Pickups- physics or cooking?

[Mike Sulzer]

Well, if Wolfram were right, an album of "easy listening" music would consist of a sine wave.

I guess this is partly true in that it's unlikely to contain instruments with harsh timbres, tritones or Giant Steps-esque leaps into the unknown. And Wolfram is probably the kind of guy who listens to a sine wave to chill out anyway.

There is an interesting resonance between this viewpoint and that of audiophiles who think a sine wave is too easy a test for a power amp, and real music is somehow more challenging. (This argument is from my audio bullshit busting hero, Douglas Self.) The amp obviously doesn't know or care about the difference, but we do.

So, prodding the hornets' nest with my troll stick again: How can I state so confidently that a hi-fi amp doesn't interact with what it amplifies on any deep level? Am I justified in following Self's lead and laughing at claims that an amp can alter the rhythm of music?

And, coming back to pickups, I've heard claims that the construction of the pickup can alter the attack and dynamics of the guitar. What could be the physics behind this?

Again, we have to be careful about projecting characteristics of human perception onto the outside world. An amp might not care what waveform you run through it, but you do. Slight alterations of transient response alter the perceived sound, and for good sound reproduction, an amp that does not change anything is good.

For a guitar, a huge change in sound occurs as you increase the gain from a completely clean signal to one where the initial transients are modified non-linearly by the amp.

So how do changes in a pickup alter the dynamics and attack of a guitar? For one thing, changing the location and height of the resonant peak changes the amplitude and other characteristics of the initial transients. These changes lead to bigger changes when you consider the non-linear response of the amp.

[kg]

There is an interesting resonance between this viewpoint and that of audiophiles who think a sine wave is too easy a test for a power amp, and real music is somehow more challenging. (This argument is from my audio bullshit busting hero, Douglas Self.) The amp obviously doesn't know or care about the difference, but we do.

add me to the list of folks who think "a sine wave is too easy a test for a power amp," steve.

now, if you said "multiple superimposed non-steady state sine waves" then i would change my stance. the easiest way to generate those is to play some music!

And, coming back to pickups, I've heard claims that the construction of the pickup can alter the attack and dynamics of the guitar. What could be the physics behind this?

as mike points out, it is likely something as simple as frequency response.

[fieldwrangler]

Well, if Wolfram were right, an album of "easy listening" music would consist of a sine wave.

I guess this is partly true in that it's unlikely to contain instruments with harsh timbres, tritones or Giant Steps-esque leaps into the unknown. And Wolfram is probably the kind of guy who listens to a sine wave to chill out anyway.

And he'd eat a tablespoon of sugar for dessert.

There is an interesting resonance between this viewpoint and that of audiophiles who think a sine wave is too easy a test for a power amp, and real music is somehow more challenging. (This argument is from my audio bullshit busting hero, Douglas Self.) The amp obviously doesn't know or care about the difference, but we do.

So, prodding the hornets' nest with my troll stick again: How can I state so confidently that a hi-fi amp doesn't interact with what it amplifies on any deep level? Am I justified in following Self's lead and laughing at claims that an amp can alter the rhythm of music?

And, coming back to pickups, I've heard claims that the construction of the pickup can alter the attack and dynamics of the guitar. What could be the physics behind this?

Added mass & that peskey string pull business??

[fieldwrangler]

(Sorry if this shows up twice, but I couldn't see it after my first attempt to post; it began as a response to Salversan's pointer to the Bobby Mc Ferrin pentatonic performance...)

That was pretty great! I have a hunch about why this group of pitches (intervals, really) is so universal. Seeing as how we've covered the 2:1 ratio,on to the 3:2 (perfect fifth.)

A while back I mentioned something about how looking at scales as a sequence of steps obscures some basic elements of these kinds of relationships, and the pentatonic is no exception. Since we all agree about this octave equivalence of pitches (in a general sense) let's look at the pentatonic as a series of perfect 5ths reinverted into the compass of one octave (crammed as close together as they can be) octave grounded to the first note in the series as a root.

The concept of Harmonic Distance, usually applied to key centers, can in fact serve as a guide to the hierarchy of intervals based on compound 5th content.

Looking at the familiar Circle of Fifths, notice that all the "major intervals" (C to D, C to A, C to E, C to B) are a function of clockwise elements that can be viewed as encompassing an increasing number of 5ths (2 to 5 in the above examples) as we go further out from our starting point. The so-called "minor intervals" (C to B flat, C to E flat, C to A flat, C to D flat) are a function of a similar counterclockwise path.

In fact, many feel that the intensity of the "majorness" increases as a function of the number of clockwise 5ths encompassed in this way. Similarly, the degree of "minorness" follows the counterclockwise track.

So, why stop at 5 notes separated by perfect 5ths? I think it's because this is as far as you can go witout introducing potential half steps in any inversion; this confers a stability and lack of potential implications of modulation.

Bob Palmieri

[David Schwab]

There is an interesting resonance between this viewpoint and that of audiophiles who think a sine wave is too easy a test for a power amp, and real music is somehow more challenging. (This argument is from my audio bullshit busting hero, Douglas Self.) The amp obviously doesn't know or care about the difference, but we do.

Testing with a sign wave would show non linear distortions, i.e. a lumpy sine wave, but you also have to test with waves like a square wave to see the skew rate; i.e. rounded over corners or ringing. Also you can't test for transients with sign waves.

I have a real world example. I had bought a new GK-800RB (bi-amped) bass amp back in '88. After my first rehearsal with the amp, I noticed that the top end seemed harsh and a little clippy. I brought it back to the store for servicing, and the tech couldn't find anything wrong with it. So I talked to him on the phone and he said he was looking at it on a 'scope running a sine wave in. I told him to plug a bass into it and play a few notes. He did, and said "oh wow! Now I see it!". It had a bad power supply to the high frequency power amp that was making it clip.

On unless you are only going to listen to sine waves all day, you need to see how an amp reacts to real music.

[Mike Sulzer]

Testing with a sign wave would show non linear distortions, i.e. a lumpy sine wave, but you also have to test with waves like a square wave to see the skew rate; i.e. rounded over corners or ringing. Also you can't test for transients with sign waves.

If you mean slew rate, a sine wave shows it up just fine. Sine waves become more like triangle waves as slew limiting occurs. This is a form of non-linear distortion; it occurs when there is insufficient current to drive a capacitance, typically the Miller effect capacitance of the second stage as seen by the output of the first stage.

[David Schwab]

If you mean slew rate, a sine wave shows it up just fine. Sine waves become more like triangle waves as slew limiting occurs. This is a form of non-linear distortion; it occurs when there is insufficient current to drive a capacitance, typically the Miller effect capacitance of the second stage as seen by the output of the first stage.

Yeah..slew rate. That's what happens when I'm trying to post when we have company! But yeah, what you said.

[Steve Conner]

You don't need the square wave though. If the amp can reproduce a 20kHz sine wave at its rated power with low distortion, it is more than good enough for music. As was mentioned earlier, music has a 1/f power density, so even this is overkill: real music never has anything like the slew rate of a full scale 20kHz sine wave.

Also, you can't hear the harmonics of 20kHz, so the amp could actually reproduce it with any shape whatsoever, and you wouldn't know the difference unless you smelt burning. 10kHz is a more relevant test frequency because the 2nd harmonic is audible.

The extra detail shown by inspecting the edges of a square wave is more than you need to know. It is of interest to the circuit designer of course, the sharp edges force the amp into slew limiting and allow you to check for any unpleasant glitches or hangup. But I think 10kHz THD at both 1W and rated power is a better predictor of the subjective quality of the amp's high end. (I'd love to see the 10kHz, 1W measurement for those old Japanese amps.)

As for the perfect fifth: My understanding is that it's the simplest ratio after 2:1. So it has the highest incidence of unisons and octaves among the partials of the two tones. For example, the 3rd harmonic of the lower tone and the 2nd of the higher tone form a unison, and so on. The other intervals are consonant to the extent that this holds true.

Of course this glib Pythagorean explanation only works for vertical harmony, how on earth does it work when the two tones sound one after the other? Mike's theory of tone prints in the brain is interesting, suggesting that melody and harmony are the same thing. But the flavour of, say, a semitone jump in a melody, isn't as sour as two notes a semitone apart sounding simultaneously.

[Mike Sulzer]

(Sorry if this shows up twice, but I couldn't see it after my first attempt to post; it began as a response to Salversan's pointer to the Bobby Mc Ferrin pentatonic performance...)

That was pretty great! I have a hunch about why this group of pitches (intervals, really) is so universal. Seeing as how we've covered the 2:1 ratio,on to the 3:2 (perfect fifth.)

A while back I mentioned something about how looking at scales as a sequence of steps obscures some basic elements of these kinds of relationships, and the pentatonic is no exception. Since we all agree about this octave equivalence of pitches (in a general sense) let's look at the pentatonic as a series of perfect 5ths reinverted into the compass of one octave (crammed as close together as they can be) octave grounded to the first note in the series as a root.

Looking at the familiar Circle of Fifths, notice that all the "major intervals" (C to D, C to A, C to E, C to B) are a function of clockwise elements that can be viewed as encompassing an increasing number of 5ths (2 to 5 in the above examples) as we go further out from our starting point. The so-called "minor intervals" (C to B flat, C to E flat, C to A flat, C to D flat) are a function of a similar counterclockwise path.

For C to D we move two fifths. This is 3/2 squared.
C to D: (3/2)^2 = 9/4; This is the second harmonic (9/4/2 = 9/8) of D. That is, 9/8 is the ratio of D to C in the just scale. That looks great! But I cannot get any of the others to work out exactly.

[Mike Sulzer]

Also, you can't hear the harmonics of 20kHz, so the amp could actually reproduce it with any shape whatsoever, and you wouldn't know the difference unless you smelt burning. 10kHz is a more relevant test frequency because the 2nd harmonic is audible.

Do not forget about intermod. 20 KHz - 19 KHz = 1 KHz. Certainly audible. You can also argue that if the program material fed to an amp contains material above 20 KHz, that will intermod down, and so something you could not hear directly will become audible as distortion products.

[kg]

i absolutely have come to believe that the behavior of an amp depends on some characteristics which are not visible on a 20-20k sine sweep.

[RedHouse]

I guess we all sit in front of a computer a large percentage of our days. ...

No.

"We" are busy making musical products, the computer is a back-burner device, often distracting to productivity, always a robber of better spent time.

[Steve Conner]

Mike, you're right, if the amp had bad THD at 20kHz, it would certainly have bad intermod too. And it is certainly a worry in this day of sigma-delta DACs that barf out ultrasonic noise by design. So I concede, maybe IMD is one of those characteristics KG is talking about.

I have tried measuring intermodulation in my hi-fi amps using a 24/96 audio interface and Spectrum Lab, but I was more or less stymied by all sorts of weird spurious responses that appeared even with the amp out of the loop, and were bigger than the amp's own contribution. Maybe op-amps, DACs, SPDIF jitter etc, all have their own HF IMD issues.

[kg]

Mike, you're right, if the amp had bad THD at 20kHz, it would certainly have bad intermod too. And it is certainly a worry in this day of sigma-delta DACs that barf out ultrasonic noise by design. So I concede, maybe IMD is one of those characteristics KG is talking about.

I have tried measuring intermodulation in my hi-fi amps using a 24/96 audio interface and Spectrum Lab, but I was more or less stymied by all sorts of weird spurious responses that appeared even with the amp out of the loop, and were bigger than the amp's own contribution. Maybe op-amps, DACs, SPDIF jitter etc, all have their own HF IMD issues.

i think the noise and crap from digital measuring equipment is a huge impediment. perhaps a good old fashioned 250mhz analog scope is best? of course frequency domain analysis is a bit more problematic..

[kg]

also, i think the DUT should be exposed to burst signal to demonstrate behavior/characteristics during high peak/avg transients. again, much like music signals.

in my experience, if i test the BAGA for steady state power output it's around 250wpc. however, the peak signal is considerably higher (more like 350wpc) for a few hundred milliseconds due to the cold bias and "soft" full wave voltage doubled plate and screen supply on the output tubes. the amp "feels" and "sounds" MUCH louder than a 500w total amp should feel and sound like because of the way it allows quick transients (ie pick attack, palm muting, etc.) to pass through. as a comparison it sounds somewhat like a low ratio compressor with a ~300ms attack/release time.

the nutshell is that i can only pick up that aspect of the amp with a burst measurement.. continuous sine/triangle/square won't show it at all.. and it has a drastic (and imo positive) effect on the response/liveliness/tone/dynamics of the amp in practice--so much so that i plan on using full wave voltage doubled supplies in future MI amps, without question.