Digital Noise Source?

Digi001 · 2015-02-23T16:33:45.000Z

Just curious, why do you think they used a Digital Noise source instead of analog?

So it is a sample of noise rather than just real noise?

Wouldn’t it be easier to just create an analog circuit for this? Or possibly this negatively coupled noise into other sections?

Gummi · 2015-02-23T19:09:27.000Z

its not a sample
its much easier to do digital
its trivial soundwise

R7 · 2015-02-23T19:33:22.000Z

It’s an interesting topic because the only issue I really have with digital sound sources is aliasing. Which isn’t an issue if the sample rate or oversampling are high enough, or mathmatic multiples to the fundamental frequencies of the note like the Modal 002, and maybe the PPG Wave.

Aliasing will sound bad if the foldback frequencies are not mathmatic multiples of the fundamental frequency. Then you get all this inharmonic garbage frequencies that sound generally awful. They can sound interesting and FM like since you are busying up the frequency spectrum but it depends on what note/frequency your playing and some notes will sound uglier because they are more obscure and inharmonic ratios to the particular note you could be playing. That’s because aliasing is always calculated at the Nyquist rate and stays constant in a fixed sample rate like 44.1 or 48khz, and is easily audible at least with notes in the higher registers or with alot of harmonic content since they are sending information up over the Nyquist frequency. It’s generally folded back in a frequency that is too high to be audible in high sampling rates like 96khz, which is as far as I’m concerned the one real benefit of working at a higher sample rate.

However, in the case of white noise, aliasing really shouldn’t matter. The frequency spectrum is already filled out completely, and nothing is really tonal, or has a fundamental frequency or harmonics of that frequency. It’s just all frequencies. So even if the white noise is aliasing you shouldn’t be able to hear it because those frequencies that would be aliasing artifacts in a less busy signal, are already present in the filled out white noise spectrum.

So it shouldn’t matter much. At least in the Rytm. The A4 has those C64 noise samples or whatever they are and they tend to be a bit more tonal but again, noisy so a bit more noise shouldn’t matter much.

Especially in white noise, noise is noise.

Digi001 · 2015-02-23T20:17:09.000Z

Ah ok. I understand now.

I was reading the FAQ on the webpage and for some reason thought they were using a sample when they said digital.

But basically they have a random number sequencer feeding a DAQ? Rather than actual white noise from a resistor or something?

Diayse · 2015-02-23T23:42:38.000Z

One thing I think I noticed is that the noise source has very little low frequency. Either the source is filtered out, or there is an individual filtering for each instrument.

In most situations, it’s probably best not to fill the low end of the spectrum with white noise, but for instance I couldn’t get a LPF filtered noise kick drum.

R7 · 2015-02-24T03:49:22.000Z

""Digi001"":

Ah ok. I understand now.

I was reading the FAQ on the webpage and for some reason thought they were using a sample when they said digital.

But basically they have a random number sequencer feeding a DAQ? Rather than actual white noise from a resistor or something?

The FAQ on elektron.com says it’s a digital noise generator, so I assume it’s not a sample. I had to look up how you would generate white noise digitally and wiki says “White noise may be generated digitally with a digital signal processor, microprocessor, or microcontroller . Generating white noise typically entails feeding an appropriate stream of random numbers to a digital-to-analog converter[url=“Digital-to-analog converter - Wikipedia”]. The quality of the white noise will depend on the quality of the algorithm used.”

So yeah that’s probably what is happening in the Rytm and the A4. They do have a lot of digital processing going on so I would think it would be easy enough to implement.

However if you did use a sample, and it was actually perfect white noise and looped properly, I don’t think you’d really be able to tell the difference between it and a generated one. They would both contain all frequencies and equal amplitude and because of this would avoid the general limitation of sample playback engines, the length of the sample becoming shorter as the pitch is raised, because white noise has no particular pitch and contains the whole audible spectrum.

R7 · 2015-02-24T03:58:43.000Z

""Diayse"":

One thing I think I noticed is that the noise source has very little low frequency. Either the source is filtered out, or there is an individual filtering for each instrument.

In most situations, it’s probably best not to fill the low end of the spectrum with white noise, but for instance I couldn’t get a LPF filtered noise kick drum.

Well now I’ve got to try and synthesize a white noise kick drum, and plug the Rytm into a spectrum analyzer and see whats going on.

Do all the engines do it and in all circumstances? From a new track, new sound and different filter settings? I could see why Elektron would cut out the low end on white noise because too much sub bass can definitely compete with some important frequencies the drum machine would be using.

Don’t know why I’ve never made a white noise kick drum, that’s a good idea! I do it all the time with white noise and a bandpass filter to make snares and hi hate type sounds.