#124

edit: holy hell, that clip was gorrrrrrrrgeous.

weird that the blips were at 30 and 60 sec

maybe? i can add a param to disable delay if it’s not being used, but the latest norns update cut CPU usage down with softcut :slight_smile:

lmk if u can reproduce the arc thing? unsure how u did it lol

edit: @ypxkap, OH I THINK I GET THE ALT THING U WERE DESCRIBING. is this it?:

  • you were entering a pattern into bank b
  • you also wanted to alt-modify the filter of bank a
  • as you held the alt for bank a mods, you were also tapping pads on bank b
  • this actually meant you were copying/pasting pads-to-pads on bank b instead of entering your pattern

yeah, feeling like a dedicated copy/paste button (or mode) would be best – alt can just be a global modifier as it is now. this way, we can avoid waiting for a thing to take effect, especially not a thing someone might accidentally trigger by lingering on a pad. dope suggestion to examine that behavior, thank you!

1 Like
#125

really hoping the next update is 1.1.1.1
:laughing:

5 Likes
#126

…this is pretty…

2 Likes
#127

I don’t have a Norns (or Arc) yet and this is making me want to get one (both).

Question: is there any way to get the multiple loops separated out, either by different audio outs (via a USB interface plugged into Norns?) or written to disc as separate files? The goal being that I could mix the different layers in a DAW later as I feel like I would need/want that to turn the ‘jams’ into ‘songs’ if that makes sense.

1 Like
#128

20 characters of loving this

2 Likes
#129

:heart_eyes: so beautiful

2 Likes
#130

Dan those envelopes are an amazing addition! Glad Marcus suggested them… Seriously what an amazing program… thank you so much for the inspiration I’m having in the studio…

2 Likes
#131

thanks for the kind words and inspiring tools!!!

1 Like
#132

there’s a way toward this, yeah! there’s a tape recorder built into norns which all audio routes back to. it records exactly what is sent to the stereo outputs.

depending on how you want to work, you can set up a few patterns on each bank and use alt + pause on cheat codes for two banks at a time – that way, one bank’s pattern plays and you record it, then you alt+pause that bank and alt+unpause another bank and keep recording. then, take the audio file from norns, which would have captured sequential plays of each of your patterns, and pull that into your DAW for splicing/layering.

typing this out, I just realized that you could feasibly hack two separate tracks with panning and recording to tape…so you could do a full two voice performance into tape and get the two separate voice’s clips by splitting the tape recording into L + R.

i’ll keep thinking on it :slight_smile:

dreams

i also have this crazy wish for the headphone jack to be a destination separate from the stereo outs, which would allow the three channels to be split, but i’m tired and that’s a little insane.

#133

Got it, thanks for this explanation!

Do you think it might eventually be possible to have a record function that records the 4 stereo tracks independently to disk? Would the hardware have enough CPU to deal with it? Or disk space? Sorry for my ignorance but I’m not sure what the hardest part of implementing that would be :wink:

And regarding the other approach, would Norns recognize a multichannel class-compliant USB audio interface? If there was a way to send the different loops out of different hardware outputs then that sounds like it could be an easier approach?

Thanks again!
B

#134

thanks for the report. if Tape is dropping audio frames, i doubt it has anything to do with the script. it would be bound by disk I/O and could (maybe) be addressed with double buffering.

Tape has a dedicated disk output thread. it is called on each audio block, and attempts to write all the frames in the audio buffer, before the next audio block arrives. adding an additional (larger) RAM buffer in between the audio process and the disk IO, could help by amortizing any disk output bottlenecks. but it could be fruitless if it this is simply about characteristics of the eemc.

if you want to open a github issue that would help me remember to revisit this when i rewrite Tape (which will be “pretty soon.”) any reproduceable details would help. (like, i wonder if some other system component is acccessing the disk when the dropouts occur.) if i were to try and reproduce this i would do so with sinewaves (making dropouts very obvious), split to a separate recording device… (to ensure that dropouts are not actually in the audio path before hitting the Tape.)

that is suggestive… is there some process writing to disk every 30s?

norns is just a linux computer. so, yes. but also, no, the norns software stack isn’t set up to understand multichannel hardware. you could use completely different software i guess.

IOW, i don’t think it makes a lot of sense to maintain the norns DSP layer to support arbitrary I/O channel counts, but you could ditch it all and just use it as a supercollider machine with a cute little video framebuffer. (i just use beefier machines for multichannel work, myself.)

i have a simple feature suggestion: a “save to disk” function that stops all record heads and writes each relevant regions of the softcut buffers. softcut API supports this, and it has a fresh new implementation which is real shiny and good, and i think it would be an easy win.

along those lines:

so that technique is working ok? i haven’t checked the cheatcodes codes. do you have to do weird tricks with metros or something?

assuming there could be issue with processing slew commands out of order, or dropping them or something, i did have an offer ready: i could easily implement a dedicated softcut command for, i dunno, play_env and rec_env, which just combines a slew and a level value and ensures they are executed in the correct sequence and in a timely manner with respect to each other. (really, very simple to do.)

not possible with norns hardware, the headphone output is just a gain stage.

6 Likes
#135

More Jamuary


I would love a way to ‘unlink’ the levels page from the pads. I was having trouble fading in and out :-
lovely work on the 1.1.1.1.1.1 btw Thanks!
6 Likes
#136

another installment…MANGL and Cheat Codes

3 Likes
#137

thanks for eyes, e.

no, and I haven’t hit this dropout thing in my own recordings :\

if I’m reading this right, that’s actually implemented already – if someone is looking to save the live audio that’s in the three buffers, there are these save live buffer triggers in PARAMS:

clipparams-M

but I think @Deru was asking to record each channel to its own audio file while they used the script. these triggers ^ would just export what audio’s in each of the three 8 sec. buffers, not record a two-minute pass of manipulating what’s in the buffers. @Deru, can you help clarify?

it’d also be ~possible to do this by sacrificing live input + the delays. could re-route the audio from the 3 buffer-manipulating voices into the 3 remaining voices and record into blank space. would only get you a few mins of recording, tho, the way everything’s set up.

re: envelopes…

yeah, i do. function and execution (630-645).

i mean, that would be really good – could all the level-y params actually get slewing? I’m currently using two voices as a stereo delay, which get “zippery” when the envelope function works on level_cut_cut. also, would be dope to be able to slew post_filters. was going to metro those as well.

thanks for voicing this – was also on the roadmap, but totallllllly forgot about it. you just want to be able to go to [levels], id an arbitrary pad (not the current playing/pressed), and raise the volume for that one pad, yeah? cuz you can also use ALT+change level to fade in an entire bank, but I’m guessing more granular control would be desirable?

thank you for sharing these @ESC + @SPIKE – sounding raaaaaaad

#138

Interesting, thanks for the info. Out of curiously, is it somewhat easy to take code written for Norns and run it on Supercollider on a Mac? Is there much translation that needs to take place? Especially if I’d like to use a Grid and Arc?

#139

Yes, I was hoping there might be a way to record the performance of the different loops/channels, not the contents of the buffers.

This sounds interesting. The sacrifice of inputs for outputs when you were ready to record seems like it might be okay, but the ‘few mins’ of recording time could be a bummer… Anyway, thanks for thinking this out with me!

1 Like
#140

I think a 6 track recorder engine that writes to a folder would be a pretty easy win and useful for many softcut scripts

actually nope, current routings can’t do it - but maybe that’s a possible future

but also like many I personally find the two-track nature of norns a helpful limitation (i.e. it can be helpful to commit, and “mix” later with eq and compression rather than seperated stems)

yep ! I was wondering where that param was. I wanted to make long glides when I switched pads with different cutoff values

2 Likes
#141

Ah too bad. That sounds perfect. Maybe in the future then :wink:

Noted.

1 Like
#142

Derp i forgot that was a new function, i think thats exactly what i was looking for!

1 Like
#143

eh, just newly mentioned :wink:

there are a lot of sleeping beasts hiding in those codes