Korg Forums

Hello,

I'm looking for the sound of the organ from Golden Brown, you can hear it being played with the harpsichord on the base hand, and 2 in between separately it sounds like a carnival organ might do, but I have nothing in my Kronos X any advise?


https://www.youtube.com/watch?v=AWAsI3U2EaE

Advice on how to make the sound... Well first off according to wikipedia the sound is, "pitched halfway between the keys of E minor and E-flat minor, possibly to accommodate the tuning of the harpsichord."
Don't know of a way to do that with CX-3 offhand but you could detune everything in the Kronos by fifty cents in the Global Basic Setup tab.
Recreating the timbre is always the trickiest part... Sequencing the notes usually helps so you can loop the music while designing. Though with organs... it might be more beneficial to take a snapshot of an isolated note with a spectrum analyzer and then try to match the harmonic structure with the drawbars.

Steps:
-Record song
-Chop out isolated organ note
-Run note through a spectrum analyzer and freeze the resulting spectrum.
-Find same note on Kronos and run it through another spectrum analyzer to try to match the harmonics between the two.

Goodluck.

I recorded the tracks and started trying different stuff, although the organ sounds didn't give me what I wanted, however the synth mallet sounded the best surprisingly,

Thanks.

The sound has a slowish attack, so worth tweaking that parameter.

Searching on Google what instrument was used by Dave Greenfield, not easy to find. In one thread (only the title), I've red Hammond L100.
Without any Leslie I suppose.
You shouldn't use CX3 engine, as you will have to edit Amp envelope. I agree with voip : there a quite slow attack.

I think it could be just a synth sound Ive sent details to Dave at Audora Audio he might know.

CharlesFerraro wrote:Advice on how to make the sound... Well first off according to wikipedia the sound is, "pitched halfway between the keys of E minor and E-flat minor, possibly to accommodate the tuning of the harpsichord."
Don't know of a way to do that with CX-3 offhand but you could detune everything in the Kronos by fifty cents in the Global Basic Setup tab.
Recreating the timbre is always the trickiest part... Sequencing the notes usually helps so you can loop the music while designing. Though with organs... it might be more beneficial to take a snapshot of an isolated note with a spectrum analyzer and then try to match the harmonic structure with the drawbars.

Steps:
-Record song
-Chop out isolated organ note
-Run note through a spectrum analyzer and freeze the resulting spectrum.
-Find same note on Kronos and run it through another spectrum analyzer to try to match the harmonics between the two.

Goodluck.

That is a nice idea. Do you guys have any spectrum analyzer to recomend for that, for Windows?

Hi tand, lots of companies offer free spectrum analyzers that will run on Windows. Blue Cat Audio, Melda, Voxengo, and Vengeance-Sound just to name a few.

I was just assuming he was being sarcastic. If he was serious, this is not good advice.

First, you can't isolate just that sound in the recording.

Second, so what if you have 2 different snapshots from a spectrum analyzer? It still doesn't tell you how to get from point A to point B.

The harpsichord is the prominent part. Get in the ballpark with the background part and focus on the harpsichord.

Hah wow, it's rare for someone to challenge my expertise.

I'll start out by saying that my method would work better for more traditional organ timbres. Not so much for faux organs like the one in Golden Brown.

Anyway a common window size for an FFT analysis is 1024 samples. That's 21 milliseconds of audio at a 48kHz sample rate; akin to the sharp transient of a snare. There are literally dozens of spots that you could pull a reliable analysis from every single second that goes by in the song. Finding an isolated faux organ part would not be hard.

A snapshot of an organ timbre will definitely tell you how to get "from point A to point B". Organ drawbars are literally harmonics. A spectrum analyzer displays these harmonics.

CharlesFerraro wrote:Hah wow, it's rare for someone to challenge my expertise.

I'll start out by saying that my method would work better for more traditional organ timbres. Not so much for faux organs like the one in Golden Brown.

Anyway a common window size for an FFT analysis is 1024 samples. That's 21 milliseconds of audio at a 48kHz sample rate; akin to the sharp transient of a snare. There are literally dozens of spots that you could pull a reliable analysis from every single second that goes by in the song. Finding an isolated faux organ part would not be hard.

A snapshot of an organ timbre will definitely tell you how to get "from point A to point B". Organ drawbars are literally harmonics. A spectrum analyzer displays these harmonics.

Thank you Charles

I can't see the spectrum analyser approach working, too much clutter in the song.
The idea seems counterintuitive too, but then again I haven't looked at such an audio spectrum.
Might have a nosey if I get the chance at weekend.

I have a friend who is somewhat of an expert on The Stranglers keyboards, will ask him for the answer to your question.
He does know Dave Greenfield so there is a potential for a definitive answer.

In the words of Arnie, I'll be back...

CharlesFerraro wrote:Hah wow, it's rare for someone to challenge my expertise.

I'll start out by saying that my method would work better for more traditional organ timbres. Not so much for faux organs like the one in Golden Brown.

Anyway a common window size for an FFT analysis is 1024 samples. That's 21 milliseconds of audio at a 48kHz sample rate; akin to the sharp transient of a snare. There are literally dozens of spots that you could pull a reliable analysis from every single second that goes by in the song. Finding an isolated faux organ part would not be hard.

A snapshot of an organ timbre will definitely tell you how to get "from point A to point B". Organ drawbars are literally harmonics. A spectrum analyzer displays these harmonics.

I will concede that if you were talking strictly a tone wheel organ with no V/C or Leslie, you could correlate frequency spectra to drawbar levels, but in that specific application, that seems like an awfully long and drawn out, possibly unreliable process compared to just moving the drawbars until it sounds right.

It’s a matter of how close you want to get to the exact settings you’re trying to replicate. The analyzer shows not just the harmonics but also the relative volume difference between the stops. I agree that a leslie or most post effects for that matter would reduce the user to doing things by ear the old fashioned way.

After slowing the track down to 30% it’s original speed I was able to snag three isolated samples of the faux organ within the first two bars. Its also worth noting that I did this after killing the stereo middle information to get rid of the bass instrument completely. As others have pointed out, CX-3 is not the best tool for the job. Even though I could match the harmonics, the attack was not reproducible.

I considered sampling the isolated parts (at a faster speed to reduce granularity) but each one is a chord unfortunately. That approach would require something like RX to delete one of the notes. Both the time stretching and spectral processing would introduce perhaps subtle but obvious artifacts too.
The best results then would come from matching the spectra with MOD-7 and using it’s envelopes and filters to fashion that attack.

Well thanks for all your input, Ive messed about with an organ sample using a parallel filter, along with guitar amp cab, room hall reverb, and a graphic eq for insert effect, I;m getting closer, so i;ve uploaded a sample of it,


https://soundcloud.com/andrew-downey-30 ... n-brownwav

Please let me know your thoughts,