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

Sample Scanner


Yeah, that’s what I’ve been doing (see my post above), to clarify, the amplitude determines how much of the waveform is scanned each cycle!

The offset bit I am less sure on because the pitch of the saw is what determines the speed at which the section of the waveform you’re focussed on cycles and an offset at the point would effectively alter pitch!


On the other hand if @Joe is right, then a width of 1 would allow the full range of the incoming saw to be applied which would make sense. When I have turned the width all the way up it seems to work and this would take out some of the variability of what’s happening and firmly seat the responsibility of focus on the saw, which also makes sense!

I need to try it again :smiley:

edit: and yes, the fade parameter doesn’t seem to solve the interpolation as I originally hoped, I think this is just because it is only applied at the beginning as per @odevices diagram above.

edit 2: if i set width to be 1 then the gain settings on the saw output need to be super low to make sense, this is probably because I am using long single cycle waveform chains, they would be fine at higher values with shorter chains. The width parameter results in less waveform being played either way.

edit 3: I am finally getting some very nice morphing results, with the help of a filter to take the noise out a bit at the top end and a grid quantiser to ‘step’ the modulating sine wave on the phase it kinda jumps nicely between waveforms - feel really close now!


Very short video on dropbox showing a little morphing, even with the little bit of noise it generates I am pretty happy with this:

ER-301 as Wavetable


Nice. Found some sweet spots and quantized to them or sequenced them?


Not exactly, that was kinda my point, with the stepped modulation the sweet spots suddenly appear everywhere :smiley:


Here you go, this is my custom unit and the relevant sample file:

Lucky Pill :file_folder:

A1 is pitch
G1 is gate
B1 is sync for the phase modulating sin osc
C1 is an additional modulation input that I plugged a very very slow channel of Batumi into

p.s. this is definitely not an official custom-unit release hehe :slight_smile:


Hey, this sounds alright with something else going on, here’s a silly 2 minute manual morph using the phase, you can hear when I turn the dial a bit too quick as there’s a warble in the pitch :smiley:

The only thing that changes throughout is the phase parameter, everything else stays exactly the same.

Lucky Pill 2 Minute Morph :loud_sound:


Nice. Think I’m hearing the natural gate again too? :slight_smile:


Ooooh no, that’s the ER-301 ADSR Unit - actually, everything is ER-301 in this recording :slight_smile:

The terrible drums are just three sample players, only the 4 voices in total, sounds like more huh!

Seems the crossover between one waveform and another is okay if you get the cycles okay and filter the top end off :slight_smile:

The more I think about it the more I think the morphing works better because I have all the other parameters set ~correctly.


Finally got to v0.3.18 - so I was hoping to make a custom distortion with the following recipe:

  1. Make 64 waveforms that go from a ramp to a clipped ramp and concatenate them into a single file
  2. Load the concatenated sample into the Sample Scanner
  3. Use a combination of centre and width to be able to shift the window around the sample over one of the 64 waveforms
  4. Input audio and the modulate the width to access the different prepared waveforms that would increase in distortion/clipping as you go from 1 to 64 waveforms (I planned to make a bunch of variant waveforms for different distortion/saturation units).

After a bit of a play I realised that the width doesn’t work quite as I had envisaged with above recipe. i.e. the width & centre determines the window size and position of the internal sample and outside of that nothing is output.

Would it be possible to have a version of this unit where the input signal scans the entire width of the post-windowed internal sample?

If the above was available you could also make a rudimentary wavetable (minus the interpolation between waves).


Yeah… I went through the same thing, I believe there is a proper wavetable unit in the works!


Went through at experiment as well. Not really suited for a wave table. If you just use single cycle wave forms, it does make a very nice, flexible wave shaper though.


For your application, use the phase parameter instead of the center parameter and use a VCA or Limiter instead of the width parameter.


Thanks Joe, yes it does work well, I was hoping to develop the idea into having different flavours of distortion based on a suite of waveforms that I could dial in instead of having to load a new sample each time.

Hi Brian, if I understand you correctly is this what you mean (I’ll simplify the sample to having 2 waveforms for this discussion):

  1. I use the phase parameter to shift the 2 sample waveform along e.g. to select wave1 use 0.25 & 0.75 for wave 2
  2. And then limit the input amplitude by / N waves e.g. 0.5 so that the input signal can only scan between 0-0.5 (using a combination of vca & limiter to get it in that range) and then using the phase to change to the different wave?

For my 64 wave recipe I would need to reduce the input amplitude by /64 and have the phase move in discrete /64 steps which I was going to achieve using the quantiser & bump scanner. Make sense?


That sounds exactly right.


Thanks Brian, quite excited to get this patch built now :smiley_cat: unfortunately I’ll have to wait until this evening before I can try this out. Cheers, Phil


(please share it when you have it, sounds absolutely brilliant! :angel: )


Agreed :smiley:

I was definitely waaaaaaaaay too tired last night and think I misunderstood your post @vcoadsr and rereading this morning it seems like a completely different post!

Thanks @odevices for everything - as always!


Sounds like a good use for rational VCAs to be exact on the /64. Otherwise if you hit in between single cycles in the wave bank you’d neither be cross fading nor interpolating but using a different waveform entirely. Which I guess could result in happy accidents. Although seems like wave folding works best with specific waveforms, in my experience.

Edit: actually sounds like a good middle layer experiment. See if a waveform control can go from 0-63 in integer increments. Hmmm…


The Rational VCA only goes up to 1/32 at the moment so a regular VCA set to 0.015 should be more than close enough for this application.