Mellotron Samples

I searched the forum but can’t find any other posts on the subject. Has anyone tried to create a custom mellotron unit? Seems like the 301 would be rather well suited for this: set up a custom unit and then drop in whichever mellotron sample one wants to use i.e. strings, choir, vibraphone…I would love to contribute by trying to build something like this up. Any hints as to where I may find some good mellotronish samples?

May have found some good ones…amazing what you find when you google! (punches self in groin)


i tried once and for some reason wasn’t happy with the results.

if you come up with something good and feel like sharing, i’d love to take it for a spin!


I found some free sample packs with each note sampled separately from A2-A6, so i’m thinking the best would be to create a lookup table of sorts…going to have to do some research on that. I think there are a lot of Neil and Joe videos in my future :thinking:


There is an Organelle Patches web site ( where people upload Pure Data custom units for this device. There are two Mellotron PD files with a bunch of samples attached in WAV format for each key for free …

1 Like

Background historical information and current how to:

1 Like

Is this unit still alive on 0.4xxxx ? Best

No, if you read to the bottom of the first post there’s an edit that says it is no longer necessary!

1 Like

Awesome. Totally forgot about this…Hopefully I’ll have to do some fiddling around this weekend…:nerd_face:

OK, I must admit I’m relatively new to all this stuff. I spent the weekend reading through the old posts, including the ER-101/301 voltages post and the 88 piano sample custom unit…I guess I’m getting lost between what had been workarounds prior to updated firmware fixes like the 12TET slice selection sample player fix etc…also, I’m not sure what I should be using i.e. variable speed player or raw player…also getting confused If what I want is really to implement a bump scanner situation…

what I tried was create one sample file with an octave of mellotron single note samples from A2-G#3. I then sliced that one file so that each slice corresponded with one of those notes…I set the slice selection type to 12TET and the play option to “play slice”. all of this was done in the variable speed player…

My goal was to use my keystep v/oct out to control…I hooked up gate out to gate in of the varispeed player, and the v/oct output to control the slice select parameter…which is where I get lost. messed about with the gain settings and bias settings but couldn’t seem to get it right so that a consecutive key would play the next slice. if I’m using a v/oct, and I have 12 slices, shouldnt it be .083V step between each slice? do I set that as the gain, or the bias?

I suppose the easiest would be to just use a one note sample, no slices, and just use the keystep to control the v/oct setting on the varispeed player, but that doesnt seem like fun at all…

thanks in advance for showing me the error of my ways…:exploding_head:

Did you find your way to this thread? It might help!

If you read through that thread all should become clear - well maybe - it’s something that can be a challenge to wrap your head around at first.

Also, please note as of 0.3.06

  • ENHANCED: -10V to 10V are now mapped to -1 to 1 (previously -10.24V to 10.24V was mapped to -1 to 1).

If you think about it you have 12 slices mapped across the range 0 to 1 inside the ER-301, this means each slice has 1/12 as an activation range which as you say is 0.083, however you are using one octave on the key step, right? Which presumably will be something like 2V to 3V.

The answer to the question is how do you get a voltage range of 2V to 3V (is that right?) to map to 0 to 1 in the ER-301, and the answer is make 2V map to 0 and 3V map to 1 using the gain and the bias.

I feel I should leave it there, because you are so close and if you work it out for yourself I think it will mean much more.

If it’s too much, Neil Parfitt did a great video that showed how to map these ranges without doing the maths, I have no idea which one it was, maybe someone else can remember?

I am doing this from memory and haven’t checked it on my ER-301 so forgive me if I have missed anything or not got it quite right.

thanks for the quick response…let me see if i can figure it out…:thinking:

Since you are using the 12TET slice addressing mode you should use a gain of 1. Typically bias would be zero.

Which octave on the key step would you play with those settings? It’s just 12 samples and they should map to a consecutive run of keys through an octave on the key step (I think!).

Or… have I had a brain fail?

I thought so. This is where Im getting a bit lost. Unlike the 88 key use case, I only have 12 slices, which I’m assuming means that I should only need to travel a total of 12 semitones. I’m starting on the keystep at A2 (which should be 2 volts) so I’ve set a -2V offset after my v/oct input into the slice select parameter so that A2 out of the keystep will generate 0V input into the slice select parameter, or slice 0 (which I’ve setup to be A2 in the sample). I thought my gain should be one, so that when I hit A2# on the keystep the total voltage will be 0+0.083. My assumption (which is where I may be wrong) is that it will select the next slice in my sample (which is A2#). I have also, as mentioned in the past posts tried a half a semitone offset, so that I dont start at the very beginning of a slice, but that didnt seem to fix it…

As Brian said on the other posts, working through this is definitely deepening my understanding of synthesis and control voltages…or lack thereof :laughing:

EDIT: Am I way over thinking this? could it be as simple as setting the gain to 0.1?

Points that are easy to make a mistake:

  • How do you know that the keystep outputs 2V when A2 is pressed?
  • How did set a -2V offset? I hope you used -0.2.

No. The gain should definitely be 1.0.

1 Like

No you have it right I think. The problem is we need to know where on the keystep is mapped to 0V. The first slice would start there. And just like you suggest, if you want to shift the slices to a different location on the keyboard, you would mess with the bias. Changing the gain would destroy the V/oct relationship, so that has to stay at 1.


As if there was any doubt, Brian is totally correct. I haven’t figured it all out yet, but I did have time to plug my keystep into O’Toole which shows that the keystep maps 0V at C0…therefore my assumptions were off…keeping the gain at 1 and bias at 0 on the 301 and starting at C0 on the keystep, I was able to get each semitone jump on the keystep to correspond with advancing to the next slice on my 2 oct sample…now I just need to work through how to set the bias so that instead of C0 selecting slice 0 on my sample (which plays an A2), the voltage of A2 on the keystep selects slice 0. I’m thinking I’ll have to set a negative bias, but sadly no time to test it out tonight…stay tuned for the conclusion of this riveting drama!