Digital guitar amps and digital distortion plugins still don't sound like the real thing.
For one reason, adding distortion adds harmonics which requires oversampling in the digital domain to avoid aliasing. There is a limit to how much oversampling is practical, and limits to the practical sound quality of filters needed for reducing sample rates back to the original.
For another reason, analog distortion mechanisms can be rather complicated even for something as simple as one transformer. Accurately modeling them in the digital domain is still something not fully perfected, especially for something as complex in distortion behavior as a tube guitar amp.
Digital models do keep getting better though. Maybe someday we won't be able to tell the difference, but not yet.
I'm talking about plugins that sell for hundreds of dollars. If one goes back and thinks about the history of it a little, for hardware there was the Line6 Pod. At first it was claimed to sound just like the best 20 guitar amps ever made all in one little box.
As modeling technology got better, each time they would claim, "this time it really does sound like the real thing." That happened so many times now they just say, "it sounds more like the real thing than ever," and, "it's more convenient and versatile than the real thing." In other words, the claims are getting closer to the truth.
The above is in reference to both hardware and software emulations of devices that distort in ways people have found to be musically useful and expressive. In terms of expressiveness, one thing that refers to is being able to play the distortion box as part of an over all instrument. For example, depending on how one plays guitar and how one manipulates the strings, a good guitar amp needs to give an interesting variety of sounds that vary with playing, and also vary in useful ways as the volume, tone, and pickup controls are operated during performance.
Last edited:
About sweetness...
Pure sine waves don't really sound pure do they? Subjectively, some synths sound purer to me. If I compare them side by side I can make an argument that the sine wave sounds purer, but when listening for enjoyment, the synth just sounds smoother, more enjoyable.
So there. Maybe that puts us 0.1% closer to an explicit, working definition of sweetness.
Pure sine waves don't really sound pure do they? Subjectively, some synths sound purer to me. If I compare them side by side I can make an argument that the sine wave sounds purer, but when listening for enjoyment, the synth just sounds smoother, more enjoyable.
So there. Maybe that puts us 0.1% closer to an explicit, working definition of sweetness.
I was talking about the history of serious engineering attempts at emulation. They can't do it at a commercial level. Not yet.
Last edited:
Some years back I worked with a company that made a fet solid state amp that accurately responded the same as a valve guitar amp. They ran some tests with players, who could tell which was which. They couldn't easily do blind tests as the amp was a combo - speakers in the same box, and people wanted to twiddle knobs etc.
They made some changes: added a fake valve warm up delay, and a bit of hum same as the valve one had, and made the boxes look identical. Lo, no one could tell the difference... Not a statistically valid test, but interesting!
They made some changes: added a fake valve warm up delay, and a bit of hum same as the valve one had, and made the boxes look identical. Lo, no one could tell the difference... Not a statistically valid test, but interesting!
Which is exactly what I said. Your definition of serious engineering and mine may however differ in scale.
My definition of serious engineering for a commercial product includes among other things designing to a price point. From an engineering perspective it can be very challenging. No point in designing a digital guitar amp if it would cost significantly more to manufacture than the real thing.
Maybe that is because pure single frequency tones don't appear in nature. Any natural sound has overtones, harmonics, one wayor another. Evolution and all that.
Jan
My definition of serious engineering is say, CERN. Or the channel tunnel, or an MRI machine. All of which by the way have cost constraints. There is no 'cost no object' since the apollo missions (well maybe Ilmor's Be pistons for F1).
There was a nice minimoog emulator done that supposedly was analysed at a circuit level. But still not the same as the real thing as didn't drift out of tune as soon as looked at it...
There was a nice minimoog emulator done that supposedly was analysed at a circuit level. But still not the same as the real thing as didn't drift out of tune as soon as looked at it...
Sure.
However, as you may recall I was responding to PMA who said,"it is not difficult to add distortion purely mathematically, by software, to the original sound file. The myth of "sweet distortion" can be very easily disproved, similarly as other myths have been disproved."
My point was that in practice it can be difficult to add some types of distortion, and "sweet distortion" may not be so easily disproved (whatever it is).
We probably could agree that in principle we think there is no reason why audio distortion of some given musical type could not be produced digitally if commercial cost constraints were not as limited as they are.
As to disproving a myth of "sweet distortion," I have not heard of such a myth. Maybe someone can provide a practical engineering definition of what it is supposed to be?
However, as you may recall I was responding to PMA who said,"it is not difficult to add distortion purely mathematically, by software, to the original sound file. The myth of "sweet distortion" can be very easily disproved, similarly as other myths have been disproved."
My point was that in practice it can be difficult to add some types of distortion, and "sweet distortion" may not be so easily disproved (whatever it is).
We probably could agree that in principle we think there is no reason why audio distortion of some given musical type could not be produced digitally if commercial cost constraints were not as limited as they are.
As to disproving a myth of "sweet distortion," I have not heard of such a myth. Maybe someone can provide a practical engineering definition of what it is supposed to be?
But natural sounds can be anything but sweet. "Nature" can't be used as benchmark for what sounds good, because it includes everything that sounds bad too.
Some ocarinas are strange. They sound quiet even if you turn the volume up quite loud. Harp is the same way. When I'm testing an amp one of the things I will do is put on a good harp recording and turn the volume up, up, up until the amp overloads or it gets too loud. It's astounding to me how loud you can make a harp without any apparent irritation from the sound.
As I understand it's the idea that low-order harmonic distortion sweetens the sound. Like tube amp distortion. The idea is that you can mostly ignore H2 to H4 in your equipment because these harmonics are usually either benign or beneficial.
I have hooked up a chunky AC trafo to a speaker and it added a sort of nice texture to the sound overall (most notable with flutes and woodwinds). Not a step up in fidelity, but if I were a musician I would definitely be interested. If you have a handful of old CMCs in a parts bin you should try hooking them up to your speaker or as line level transformers to see how they act. You can get a lot of neat effects.
If you put a CMC in series with a speaker, depending on it's size you will get silence interspersed with bursts of sound when the CMC is saturated by low frequency content in the music. Or if you have a low value CMC this effect will be shifted into higher frequencies.
But as far as a scientific correlation, I don't think you will find a relationship looking only at harmonics. I think the sound modulation envelope matters as well as frequency response, attack, decay etc. To only look at the first 3 harmonics of a constant sine wave signal doesn't seem to have much correlation with how something actually sounds.
If instead of the chunky power trafo I used a 12V step-up trafo from an old UPS, the sound is anything but sweet. I don't know exactly why, but this trafo always added an irritating edge to the sound whereas the power trafo was usually pretty benign.
Last edited:
Pavel is right, these days adding distortion is trivial. You are talking about modelling hum, buzz, noise and all the nasty driver resonances. Still not a huge problem for a species who has reached the moon and detected graviational waves, but not worth the bother. I wonder how many fag burns is optimum for a 12" guitar driver and if that should be a slider or a GUI where you put mouse clicks were the burns are. And of course did they spill stella or Grolsch into the unit on the previous gig!
There are some real cost no object work going on..... all in the military. For example, when a military 'device' was to be tested to assure it provided the yield etc desired the whole event was pay what ever it cost.
THx-RNMarsh
- Status
- Not open for further replies.