Octave Leslie simulation

[elif]

Once upon a time I measured a set of impulse responses of the horn and drum of my 142 Leslie.  They were measured at each of 16 angles of rotation.  It was about twenty years before I did anything with them.   After I retired in 2018, I worked on the simulation for three months, eventually porting it into a TI TMS320C6748 low-cost development kit (floating point DSP).  I was going to make myself a Leslie simulator.  Then just after that I bought a Ventilator.  Why build what you can buy right?  The thing is, though it could use some work, I like the Octave simulation better.  To me, it sounds more like a Leslie in a room, probably because it was from a Leslie in a room, my living room.

 

The following is a link to a google site I created to allow someone to audition these.  The mp3s are hosted on my google drive.  I presume there should be access to them.

 

Leslie speaker - Octave sim

[KuruPrionz]

I couldn't get the 1st track to play, silence even if I moved the cursor into what should be the sound track.

2nd track played and sounded great, I like it!!! Kudos!!!

3rd track also sounded very good but I preferred the 2nd track to be honest. Well done. 

[Tom Williams]

First track played for me, as did the other two.  Octave sounded quite natural -- I'd happily use it.  Vent seemed to have (add?) a little more lower-mid presence than the octave; that felt a tad warmer.

 

I'm quite happy with my Kurzweil Leslie, but I think both Octave and Vent sound even better, especially in isolation from the noise of a band.

 

Amen to KuruPrionz's Kudos!

[Mighty Motif Max]

That sounds really good! All three files work for me - I find I like the sound of the Octave sim best.

[kenheeter]

Your Octave simulation is very good!! Any chance you'll market it?

[elif]

More about the simulation (though the details are quite hazy):

• The bass driver was a P15LL reconed by Gordy Kjellberg.
• The horn driver was one sold by Goff at the time, equiv. to an Atlas PD5A.
• The IRs capture the room response as well as the direct response.
• The IRs were captured with an HP3562A Dynamic Signal Analyzer at 65536 sps.
• The microphone was an MD421 spaced approx. 2-3 ft from the cabinet.
• The crossover was bypassed for the measurement.
• The crossover in the simulation is a LR4 @ 800 Hz (the stock x-over has ~ 6 dB dip @ 800 Hz)
• The IRs were measured at low power (3W to 5W) using a solid-state amp.
• The IRs were normalized as a set -- the same gain value was applied to all IRs.
• There was no effort to optimize the balance between drum and horn. The gain of each is as it was measured.
• The left and right signals for a rotor were generated corresponding to a pi/2 angular offset, equivalent to a 90 degree microphone angular separation. This could be easily changed to any pi/8 angular separation.

The reasons I wanted to play with this again:

• To capture the IR of a stock Leslie horn driver
• To get a more realistic room response for a stereo presentation using a figure-8 microphone in an X/Y configuration
• Different "rooms"

[Delaware Dave]

Are you planning to create an outboard pedal for this, a VST, or just do nothing?

[jverghese]

Very pleasant and natural sound. Always nice to see a sim that is not 'obsessive' with the Doppler effect. Percussion sounds great, too. It would be interesting to hear how it sounds with brighter drawbar settings.

[elif]

1 hour ago, Delaware Dave said:

Are you planning to create an outboard pedal for this, a VST, or just do nothing?

I haven't thought about it, but now that you mention it...   If I do anything (rather than nothing 😉 ), I suppose a VST would be the obvious choice.  Developing a commercial pedal seems a little daunting.

[KuruPrionz]

3 hours ago, elif said:

The reasons I wanted to play with this again:

• To capture the IR of a stock Leslie horn driver
• To get a more realistic room response for a stereo presentation using a figure-8 microphone in an X/Y configuration
• Different "rooms"

Consider using a small guitar amplifier to drive the speakers since you are recording at reasonable volumes. A Fender Champ (good chance you can borrow one!) or Vox AC5, all tube, 5 watts and both have small inexpensive speakers that do not showcase just how good they are. I used to gig with a Champ head plugged into an EV 12L, monster tone!!!

[jverghese]

By the way, when you say 'Octave', are you referring to the software package that is similar to Matlab?

 

I'm a bit surprised that 16 angular positions is enough for such an authentic sound. I assume you interpolate between adjacent pairs of IRs, in a similar fashion to what the vibrato scanner does. Great work anyhow.

[elif]

55 minutes ago, jverghese said:

By the way, when you say 'Octave', are you referring to the software package that is similar to Matlab?

 

I'm a bit surprised that 16 angular positions is enough for such an authentic sound. I assume you interpolate between adjacent pairs of IRs, in a similar fashion to what the vibrato scanner does. Great work anyhow.

Yes, Matlab -> Octave. Regarding the number of impulses. I was surprised too, but it's what I had to work with.  A simulation I did of an A-100 chorus/vibrato, here, is what led me to try the same approach on this simulation, essentially cross-fading FIR outputs, complicated by the two different rotational velocities and the accel/decl between them.

[Macsaint777]

Yours sounds excellent. But I would say that the Vent II version has the vent set way slower and further back than most of us would expect. 

[Delaware Dave]

3 hours ago, Macsaint777 said:

Yours sounds excellent. But I would say that the Vent II version has the vent set way slower and further back than most of us would expect. 

This....

[elif]

It's obviously maladjusted. It may have been recorded when I first got the Vent in 2018 (and stopped working on the simulation).