Leslie simulation revisited

[Mark Hammer]

As I eagerly await an imminent exchange with a local fellow of a small Leslie for a couple of old repro pedals, it started me reminiscing about the Leslie (I used to have one exactly like the one I'm re-acquiring and I have missed it for  these past 23 years).  It also got me thinking about simulation of rotating speakers.

Most analog simulations of these doppler effects use BBD chips for time delay, and more realistic simulations incorporate ramp-up times to mimic speed switching and the time required for the motor and rotor to accelerate or decelerate.  Even more realistic simulations will incorporate band splitting to mimic separate bass and mid/treble rotors.

What you never really hear about is the way that rotating speakers also have some mild amplitude and timbral modulation.  Certainly the doppler effect produced depends on the movement of the speaker, but notches are not all the rotation produces.  Bear in mind that as the speaker, or rather the surrogate version of the speaker coming through the rotor, rotates away from you, the level is slightly attenuated and you also lose a bit of treble at the point where it is facing away.  Stand behind a closed speaker cabinet and you'd get the same thing.

So, what I'm wondering is whether any of the more successful or authentic-sounding Leslie simulators incorporate amplitude and timbre-modulation in sync with the time-delay sweep to mimic the throbbing of true Leslies?  

If not, then a more authentic simulation isn't all that far off.  In some respects, all you would need to do would be to harness a simple lowpass filter and attenuator network to the existing LFO.  

What is confusing me is keeping track of what goes high when other things go low.  Too damn many double negatives.  So, when the LFO swings more positive (i.e., goes high) does that mean less or more delay?

Presumably, less delay mimics the rotor facing you and more delay mimics the rotor facing away.  As the rotor faces away, you'd ideally attenuate the signal just a tad, on top of some modest lowpass filtering.  the lowpass filtering would reduce the apparent level anyways, which is why I would suggest only modest attenuation over and above that.

A single LED and pair of LDRs tied to the LFO output could do the job.  One LDR would be part of a simple passive resistive divider *after* the mixing stage (or it could be part of a resistive network determining gain/level in an op-amp stage), and the other would constitute the R in a simple 1-pole RC lowpass filter at the same point.  Both of these would be after the mixing stage because you need to maintain equal portions of dry and wet to produce the notches.  Any simulated "throbbing" is laid over top of that.

Given that they come after the basic flanger/Leslie device, you could actually design a generic module (not unlike the Geofex LERA) that tacks onto the output of a flanger and is tied into the on-board LFO.

It would surely requiring some tinkering with values and modulation amount, but the basic premise seems reasonable.  Comments?

[Solidhex]

I was just looking around for info on Leslie sims and found this thread from eons ago.. I still think the ultimate leslie sim has yet be built. I was wondering if you had put anymore thought into the subject Mark?

--Brad

[MartyMart]

This has come up several times, I think Marks thoughts from 2003 are still valid, though there have been a number
of improvements in the "digital" versions since.
Notibly the Line 6 Roto Machine, which seems to have all those elements of pitch/volume pulse/separate rotors
and even offers some breakup/ different leslie "cabs" and ramp up/down too for not too much money.
DIY ? yes possible but a little beyond my skills, so to be honest, I was happy to spend £80 on a Roto machine !
There are also a number of good Leslie emulations in software now, from stand alone sections of the "B4" plugin and
the Logic Hammond, as well as Guitar rig and Amplitube.
Having sat infront of a few Leslie cabs whilst playing a real hammond, I can say that they still DONT capture that
"moving air" sound, nor the very analog "growl" that comes from these wonderful boxes
MM.

[petemoore]

So, when the LFO swings more positive (i.e., goes high) does that mean less or more delay?
  A lot of algorythmics.
  Tremolo effect... would be more apparent in the HF range, bass tends to carry.
  Time Delay...[a new leslie sound product 'mimic-ment'? possibility?]..it'd be best served as a chorus type of ramp.
  Pitch shift...because speaker excursions would be 'bent' by the moving air source, similar to if a cabinet were moving foreward and back a foot or two [at the speed of the leslie].
  The thing that would be hard to get is the assyncronious behaviour of treble to bass in the pitch shifting, tremoloing and delaying.
  Almost worth considering a physical turning wheel which triggers taps which can be set for 'oclocks', then rotate the triggers for all the 'separate' effects of shift/trem/delay...set the ramp up and ramp down points as a 'time' [like 6;00 to begin delay ramp up, 12:30 to initiate ramp down]...just for the learning process of having all the 'schtuff' as adjustable, then you could start working on the ramp shapes and sized...to cajole a stationary speaker into mimicing a leslie through electronics [youch].
   Not to be overly pessimistic about speakers ability to do this, if past attempts have anything to show us, something that sounds cool and leslie-like, makes a very usable effect that can be called 'leslie type effect'.
   Something with frequency separation in the form of Hi/Low channels and Hi/Low frequency drivers, or even more elaborate, hi and low frequency drivers in a '180' cabinet, drivers pointing both ways, perhaps a 'split dissipation' on the back to shoot the sound out more to the sides a little.
   

[Mark Hammer]

Having sat infront of a few Leslie cabs whilst playing a real hammond, I can say that they still DONT capture that
"moving air" sound, nor the very analog "growl" that comes from these wonderful boxes
MM.

I've probably stated this elsewhere in the forum...too many times... but my sense is that Leslies are a sort of "post-production effect" in that they are imposed at the very end of the signal chain.  The same way that taking a mixed down track with overdriven amps and running THAT through a flanging process sounds different than plugging your guitar into a flanger then an amp, sticking a Leslie simulator earlier in the signal path and then running that out to a stationary speaker or two isn't quite the same thing.  I'm not saying it can't come close, just that what makes up the difference is the fact that a real Leslie is at "the end", and things added "at the end" tend to sound different for a whole host of reasons.  In the case of the true Leslie, there is fact of a tube amp and transformer just ahead of the speaker/rotor, not to mention the manner in which the rotated sound bounces off whatever reflecting (or nonreflecting) surfaces may be present.  Both of these aspects serve to further filter the sound in ways that can't happen if the simulator is placed before the amp (and whatever harmonic content it adds to the "leslie'd" sound) and the speakers are pointed right at you.  I like my Roto-Machine as much as the next guy, but I suspect it might sound even better if I mic'd my amp and fed that to the pedal and a mixer, instead of simply feeding the guitar to the pedal then to two amps.

Pete,
There's something in your posting (and e-mail), but there's a little too much enthusiasm and detail packed in there for it to spell out clearly to me at the moment.  I'll ponder it for a bit and get back to ya.

[Paul Perry (Frostwave)]

While it is true that the Leslie is usually at the 'end' of the chain, it is also true that - more often than not - the guitar doesn't go through it at all.
Whenever I have seen them used, it was either in a small venue where the keyboards went thru the leslie & the leslie speaker was the final for the keyborad (while the guitars were going through separate amps & that was it) or , if it was a larger venue with a PA, the leslie was stereo miked & fed to the mixer.
As for amplitude modulation, leslie simulations usually have that, I believe.
The big problem - for simulating - in my opinion, is that the path length varies with the rotation of the horn, and this implies pitch modulation. So you have
1. phase variation
2. fundamental frequency variation
3. amplitude variations arising from interference between multiple audio paths.
If my life depended on putting a realistic leslie effect on disc, I'd use a dummy head stereo mic. I realise that's not an option for live, though.

[kahel]

The Dynacord CLS-222 was probably the best attempt at recreating the leslie sound with pure analog circuitry. It has four BBD lines and four variable gain amplifiers; two LFO's for hipass and lopass with separate ramp speeds up & down, all in stereo. It sounds pretty cool.

But it's not a Leslie.

K

[Solidhex]

I find a lot of times in my opinion that there's not enough attention is payed to the overdrive sound in leslie sims. Usually just a bit too muddy. Washes away the detail of the effect. A clean leslie though doesn't seem to have enough animation in the sound to me...I'm always looking for that "No Matter What" Badfinger tone. I do have a motion sound rotary guitar amp that basically does the job for guitar. Its great for recording. I used it on the intro the song "No One Knowing" you can hear it here... www.myspace.com/braddavis. We threw a Danelectro 12 string through an Analogman Bi-compressor mic'd in stereo.
  I was thinking about getting one of those Line 6 Rotomachines though.

--Brad

[Paul Perry (Frostwave)]

I'm sure Solidhex is onto something here.
Bi-amping - and the consequent absence of bass intermodulation with the high end  when overdriven - is doubtless part of the character.

[petemoore]

  Time Delay...[a new leslie sound product 'mimic-ment'? possibility?]..it'd be best served as a chorus type of ramp.
  I must have run out of time to edit...
  THing is with leslie sims, they're leslie sims.
  Getting the swoop off the inside of the leslie cabinet..when the horn whips by it, there will be many things happening, cancellations, re-inforcements, and delay like chorus which for better analogy would be like an rotory echo on a brake rotor which is warped, only at a certain index of rotation does the brake pad stick, or does the short, accellerating/decellerating delay sound occur...while ...at the same time when the horn is whipping around and coming at you, the doppler is + but moving, the pitch is + a bit as a result, when the horn is whipping away the doppler effect is - as in the pitch is 'behind' or dropped a touch, and the speed of the moving sound is greates when the horn is at the two 180's [like the power stroke of a bicycle pedal] and 0 speed when the horn is facing you, accellerating/de-accelerating during the in-beween portions.
  Getting that would require some seriously 'complex clockwork' as in the LFO triggering sequences of events other than typical phase/echo/tremolo, or some slick trickery.
  Then, after concocting extensive electronic cobble-izations [well past the effort of getting a leslie], you could get the leslie out to have something to compare it to.
  If you could get that all going on to a PCB layout, things would be good, perhaps everyone would want at least one.
  I should refrain from contributin to leslie simulator discussions other than 'I've built various types of leslies out of junk'.
  that said...a mechanical LFO rotator, one with various triggers strewn around the hole it is in, used to trigger 'events' when to start [or when to start then stop], and multiple adjustable rotational indexings possible, as the triggers are on clips...
  example... "to change parameter 1-a just move trigger clip 1-A from the 2:00 position to the 4:00 position, you can leave clip 1-B in the 6:00 position.
  Maybe if you can sort through enough events to get a 'real-er' leslie simulator, then replacing the *mechanical device [which would be less tricky to 'adjust'] which triggers various selectable/adjustable electronic circuits, into a *digital or *analog relative.
  Getting the room walls sounds that aren't around the [front loaded?] cabinet..I'll leave to the next revisitation.

[King Tubby]

I got this a few days ago... A well used Little David, Odd looking little  thing.







errr havent actually tested it yet as it requires 110v and havent got round to sorting it out for use on UK voltage 

hehe the design means It can only take 'L' jacks ...unless you put it on it's back... odd hehe

[Solidhex]

I've seen those before... Maybe we should figure out a DIY project for a small rotary speaker amp! Just need an amp and a quiet motor to rotate a baffle in front of it.. dual horns would be best but maybe some baffle shape would work well. Some light weight material that is easy to cut and reflective sound wise?


--Brad

[Solidhex]

I bet salvaging parts from an old record player (motor, pulleys. etc) could be adapted for use in front of a speaker....


--Brad

[gez]

Been quite a while since I spent any serious time working on my own designs, but I think the following is right (memory is a little hazy).

When the rate is cranked on a phaser, the sweep covers enough range (and at a fast enough rate) to cause some amplitude modulation.  Makes sense if you think about it as notching involves partial/total cancellation of a signal at a specific frequency (band).  It's not as noticeable with common or garden phasers, but you notice it more if you shift the range of the sweep up. 

We ten to hear from top to bottom, as it were.  It's why the average person hears soprano over alto/tenor/bass in part-writing.  With phasers, I suppose it's a bit like boring holes in a brick wall.  Do it low enough and you end up with arches, like in a bridge.  You can have a fair bit of notching (arches) going on, but you can still walk across the top of the wall unhindered.  In other words, the ear latches onto harmonics etc and we don't notice the amplitude modulation going on underneath.  Bore the holes towards the top so that they 'scoop' out holes in the top of the wall, and you're in for a bumpy walk.

Shifting the range of a phaser upwards is desirable when trying to (easily) sim a rotary speaker as it means you get built in amplitude modulation, but it also means you don't get much notching down below (and I'm not talking Y-fronts).  So, it helps if you add a little low-band 'rumble'. 

[Krinor]

I've seen those before... Maybe we should figure out a DIY project for a small rotary speaker amp! Just need an amp and a quiet motor to rotate a baffle in front of it.. dual horns would be best but maybe some baffle shape would work well. Some light weight material that is easy to cut and reflective sound wise?


--Brad

Great idea !!!

And remember those old Yamaha, Akai etc. cheap organs from the late 70's ? You'll find them dumped everywhere. Some of them actually contain cute little rotary speakers complete with horns, motors and all. I guess they can be salvaged and used for such a project. And if it doesn't actually sound like the real deal, at least it would be great fun ? Oh, and these old organs usually contain some pretty okay 12" speakers too. I found one with an alnico magnet which I use in my champ amp. Very nice.

[petemoore]

  I made the spinning baffle from 1/4'' plywood, the curved part was just curved 1/4'' plywood with rows of partial depth cuts running across it to make it curve easier, then I did about the same thing [matching-indexed, partial depth cuts] to the stop board to make something to push the curved board into...anyway, I ended up with two flat sides.
  That went on top of a record player turntable [16, 33-1/3, and 78 RPM settings] mounted in a leslie cabinet [looked a little like the Beverly Hillbilly's truck bed sides, on which a 1x12'' speaker enclosure was set, down facing into the baffle, very close [shimmed to as close as could be without touching]...worked pretty good after counterbalancing the baffle, stain was put on the speaker box to mark where the baffle would rise because of off balance/centrifigual force, then screws were placed to counterbalance it.

[David]

(Ladies and) Gents, we might be working too hard here.  Recall this thread, if you will...

http://www.diystompboxes.com/smfforum/index.php?topic=25130.0

Also this one:

http://www.diystompboxes.com/smfforum/index.php?topic=17465.0

[Mark Hammer]

The Korg CX-3 document is still posted.  The file of key interest is called S3.gif.

[snoof]

Here's an example...

http://www.songworks.com/images/DIY_Rotary_Wave_Assembly_Instructions_3_7_04.pdf

There's some good food for thought in there.


if you added a horn setup as well it might come close.

Diff approach...

The real parts are easily found.

http://www.tonewheelgeneral.com/build_page.php?item_no=HS-HRN

high freq driver and a motor, split at 800hz, send the lows to a trem.  I had a Motion Sound unit, same sorta setup, real spinning highs, simulated lows.

I had an old Leslie type deal from the 70's that used a real spinning les horn, and crossedover the sig, and sent the lows out to an ext speaker, no simulation(took the speaker out from your amp as the sig source).

[snoof]

The Korg CX-3 document is still posted.  The file of key interest is called S3.gif.

My CX-3 sounds pretty bad on the fast setting(high freqs).  but it might sound OK for gtr use though.