Old RingMod device (randomly) breaks after a while

This instrument is made using the good old tuned ringmod tricks with the older version of ringmod and it doesn’t seem to behave exactly like before.

I’ll describe the problem:

I open this instrument, play with it for a while, turning macros and playing some keys, then it just starts to sound worse and worse until the main synthesis source turns into only awful DC signals.

To get it back to normal i have to save the instrument in the current ‘buggy’ state and load it back up, then it sounds just fine again.

Here’s the instrument:http://forum.renoise.com/index.php/files/file/398-buggy-cymbal-synth/

I can confirm that I’ve had this problem with the legacy ringmod. I haven’t found anything that triggers it, so it just seems that it just quits after some amount of time. I was using it on a simple track that just had a snare sample, and every time the device failed I had to copy it, and delete the old one.

I wonder if this is being looked into or not, i understand if you don’t prioritize depricated DSPs, but it kinda breaks the backwards compatibility for a lot of my work. So can you please tell me if you are going to spend time on this or if i should just move on?

I’ll bump this untill i get an answer. It shouldn’t be to hard to just say yes or no to this?

I have made lots of resources using the old ringmod and i really want to know if there is any point in developing these any further or if they should be scrapped.

I played around with your instrument for a short time but I didn’t manage to trigger anything particularly horrible yet. Then again, it’s quite a strange sound to begin with, so I’m not totally sure what I’m supposed to be looking/listening out for here. I definitely didn’t manage to make it produce any horrible DC signals so far, it simply always sounded like a weird noisy cymbal-ish thing (that eats a shitload of CPU, by the way!).

play with it for a while, turning macros and playing some keys

How long is a while?

Which exact macros are you tweaking, and how?

Are there any specific macros or values that seem to make it turn bad faster?

Any keys in particular that you’re pressing?

Anything else you can think of that’s worth mentioning?

Do you perhaps have the ability to do a screen recording of you triggering the problem live?

PS. Can you manage to recreate the problem with a much simpler set up? Maybe just a single RingMod device that reliably goes bad under much simpler testing conditions? That’d really help to remove some of the random factors here, and make our lives a bit easier.

Try to turn the “transient” slider to 0.00 because it is a sample that is not touched by a ringmod. Then write a note in the pattern editor and hit play. After a while (5-10 minutes or so i guess) all you hear are small clicks as in DC noise.

After this the instrument sounds useless, but if you duplicate it, the new one sounds just fine.

(that eats a shitload of CPU, by the way!).

No, no, you need a new computer. :smiley:

After a while all you hear are small clicks as in DC noise.
After this the instrument sounds useless

~~Once again, how long is a ‘while’? After a few seconds? A few minutes? An hour?

Any particular notes? A C-3, a C-4, something else?

Is it something really obvious that sounds terrible, like a night and day difference, or is it more subtle?

If it’s really obvious and is supposedly easy to replicate, then I’m sorry to report that I simply didn’t manage to witness it so far.

So I think you better make an XRNS for us to test, to hopefully replicate the exact scenario you have over there.

Please also mention exactly which audio settings you have: sample rate, latency, and so on.

If you can also render a good vs bad audio example, just to be extra sure, that’d be great.~~

Edit: Nevermind, I finally managed to make it happen. Will try to investigate further. Cheers.

Edit: Nevermind, I finally managed to make it happen. Will try to investigate further. Cheers.

I just figured it out myself what to do to recreate it, so i didn’t know how long it would take, but seems like it starts happening after 5-10 minutes of continuous playing, like one ringmod cuts out after an other just all of a sudden.

As far as I can tell, the DC crap comes from the trigger samples themselves simply playing in their raw un-modulated form.

At some point, for some unknown reason, the RingMod device gradually stops modulating the DC signal into an actual waveform, until it’s finally just spitting out the flat-line DC from the sample.

So it may be that the RingMod’s amount parameter somehow gets borked, or maybe it’s related to the phase, or who knows.

It can at least replicated after a wild bit of random parameter spamming, so it should be possible to track down what’s going on.

Cheers.

As far as I can tell, the DC crap comes from the trigger samples themselves simply playing in their raw un-modulated form.

It’s a lovely method you invented with your tuned devices. :smiley:

Thanks dblue, i made so much stuff based on your tuned ringmod, so it’s a bit sad to let the thing go to waste. :slight_smile:

It can at least replicated after a wild bit of random parameter spamming, so it should be possible to track down what’s going on.

You don’t need wild spamming of any sorts. Simply write one note at the beginning of a 64 line pattern, press space and let it loop for 10 minutes.

Sit back, relax with a cup of coffee and listen to the amazing cymbal sound loop for 10 minutes, i know you can do it. :stuck_out_tongue:

Don’t know if it is related, but I found somthing funny, too. Might be related.

Steps to reproduce:

0.) Fire up renoise, standard default song

1.) Turn off your audio. You not want to hear this. Better just watch the effect in the scopes view.

2.) add old ringmod device to channel one, set amount to 100% and freq to max, phase to 0% (all other settings could work, too)

3.) create 1 sample loop in an empty instrument, use the drawing tool to make sample maximum dc

4.) add adhsr device to the standard modulation, set release to 60 seconds (maximum), maybe sustain to max also to accelerate the “effect”

5.) while watching the scope add c-4 notes of the dc hold instrument repeatedly to the channel with the ringmod (I did with a midi keyboard, maybe this matters as for the notes playing on in release-held mode)

at my first and only try on this, the ringmod got fucked just before the standard 0x40 line pattern was filled with notes - it would not produce full amplitude notes anymore, but strange low volume stuff I didn’t dare to listen to, maybe the coolest glitches of the world, but maybe the most ear destroying, who knows, I don’t…deactivating the ringmod showed that the dc instrument built up an enourmous dc beyond the scopes range. Heh, yeah, it maybe broke the internal state of that instance of the ringmod plugin. Maybe it will simply get an overflow somewhere internally with too large signal values (floating point numbers can climb the nuts, almost limitless if not capped, and very large numbers might break one or another calculation with rounding errors as 23 bits precision is the same regardless of the number scaling, I know this pain…).

And I’ve seen the cymbal synth here has release active and isn’t in mono mode, so it might be the same breakage, but just happening a lot slower because the dc buildup takes more time.