Home | ER-101 | ER-102 | ER-301 | Wiki | Contact

V0.3.x Firmware Workout: Better late than never

feedback

#961

I’m getting pretty severe CPU spikes in Clocked Stretch as I increase clock rate. Is it fair to say that the CPU usage will be related to file size in this unit?


#962

v0.3.20 CHANGES

  • FIXED: Loading a sample chain with a mix of mono and stereo files would yield reversed slices. (@giftculture)
  • FIXED: Manual Loops > Bias fader range was 0 to 1 when it should have been -1 to 1.
  • ENHANCED: Manual Loops > added a parameter called max.len which sets the loop length (in seconds) that corresponds to a value of 100% on the (loop) len parameter.
  • ENHANCED: Manual Loops > Added contextual waveform view.
  • ENHANCED: Manual Loops > Added (sample) edit command to the unit menu.

#963

What do you mean by severe?

I suppose at some point as the clock gets higher and higher, every grain is going to be in a completely random part of the file. Memory performance will suffer somewhat because L1 and L2 cache hits are constantly missing.


#964

Haha. I think I’m misunderstanding the conditions of the bug also. Let’s focus first on the Spread Delay unit. I’m scrolling the fdbk parameter back and forth and I can’t get it to change to “nandB”. I’ve even tried modulating it with 10x white noise. May I ask you to describe the patch in more detail? Perhaps it has something to do with what you are feeding to the input of the delay…:thinking:


#965

23% to 93% with a few mixers and 1 instance of edit: sorry Clocked Stretch as I increase clock rate.

The sample was large (240 MB) and the CPU spikes reduced on replacing it with a smaller sample.


#966

48kHz or 96kHz firmware? How fast was the clock (in Hertz)?


#967

48kHz, just trying to test now clock speed at which it’s happening. I’ve replaced the smaller sample with the larger and it’s pretty reproducible, but I’ve been watching the CPU meter, not the clock speed…was trying to recall how to take a screen shot to show you the meter.

If I dial the clock speed through a variety of rates i get small needle like spikes on CPU, but if it sits around 40 hz. Oddly though (perhaps because of the CPU load?) the clock speed is bouncing around 35-50hz. The clock source is a LFO out of a Mother 32 so should be stable.


#968

This helps a lot. Thank you. One more question. What are your settings for the following?

  • Play Duration
  • Play Extent
  • Slice Polarity

That’s a known problem with the frequency display. Don’t worry it is not your clock.


#969

I actually don’t have any of those settings?!?


#970

They are in the unit menu. Well I guess that means you haven’t changed them.


#971

No, have not, but will investigate. Is there anything else I could make clearer?


#972

No. I think I have the full picture now. Investigation has begun!


#973

once, slice and symetric…for the record.


#974

ok here you go, there’s no patch, just a stereo chain with a mixer and here i’m adding the spread delay:

i can scroll to nandB and it instantly stops working with occasional monsternoiseblast.


#975

just so i understand this all correctly and i don’t appear as a fool, this should not happen, right?
i’m a bit confused because of this:

so i checked all delays and this happens with every feedback parameter and the pedal looper starts with feedback as nandB:


sorry, video orientation should change soon…

#976

Wow! Thanks so much!


#977

Absolutely not a fool :astonished: The fool would be me because initially I thought you were saying that you had a case of runaway feedback where the signal amplitude just kept growing and growing. AFAIK there are 2 camps of people, one side wants their feedback to be tamed so that there is no way that it can blow up, and the other side wants full control even if it means danger of an unstable feedback loop. So I wanted to know what side you were on.

However, it turns out this is not the issue at all, or at least I don’t think so based on your (much appreciated) video. In fact, I’m quite perplexed by this one because no matter what I do I can not get the behavior you are getting. I wonder what I’m missing…we are on the same firmware, right? I see that you have your Spread Delay unit inside a mixer being fed by some signal. Does the same thing happen when you just insert at the top-level with no signal coming in?


#978

Actually, I noticed that in your Spread Delay video, as soon as you insert the Spread Delay unit, it is already weird: fdbk parameter says -36dB but this should never happen, any value below -35.9dB should show as -inf dB. Something is very weird…may I suggest first re-downloading the firmware zip file and updating your ER-301?

Also, you haven’t, for example copied, your own version of the builtin units into the ER-301/libs folder right?


#979

hey brian, i updated from 0.3.19 to 0.3.20 just now and it seems the nandB doesnt show up anymore, however my spread delay still starts with -36dB, actually every delay with feedback parameter.
pedal looper looks fine now. and no, i haven’t manually copied anything to folders.


#980

The combination of the new Clocked Grains unit and the ability to select multiple files that automatically get sliced is a really big step in improving beat-oriented workflows on the ER-301.

I have a couple ideas of what might make the clocked players even more powerful. I think it’d be really awesome to have the option to add some automation to slice switching. Two things jump out to me…

One: a mode where when the slice index has changed and the currently looping splice finishes, have it automatically switch to the slice that the slice parameter currently points to. I think this would make it a lot more playable as far as switching between looping samples so you don’t need a perfectly timed-retrigger to have sync’d changes.

Two: even cooler would be if changing the slice while looping would jump to the next slice instantly, but have the playhead in the same relative position, taking into consideration the clock-sync settings. I realize this would probably require some kind of microfade to avoid clicks, but it would open up a way of stacking a bunch of synced loops into slices and easily move between them.

Anyway, I’m sure I could rig something together in teletype to handle #1, but I think it’d be wonderful out of the box. It could also be easily implemented with that oft-requested end-of-slice trigger :wink:

Edit: Hmm… upon further consideration, it seems like just feeding the same clock to the trigger is an adequate solve for the first idea so disregard that !