DigiChain - web app for making sample chains

For the Rytm, the number of slices per chain must be divisible by 120, and as @orionka said you will also need to use the ‘spaced chain’ which will ensure all the slices are the same length.

E.g. for a gap of 8 on the sample start/end, you would get 15 samples in a chain.

You will see, the default options don’t have 15, so, Ctrl+Click on any one of the numbered options to enter your own custom number, [15], then create spaced chains.

The Rytm (and model:samples) don’t have a slice grid, so a bit of math is necessary unfortunately.

And there’s a list of all the keyboard shortcuts here.

1 Like

Thanks for the heads up! Will be the first thing I check out tomorrow

1 Like

@phaelam just moving this over here so as not to clutter the Octachainer thread - could you try buidling the same chain but using Chrome? (or, if you don’t mind doing so, send me a zip of the source audio files, and the outputted chain file so I can can do some testing - I will delete anything sent, and won’t use any of the audio files other than for trying to reproduce the issue).

Thanks!

1 Like

No, I won’t install other browsers. Chrome is brutal.

It would be extremely helpful if you could see the slices in the slice editing mode on the OT.

You said it happens more on bass frequencies, so I’ll do some testing with bass hits later this evening, see if I can get any clicks/pops to appear.

Are you by any chance doing sample-rate conversions? (e.g. going from 48k to 44.1k), it’s possibly rounding errors when that happens (which would also explain seeing it in Octachainer).

hey man…I was looking at the .WAVs in Audition, and tho I had normalize and peak off, the WAVs were pretty hot. so I just dialed down the volume in Audtion and chucked that into the OT, and it seems to be good so far. ill try the same with some other sample chains.

re: the second paragraph in your response, ya, the original samples were 48K
and now that you’d mentioned that, the set I JUST did that was successful was 44.1K initially. SO…with that im going to take the same samples that failed, and batch process them to change to 44.1K. super fast to do vs hand chopping 64 samples.

so ya, perhaps theres issues with 48 to 44.1

thanks for follow up :+1:t6:

1 Like

ya so after a quick runthru…I took that same sample set, batch converted to 44.1K 24bit and ran thru the digichain…much better :+1:t6:

no need to adjust volume

1 Like

I did some work on this last night, it looks like the native web audio decodeAudioData does a poor job going from 48 -> 44.1 on low frequencies, so I’ve made some changes to resolve that, and also added a declick on the end stage wav generation (configurable in the settings panel).

I have some more testing to do around this, as I still need to fire up my MacBook to test my work in Safari. But I should be able to get this fix out over the weekend.

2 Likes

MUCH appreciated :+1:t6:

1 Like

Just pushed up v.1.4.4 update to the main URL:

  • Minimizes reported click/pops on exported chains when importing wav files of differing sample rate to the target sample rate, or when using audio interfaces configured with small sample buffers. On by default at a 40% threshold, user configurable thresholds and can be disabled via settings panel.
  • Added fuzz selected actions action to add fuzz/noise into the selected samples.
  • Added fade-in fade-out selected actions to fade in/out the first/last 256 samples of each selected sample.
  • Blend panel to blend/interpolate between the selected samples (works best when the samples are similar in duration).

Which means a few more options in the settings panel for the de-clicking.

The ‘selected actions’ options got a move around to accommodate a few additions:

image

And I pushed up the new Blend panel that I had been noodling with for a while, this blends between the selected samples by the total number of steps chosen. Use this with single-cycle files to create wavetable-ish sounds, files generated by blend will (should, it is experimental!) loop, as the first also blends into the last - so you can select more than just two files to blend between, and they will blend in the order they are in the list.

e.g.








While this will work with any samples, it works best with smaller length files - and the buffers can get quite large as its the number of steps chosen * the length of the files.

Another example with three trimmed pitched vocal stabs.

8 Likes

This is such a nice app.
One wish: Is there any way to finetune individual samples in the list by cents? I’d like to make some microtonal multisampled synths to be played with the Slice machine.

2 Likes

Thanks!
At the moment, the editor only adjusts by semitone/octave. I’ll take a look at adjustment by cents and see where it would fit in on the panel.

2 Likes

That would be really cool.
Thank you so much for the immediate reply!

1 Like

One request: Please implement a select-all/deselect-all function unless this is already implemented. I potentially have thousands of one-shot samples I might need to select or deselect en masse and have not yet found this button. This would save a lot of time for the user.

Thank you, Brian. And thanks again for that randomize function. I finally got my laptop fixed and about to scramble-up my OT sample chains. Wish me luck.

And yes, I feel selfish as hell asking for something for free from a guy that lives a whole ocean away from me but, what the hell. Can’t hurt to ask.

1 Like

Sorry, I know some of the shortcuts aren’t obvious :slight_smile:

CTRL+Click the Selected row header to toggle the selection of all samples.

The OFF button will also deselect all the samples in the list, and the i key inverts the current selection, so that also selects all when there are no samples currently selected.

There’s a full list of supported keyboard shortcuts here.

The import is limited to 750 files, if you want to drop in more than that, there’s a setting in the settings panel to turn this limit off, but it may cause timeouts on computers with low available memory.

And when creating files for the OT, make sure you start off in a 44.1khz export option from the drop-down, and have the .ot file generation turned on in the settings panel :smile:

2 Likes

You are awesome before I realize it, always. You have a friend here in Chicago named Dicky. You’re welcome. (And thanks again, of course.)

1 Like

No problem, and thanks!

You say you live a whole ocean away? There’s no ocean between Illinois and Pennsylvania :upside_down_face:

Why did I think you were in England? Hm. I’m not sure exactly.

There probably should be

I’m from England, just haven’t lived there since 2020 :slight_smile: I must still be typing with an English accent!

1 Like