Roger Mayer (Inventor of Octavia) Talks Tone, Amp Modelers

László

Only Human
Joined
Apr 26, 2012
Messages
35,859
Location
Michigan
Today I got an email from Lava Cable that had a link to this interview with Roger Mayer, Hendrix' pedal creator, inventor of Octavia, pedal maker to Page, many stars, done a few years ago.

http://lava-cable.myshopify.com/blo...tm_medium=email&utm_campaign=Reading+Material

He mentions two things I've been saying for a long time, and these statements resonate with me:

1. Chasing the tone on recordings is not the same thing as having the player's tone in the first place, and is pretty much a waste of time considering the sound sculpting done in the studio, during tracking, mixing, mastering, pressing, etc; and,

2. He uses a different analogy, but the gist of what says is the same basic thing I've been arguing for years - namely, that the creation of sound with tube amps can't truly be modeled because the nonlinearities inherent in producing sound from an analog, tube amplifier are too chaotic and unpredictable for a model to truly achieve the behavior of the amp:.

"Imagine you are beside a lake with an unpredictable wind blowing across the water and are holding a stone that you intend to throw into the lake. How could you possibly predict exactly what the pattern from the splash of the stone would be before you threw it into the lake also taking into account the effect of the unpredictable wind. You couldn't possible [sic] compute it ever. There are too many unknown variables involved."

Another thing he mentions got my attention, he makes a distinction between gear used for live work and max volume sessions, and low volume bedroom gear:

"I specialise in producing equipment that has the capabilities to produce the sounds that can be recorded to make hit records and some devices are designed expressly for live stadium type gigs at maximum volume with the back line amplifier then being miked up and distributed with a massive house PA. I do not make pedals that are primarily designed to be used at low volume in a bedroom as many people will tell you when they move to a real gig that the sound they thought was great at home just does not cut it in front of an audience and used in anger. The situations are completely different and require the correct solution."

The reason this got my attention is that I've often put a pedal in place that sounded good while I was practicing, but when I cranked the amp for a session for an ad, the tone was not at all like what I expected, and I was disappointed. And the reverse is also true; some pedals have sounded great cranked, but at practicing volumes, not so much. Very few do both well.

I think this is one reason that I prefer pedals from certain makers, and perhaps not others. It's something to think about anyway.
 
Yeah, I don't know. I agree I have yet to hear an adequate emulation, but it can be done. I think they just don't take everything into account -- for example, if you've been playing through the amp hard, it's going to react differently at the next time-slice than it will during your first time-slice after having turned it on and letting it warm up.

Even turbulence is predictable if you take enough things into account, the problem is that it's not yet able to be predicted in the time it takes to just wait and measure what happens next -- sort of like weather. I still think it can be done. Non-linearities can be simulated.

My belief stems from the fact that it can be recorded digitally. If it can be recorded and played back, it can be emulated -- that's the logical next step. Again, maybe not in real time. And maybe not without a supercomputer, but it can be done.

I mean, think about it -- worst case, it's a table lookup. Now hash the table lookup. Now hack the hashed table lookup. Voila, you have an algorithm.

And I am by no means a slouch -- I have programmed advanced physical modelling on the computer, though I have to admit, nothing as advanced as this.
 
My belief stems from the fact that it can be recorded digitally. If it can be recorded and played back, it can be emulated -- that's the logical next step. Again, maybe not in real time. And maybe not without a supercomputer, but it can be done.

One problem is that the recording isn't perfect. First of all, you have the microphone, that doesn't and at this stage of technology, cannot capture the true sound of the amp in the room. You can do some nice things with mics, and you can get some cool recordings, but play back any recording, on any sound system, and no one listening will believe there's a real band in the room. Listeners can tell the difference between a recording, even a very good one, and a live instrument.

I fight this problem all the time!

Making a recording of an event is different from generating the sound itself. The dynamics are different, the details are different...it's just not the same thing.

Which would be a less costly buy, the supercomputer that might come close to modeling them, or a few really good genuine tube amps? I guess my point is, why not just buy the real amps? Real amps are 100 percent of what they are. No modeling necessary.

And honestly, no truly creative musician needs every amp ever made, in a digital box. Hell, I am asked to do different styles all the time, and I do them with two amps and a few pedals. It ain't...rocket...er...computer science. ;)
 
Last edited:
Which would be a less costly buy, the supercomputer that might come close to modeling them, or a few really good genuine tube amps? I guess my point is, why not just buy the real amps? Real amps are 100 percent of what they are. No modeling necessary.
You know my answer -- the tube amps, of course. But based on the success of the Kemperer, I can see why it would interest companies from a commercial perspective.
 
You know my answer -- the tube amps, of course. But based on the success of the Kemperer, I can see why it would interest companies from a commercial perspective.

Yup, that's true.

I know that some big-time bands are touring, or considering touring, with Kempers simply to save money on shipping and cartage costs. Bands often spend six figure money on shipping costs for longer tours. So there's a reason to do digital modeling right there.

On the other hand, you could accomplish the same thing with a head and a cabinet simulator/power soak/direct box, like the Suhr, the Mesa, the Two-Notes or one of the others that are on the market, and there really wouldn't be a shipping cost penalty, since you either ship the Kemper or you ship the head.

So I dunno.

For me, it's an easy choice, I never tour anywhere!

In any case, here's to you, Dusty Chalk, for making this thread an interesting discussion!
 
Last edited:
Here's the odd thing, as far as I'm concerned: while I don't care for modeled amplifiers, I have no problem liking some modeled studio gear.

Today Plugin Alliance/Brainworx came out with a really good emulation of a Neve VSX console channel strip.

What's different about it is that they didn't just model all the bits and pieces that went into the console's channel; lots of companies do that. Instead, the plugin models each of 70 channels of the console. This is cool because one of the things that makes an analog board interesting to work with is that every channel sounds a tiny bit different because of parts tolerances varying from channel to channel, even right and left stereo pairs. Thus on a real console, the nonlinearities add up as you add channels to a mix. It makes a console sound interesting.

So you can specify different channels for every instance of the plugin, and I think it sounds more realistic than any other mixdown plugin I've tried. In fact, as far as my ears are concerned, the difference is startling!

I ran a single instance on the drum bus today, and it sounded great. Then I broke the drums into their respective tracks, and ran the plugin on each track - i.e., kick, snare, hat, toms, overheads and cymbals. As good as the single instance sounded on the drum bus, the added instances on each track sounded even better, more real, more juicy. This doesn't happen even with the very good Slate VCC, and separate plugins for EQ and dynamics, because all you get is the same thing on each channel.

It sounds an awful lot like my Neve summing mixer did. I'm sure others will copy this approach soon, but it's always fun being an early adopter! ;)

I'm very impressed with this plugin. I have Plugin Alliance's "All-all" bundle, and I do try to keep it up to date. I don't use the amp modelers, but most of the plugs are superb, and they've replaced an awful lot of highly regarded plugins - and hardware boxes - in my studio.

I think it also sounds more realistic than my former "console sound" benchmark, the Slate Digital VCC. That's saying something, because I do think Fabrice Gabriel, Slate's designer, is brilliant, and he has created some stellar plugs.

In addition to the Slate and P/A stuff, I use most of the Universal Audio UAD plugins (many of which are created by Brainworx). as well as the Softube plugs.

The advantage of the UAD equipment was that it ran independently of the computer, but today computers are so fast that native plugins do the same thing. However, the UAD has one advantage - the processing in the UAD hardware is up-sampled to 96 kHz even if you're running your DAW at 48kHz, and it sounds noticeably better. I have no choice, the TV video standard in the US is audio at 48kHz, and while I could work at 96 or higher and downsample, I've found things simply sound better keeping the whole project at the same sampling rate.

This also prevents accidental audio to video synchronization errors - once I lock to picture, I'm good to go. And if you're doing music and sound design for picture, that's pretty important.

As a result, I've actually duplicated some of the P/A BX plugins with the UAD versions, because once you start processing audio at the higher sample rate, it really does sound more three-dimensional and detailed in a way that's hard to describe in words.

Anyway, I guess my point is that there's digital studio processing that works for me, yet the amps don't, and maybe that really comes down to sound creation vs sound reproduction. There are tradeoffs I'm willing to accept in mixing that I can't live with in tracking.

It's odd, but also true.
 
Last edited:
Great thread. Love getting the thoughts of those who have been there and know the deal.
 
Thanks, Coyote!

On another front, speaking of modelers...

In today's email I got a note that Pete Thorn demo'd the Axe FX FX8 pedalboard modeler. I love Pete's demos, so I checked it out. He ran it into the PT 100 amps, at times in stereo with two amps, some settings in the loop, others in front of the amp.

As always, Pete's writing and playing were stellar.

However I did not dig the tone of the FX8. It made everything sound watery and over-processed. Granted, it just may be the way Pete set it up. But even the factory presets on my old TC Nova System sounded a whole lot better to me, and they (like most factory presets) were a bit over the top in terms of the processing, and had to be dialed back.

The Nova was a $350 box. The FX8 is a $1400 box. So I dunno, you tell me. I listened on my studio's monitors, it's not like I was making do with some crummy computer speakers. It just didn't sound right to me.

This isn't to say digital effects are garbage, they can sound great, especially modulation, reverb and delay. I use an Eventide H9 on lots of my tracks, and it's got a spot on my pedalboard for pitch, modulation and delay. I think it, too, sounds a lot less processed than the FX8.

Again, maybe it's just the way this was set up for Pete's demo. But he's usually very good at getting the most out of a piece of gear.
 
One problem is that the recording isn't perfect. First of all, you have the microphone, that doesn't and at this stage of technology, cannot capture the true sound of the amp in the room.

Everybody is always talking about the amp in the room sound with modelers. At every point in your life as a musician you are the only one who's ever gonna hear that sh!t. Nobody else does. They're listening to your guitar processed through microphones, mixing desk, high-pass filters, compressors and limiters, some sort of amplifier, and then finally through some speaker cones before it gets to your ears in the studio, on stage, audience, iPod, or car stereo.

That stuff's only a problem for at-home players and dudes who don't wanna sit in the control room.

There are a few reasons not to like modeling, but that argument shouldn't be one of them.
 
Last edited:
Everybody is always talking about the amp in the room sound with modelers. At every point in your life as a musician you are the only one who's ever gonna hear that sh!t. Nobody else does. They're listening to your guitar processed through microphones, mixing desk, high-pass filters, compressors and limiters, some sort of amplifier, and then finally through some speaker cones before it gets to your ears in the studio, on stage, audience, iPod, or car stereo.

That stuff's only a problem for at-home players and dudes who don't wanna sit in the control room.

There are a few reasons not to like modeling, but that argument shouldn't be one of them.

That's a load of garbage, and you of all people ought to know it.

And FYI, my statement wasn't an argument against modelers! It was a response to a post stating "My belief stems from the fact that it can be recorded digitally. If it can be recorded and played back, it can be emulated -- that's the logical next step." It's far more difficult to create the original sound of a real amp than to reproduce the miked sound of said amp, for the reasons I stated. Just as it's easier to get a recording of a violin than to create a convincing digital violin model, that sounds like a violin played live in a room somewhere.

Now, the fact is that if you play live with a band, the sound of the amp in the room you're playing in matters a great deal, because it needs to cut through the band and be heard. We've all heard amps - analog and digital - that are lost in a mix. Not only that, but it also is better when the amp not only cuts, but sounds good doing it.

I would submit that a recording of an amp in a studio will be a pretty crappy recording if the amp sounds bad in the recording booth or the studio's live room. And this is especially true if you're cutting a band live.

There are times your arguments clearly demonstrate that you're out to do no more than give me a hard time, and this is one of them. You obviously didn't read my original post or the post that I replied to very carefully.

And you clearly know - or damn well should know - that in a recording situation it matters very much what an instrument, vocal or amplifier sounds like in a room so that its sound can be captured with mic placement - or messed with if that's the creative goal. Furthermore, if you've ever tracked a live band in a room, you wouldn't ever make a statement that "the stuff's only a problem for at-home players and dudes who don't wanna sit in the control room."

I hear band after band live, and so many recordings of live shows, where someone's modeler or crappy amp is background mush and doesn't do what the person bought the amp to do, namely, make the instrument heard in the context of music.

You're looking for an argument? Now you've got one.
 
Last edited:
T
And FYI, my statement wasn't an argument against modelers! It was a response to a post stating "My belief stems from the fact that it can be recorded digitally. If it can be recorded and played back, it can be emulated -- that's the logical next step." It's far more difficult to create the original sound of a real amp than to reproduce the miked sound of said amp, for the reasons I stated.

I guess I still don't understand the context of that particular reply. That's okay, I'm not the smartest dude, I'll re-read it, and hopefully get it.

What I thought the comment (and thread) was about was the use of modeled effects/amplifiers/etc. may be useful for certain applications and not for others. My comment about bedroom players and dudes who don't want to sit in the control room was not meant to be disparaging, as I am both of those things every day.

I suppose the thing that got me flustered with the whole "amp in the room" thing is that the subject always comes up when people talk about modeled amplifiers, and maybe I just don't understand what they are talking about. When I'm recording an amp I'm ultimately hearing it through the studio monitors. When I'm playing or watching a live performance (unless it's like a jazz trio or a punk rock basement show with no PA) I'm hearing a mic'ed up amp through the monitors, side-fills, and FOH. I'm barely ever allowed or encouraged to turn up my amp loud enough to fill the room. The last time I played a 1000+ theater I had my power amp on the 17 watt setting and the volume on 2. It made me question why I had bothered to bring a 4x12 when my amp was little more than an oversized personal monitor, and it ultimately helped me decide to sell it.

I'm not looking for an argument and I'm not trying to give you a hard time (except maybe that one time about coffee pods), I'm genuinely curious about what makes some people so dissatisfied with the tones from modeled gear. I mean, I really do like and use it all: tube amps, SS amps, consoles, soft synths, analog synths... maybe not for everything, but I don't feel like any one thing is the answer for every application, that's all.
 
Ok, I gotta know, what's wrong with coffee pods, and whose side am I on?
 
What I thought the comment (and thread) was about was the use of modeled effects/amplifiers/etc. may be useful for certain applications and not for others.

My comment was that creating the sound of a digitally modeled amp - one that would really sound like an amp in a room - is a different thing than recording the sound of an amp with a digital recording system.

My comment about bedroom players and dudes who don't want to sit in the control room was not meant to be disparaging, as I am both of those things every day.

Um...I dunno about that one.

I suppose the thing that got me flustered with the whole "amp in the room" thing is that the subject always comes up when people talk about modeled amplifiers, and maybe I just don't understand what they are talking about...I'm genuinely curious about what makes some people so dissatisfied with the tones from modeled gear.

I can't say for sure what other people are talking about; when I say "the sound of an amp in the room" I mean an amp that sounds like a tube amp sounds when you're with it in the same room.

Whether it's an instrument or a model of an instrument, if it doesn't sound real when played, then it fails as far as I'm concerned.

Here's what I have found over the years, and still find, with modeled gear - and I should mention that I got into modeled sounds in the early 90s with the very first Yamaha VL-1 synth: if it's in a mix (live or recorded) with real instruments, it disappears even if it's at the same volume level!

Here's something even stranger: when I record my real Prophet 12 (a synth that uses digital waveforms but processes them with analog filters, much like the old PPG Wave series), it makes the modeled synths I use disappear in the mix, too. Then again, models will also disappear in the same mix as a DX-7, an all-digital hardware synth. Strange, hard to fathom? Yes. But also true. Hardware sounds different.

And this happens even though the modeled synths sounded fine beforehand, and it happens even when I am very, very careful to set the levels correctly.

When I play my guitars through real amps, and then try to blend them with modeled amps (yes, I have quite a number of those), the modeled amp tracks also disappear in the mix. Same thing, despite very careful level setting. Why? I think I can answer that question, and will try to do so below.

When I go out and hear a live band whose guitar player uses a modeled amp, even if it's miked up, all I wind up hearing is the vocals, the bass, the drums, and some mush; mush that is masquerading as a guitar amp. Why? Again, I think the answer has the same cause as the one about modeled synths and modeled amps in recordings.

I will also state preliminarily that sampled sounds do not seem to do this to nearly the same degree.

So here's my theory: The dynamic range of a modeled signal isn't right. The attack portion of the waveform doesn't cut through.

Alone, the instrument sounds like it oughta be right, but in context with other real instruments, it fails.

Then too, many modeled instruments just plain suck and sound like cardboard, but that's another matter.

This is why I'm dissatisfied with modeled signal-generators, like instruments and amps.

Signal processors, however, are a different story. They're not creating sound, they're modifying it. And digital seems to work fine for this.
 
Um...I dunno about that one.

WTF? When have I ever made fun of someone's skill level or reason for playing guitar here? Never. They were examples of playing when you're actually in the room with the amp.

I should have had the good sense to stay out of another one of your multiple threads about hating modeling like I normally try to do.
 
WTF? When have I ever made fun of someone's skill level or reason for playing guitar here? Never. They were examples of playing when you're actually in the room with the amp.

I should have had the good sense to stay out of another one of your multiple threads about hating modeling like I normally try to do.

Or perhaps you should have had the good sense to know I'd take offense if you talked down to me about bedroom players and dudes who don't want to sit in the control room.

EDIT: I just killed the rest of my reply because no one else needs to suffer through reading a personal dispute, and I'm over it.
 
Last edited:
Or perhaps you should have had the good sense to know I'd take offense if you talked down to me about bedroom players and dudes who don't want to sit in the control room.

EDIT: I just killed the rest of my reply because no one else needs to suffer through reading a personal dispute, and I'm over it.
Bummer, I was just about to go make some popcorn :)
 
Back
Top