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

Is it possible to record into the same buffer as you are playing without any silence or glitch when looping?


Is it possible to record into the same buffer as you are playing without any silence or glitch when recording is over or the sample start over again in loop mode?
This is a problem with octatrack and was my main reason for selling it.
So now Im eying the 301.
Sorry if this has been discussed before but I cant find it.
Thank you



Please see this thread: Looper and sample player unit behavior with external audio input


Thank you!


Yes, as many as you like.

I would caveat that the shared buffer thing should be fine as long as you have the play and record heads moving at the same speed. If the play head crosses the record head (e.g. because it is moving faster) then you will likely get some artifacts (which makes sense if you think about it).

I know that @odevices has been thinking about ways to potentially make this work even with different playback speeds. I have no idea what kind of magic it would take to pull this off, but I am hopeful.


Thank you. Sorry for the sub q deleting but I found out.
If there allready is a sample in the buffer there will no artefacts?


It should be fine. Can you describe in a little more detail what you’d be doing?


ok, lets say you have 30 sec buffer and you want to have lets say 4 play heads and 4 recording heads running inside this 30 sec buffeer zone. You beeing able to move around both player head and recorder head.
Like the program RoSa from Steim http://steim.org/product/rosa/ can do.


as soon as the record head collides with the play head you will hear a short click, so by changing speed on the play head or the position of that you will run into this problem, it has been discussed in this thread:

brian spoke about a possible solution to this, so fingers crossed that he finds a solution!

Newbie issues with clicks when using Feedback Looper and Manual Grains

I’m not really sure I fully understand Rosa without diving further in. But it’s pretty much like @kilchhofer says. If you can prevent a moving play head from colliding with a moving record head in the buffer, you should be good. If they collide, you’re very likely to get some kind of artifact or unwanted noise as of the current firmware.


But you can’t record into a buffer and use that buffer for exact convolution of the incoming audio.


That is correct because the raw samples are not used for the convolution (as you would in a naive convolution). The Exact Convolution unit uses the much more efficient Uniform Partitioned Overlap-Save algo which gets its computational savings from pre-computing (and re-using) the FFTs on the impulse response.


Understood. I woke up in the middle of the night with idea for a patch only to realize it isn’t possible (yet?).
How long does the pre fft take?


Im sorry, I dont know the lingo but the way RoSa works is that there is an empty timeline that you record into and play from in whatever place you want. I guess that is not how the ER 301 works?
Thank you


If you can point out which part of Joe’s or kilchhofer’s answers that you did not understand, then perhaps I can help provide more details. If you keep asking “Does the ER-301 work like RoSa?” I will not be able to give good answers because I have never used RoSa. I don’t think the others who have answered so far have used it either.

Oh and I have read the RoSa webpage and manual. Unfortunately, there is not enough detail to say anything other than what Joe and kilchhofer have already said.


Its ok, i just wrote down the way RoSa works from a user point of view. Im not a programmer or anything so I dont understand anything about convolution/FFT/Uniform Partitioned Overlap-Save algo bla bla. That is why I asked. Sorry to bother you
Take care.


I am not bothered :wink: You are misunderstanding the conversation here.

You should ignore auxren’s statement. That has nothing to do with your question.


you can record into a buffer and play that same buffer at the same time, with a granular player or a normal sample player, you just attach the same buffer to both, recorder and player. you can then do everything you want, have multiple players play at different speed or pitch, different spots in the buffer, and at the same time record into your buffer in real time, that’s all possible but again, as soon as the play- and recordhead collide, you will hear a click, i guess that’s what you would call a glitch. so to answer your question with regard of the glitch part, no, it is NOT possible without that artefact.


Great, thank you!


@odevices sorry if u answered this somewhere already, but is the is this ring buffer mentioned in the other thread still in the works? or planned to be?




Definitely still in the works!