"Velocity-sensitive" Rhodes sample for Stage 2 and Wave

[karri]

This is basically a follow-up to this post: http://www.norduserforum.com/post21078.html#p21078. After experimenting with JaZzMaN251's Rhodes sample, I got the idea that the velocity-controlled filter in the Stage 2 could produce much better results if the sample was pre-EQed to neutralize the filter at some medium velocity. So I analyzed the frequency response of the LP12 filter with a given resonance, and created some EQ curves in Audacity that are (more or less) the inverse of that frequency response. (I needed lots of different curves because of keyboard tracking.) The idea is that when playing normally, the filtered sample should sound somewhat close to the un-modified sample without filtering, and when playing harder or softer, the sound should change appropriately.

Then I applied these EQ curves to JaZzMaN251's recordings, which he had kindly sent me, and after a lot of trial and error, as well as some improvements to the original sample, here is the result:

• Rhodes MKII LP12.mp3

(The demo was recorded using slot B, which includes effects as well as an additional LFO-based tremolo.)

Of course, this is not a substitute for true velocity layers, but at least it is highly tweakable. For example, to adjust dynamics, change the base filter frequency and mod envelope amount.

If someone wants to try this on a Nord Wave, the settings are:
Filter: LP12
Keyboard Tracking: On
Resonance: 3.0
Base filter frequency: 35 Hz
Mod env amount (for filter): 5.6
Mod env attack time: 0.5ms (minimum)
Mod env decay time: 8.57s
Mod env release time: 110ms (doesn't really matter)
Amp env attack time: 0.5ms (minimum)
Amp env decay time: 7.68s
Amp env release time: 10ms
Both mod env and amp env controlled by velocity

Have fun!

[Frantz]

Thank you for sharing your work FunKey
I'd like to learn more about what you did, not sure I understood well, did you apply the eq curves on the wav source files (not the nsmp) ?

So I analyzed the frequency response of the LP12 filter with a given resonance, and created some EQ curves in Audacity that are (more or less) the inverse of that frequency response. (I needed lots of different curves because of keyboard tracking.)

How many curves accross the keyboard were needed ?
Why not just only deactivate "keyboard tracking" then ?
Cheers,
Frantz.

[FunKey]

Yes, I applied the EQ curves to the WAV file. (The sample happened to be a single file, but that doesn't make any difference.) The curve is different for each sampled note, and there 13 such notes, although I don't use two of them because they are too different from the rest.

Turning keyboard tracking on or off was part of the trial and error phase. Basically, a fixed filter frequency (except for the velocity responsiveness, i.e. no keyboard tracking) creates problems for high notes, whereas keyboard tracking creates problems for low notes. Without keyboard tracking, if you want the filter to close itself enough for softly played notes to become noticeably quieter, you have to choose a base frequency that is actually lower than the fundamental of the highest notes. So when you apply the EQ to the WAV file to compensate for the filter, these notes are amplified so much that you have to reduce the volume of the entire sample by more than the maximum gain of 9dB which can be applied in the sample editor. It also means that these notes vary between really quiet and really loud, whereas low notes are affected very little by velocity. In contrast, with keyboard tracking, in theory all notes should be affected equally, but in practice lower notes have stronger harmonics, and if the filter reduces their volume too much, there is no chance to compensate for that. All in all, keyboard tracking turned out to be the lesser evil, if the "normal" frequency (the one that the EQ is tuned to) is chosen sufficiently high.

I did some experiments with a Rhodes sample from another keyboard first, with totally different results. In order to make it sound good, I needed to split the keyboard into a high zone with keyboard tracking and a low zone without keyboard tracking. Also, with the low-pass filter, it sounded awful at higher velocities; the band-pass filter was better but not perfect either, since it reduced the overall volume again at very high velocities.

Apparently the Nord Wave has two settings between no keyboard tracking and full keyboard tracking, so it should be possible to create a better sample that is designed for one of these intermediate settings (if anyone actually wants a Rhodes sound for their Wave).

[Frantz]

Thank you for your explanation FunKey.
I wish we had two new knobs :
- one for the filter slope, from 2 to 32 db/oct
- one for the kb track, from a fixed frequency to each pitch fundamental
dunno if that could make sense, but I'd like to tweak and hear !

[fjhridder]

Hi - I am a stage 2 newbie - I listened to the mp3 and it sounds awesome. When I loaded it not my stage 2 it did not work - I guess I do not have the right sample loaded. Can anyone help?

Thanks

[neolithic]

Hi fjh- welcome!
Check this other post out, as it's important to know that when you load a sample it is not the same as loading a patch. A patch is ready to eat, but a sample is like a frozen meal ... Could be the same problem you are having.
general-nord-forum-f29/sample-editor-in ... t8110.html

[pablomastodon]

[quote="neolithic"A patch is ready to eat, but a sample is like a frozen meal[/quote]

nice! I may have to borrow that...

bless,


Pablo

[neolithic]

there's a better metaphor out there I'm sure! Must have been hungry when I wrote that!