when there is no synchronization with the MIDI clock (as said in my very first post of this thread).

A workaround for this problem would be to add an option concerning the MIDI Tempo interpretation speed :
  • FAST (as today)
  • SLOW : smoothes the Tempo variations (and so ignore local errors). In ly case, 130 will always remain 130

Another idea: An editable threshold in bpm : 0 bpm to 10 bpm (for instance). There, below this threshold, the amount of the Tempo variations are always ignored and the tempo keep the mean value
 
Budda said:
Why not just set the axe fx bpm to 130 and use that?
Click to expand...
That is equivalent to what i did (and explained in my very first post).
But this does not solve the bug at all. I meet this problem on all my songs, on several tempi.
Nevertheless, my case is just an illustration of the problem : The actual behavior is not expected and in other MultiFx, this problem does not occur.

My post is to describe the bug, not to find a workaround
 
The Axe-Fx III time stamps the clock message so if the tempo is changing that means your message time resolution is poor.
 
psimon666 said:
That is equivalent to what i did (and explained in my very first post).
But this does not solve the bug at all. I meet this problem on all my songs, on several tempi.
Nevertheless, my case is just an illustration of the problem : The actual behavior is not expected and in other MultiFx, this problem does not occur.

My post is to describe the bug, not to find a workaround
Click to expand...

Ah, gotcha.
 
MIDI beat clock has been problematic since its invention in the 70s. If you can insert a SysEx message in your DAW project, you can command the Axe-Fx to change precisely to any tempo you want. @Michael Pickens once created a spreadsheet of these messages. I'll see if he still has it.
 
The III and FM3 use a completely different message. I asked Michael.
 
Admin M@ said:
The III and FM3 use a completely different message. I asked Michael.
Click to expand...
I asked about this long ago (in the thread where I reported jumpiness that Cliff improved) but it wasn't available at the time.

Would love to have it...:)
 
I've noticed when I tap a tempo into the FM3 I can get strange delay artifacts, more audibly with the Plex and Multi delays than a straight up delay. All the talk in this thread is about the MIDI changing the Axe tempo slightly (sorry to highjack), but nothing was said about a way to make the delay deal with slight changes in a better way. Is this just not possible?
 
I just tested this to make sure everything is working properly and it is. However... it all depends on the accuracy of your DAW. If I set the BPM to 130 and the accuracy is 5ms it will occasionally go to 131 BPM on my system here. If I lower the accuracy to 2ms then it doesn't do that.

And this is as would be expected. 130 BPM is 2.167 beats per second. MIDI Clock is 24 messages per beat. So that's 52 messages per second or a message every 19.2 ms. 129 BPM translates to 19.38 messages per second. So your messages need to be accurate down to fractions of a ms.

So you need to find if there's a way to increase the MIDI clock accuracy in your DAW.
 
You must log in or register to reply here.
Back
Top Bottom