Tube amps have a signature sound

If there is a 'signature sound' it is the low levels of odd-ordered harmonic distortion which our ears are evolved/created to detect. Odd-ordered harmonics are how the ear detects loudness. Tubes make far less of this than transistors. Transistors don't make much either, but the hundredths of a percent that they *do* make is a demonstration of how sensitive our ears are to this type of distortion.

By contrast our ears do not care so much about even-ordered harmonics, but a properly set up push-pull tube amp is not going to make a lot of those either. So if the amplifier is set up right, and other important design considerations are met, a tube amplifier will sound more natural to our ears than a transistor amplifier will.

So IMO, the 'signature sound' is music.

Rodman99999, it is easy enough to demonstrate that in fact odd-ordered harmonics are indeed what the human ear looks for as a loudness cue. Its an easy test.

Get a sine/squarewave generator, an amp, a speaker and a VU meter. Run the sine through the amp and speaker and set the level so the VU meter reads 0VU (the meter will be across the amp terminals, this is not a loud test but it can be if you want). Now run a squarewave, and set its level such that is sounds to be at about the same level as the sine wave. You will find that the equivalent level is in the neighborhood of -20 to -24 db to get the same effect of volume. That's at least 1/100th the level of the sine wave!

Try it!

Rodman99999, its OK. Square waves are by definition made up exclusively of odd-ordered harmonics of the fundamental waveform (that's why they can sound so obnoxious). Maybe I should have mentioned that earlier.

Rodman99999, the clipping of an amplifier does produce a square wave, and those 'corners' are where the odd harmonics reside.

If you compare a tube amp to a transistor amplifier when both are clipping, the much harder clip of the transistor amplifier is visible on any oscilloscope and the obviously higher quantity of odd-ordered harmonics is easy seen. This is why transistor amplifiers have a harsh overload character and also why most guitar players prefer the smooth overload of tube amplifiers.

Mrtennis, it is possible to get more detail without an excess of treble energy. This is where understanding the rules of human hearing comes in.

The traditional way to get more detail is by reducing distortion. Distortion 'masks' detail ('masking' is a quality of the human ear; one of those rules...). Usually this is done by adding negative loop feedback. While this reduces distortion overall (and thus increases detail), negative feedback will also augment certain odd-ordered harmonics slightly, but enough that our ears detect the augmentation has hardness or brightness (another rule of human hearing: we use odd-ordered harmonics to detect loudness and we are **very** sensitive to them, even in amounts of only hundredths of a percent).

Negative feedback frequently tends to compress dynamic contrasts also. Incidentally an amplifier that employs lots of feedback might measure very flat, but is very likely to sound bright.

So the way to get more detail (and impact) without brightness is to reduce distortion without using negative feedback as a crutch. This results in a detailed amplifier that is also at the same time very relaxed and lacking artificial loudness cues of the type you mention.

IOW: you can have your cake and eat it too.

MrT, I interpreted your comments correctly. I want you to understand that I was not contradicting you either. What I was pointing out is that what you were reporting in your post has an underlying design issue, and what can be done about it.

IOW your perception of a frequency response issue is in fact also related to detail in many amplifiers. Here is your comment:

i find many in-production tube amps unbalanced in frequency response and overly detailed. there is often too much treble energy. i find this is the case with solid state amps as well.

I was commenting to the design issues that cause 'brightness' and 'overly detailed' to be related. It is negative feedback.

Chris, its on-topic as far as a signature sound is concerned, on account of the fact that tube amps use far less feedback or maybe none at all compared to transistors. I am referring to global or loop feedback.

Local feedback can refer to loop feedback around a single stage, which can still be problematic, or it can refer to degenerative feedback, which does not involve looping around a gain stage and so does not cause trouble.

The ills of loop feedback have to do with a phenomena of all amplifiers called propagation delay. This is the time it takes for a signal to propagate from the input to the output of an amplifier or from the input of a gain stage to the output. The delay is a constant for a given amp and does not change with frequency. From this it might be easier to understand that a loop feedback signal will never arrive back to the input in time to correct the signal that it is supposed to.

As frequency increases, so do problems caused by the lateness of the feedback signal. Eventually this can cause the amplifier to oscillate, so wide bandwidth amplifiers usually have a mechanism to reduce feedback above a certain frequency where oscillation might be possible.

Anyway, negative feedback is a *destabilizing* factor in an amplifier, but will improve overall distortion and sometimes increase bandwidth, at a price: it will, while decreasing most distortions, actually *increase* certain odd ordered harmonics, not much (actually we are talking hundredths of a percent here), but enough so that the overall effect is as MrT describes above as a perception of brightness. You might get with it more detail (as getting rid of distortion reveals detail underneath), but the amplifier may be painful to listen to.

My own opinion is that feedback has to be avoided if you are to have an amplifier that will lack loudness/harshness cues. You still want to get detail, so you still have to reduce distortion- you do this by using linear amplification devices and techniques, so distortion creation is minimized. A couple of examples are triodes and class A operation. Of course, I think a major contributor is the output transformer :)