"Gain", "Distortion", "Output": What's the difference?

[Mark Hammer]

I see these terms used interchangeably in ways that make me wince.  Of course, the more interchangeably they are used by folks who don't know any better, the more they are likely to be used by those who come along after them and draw their information from the young, under-informed and enthusiastic.

Now this is NOT a condescending slap on the wrists to those who are eager to learn and mod and rock the world.  I salute your intentions.  But you'll reach your desired objectives more efficiently and effectively if the right words are used for the right things, especially when asking questions of people or describing what you want.

So, to start, gain, distortion, and output ARE inter-related, but they are not the same thing.   Overweight people weigh a lot, and so do those with large and tall frames.  Obviously, "more person" = more weight.  But if you weigh 240lbs, I can't tell if you are 4'10" and horribly overweight, or 6'6" and muscular.  One of those is certainly a possibility, but it's not a certainty.  Similarly, I cannot automatically presume gain from output level, or output from distortion, and so on.

Gain refers to how much the input signal is increased within a stage.  Gain is multiplicative, such that a stage with a gain of 10x, followed by another 10x gain stage, with nothing between them to get in the way, results in a total gain of 100x.

Because the stage/device's ability to provide gain is limited by the power supply, and properties of the device (like how wide the device can "swing" within the boundaries of the power supply voltage, or how much maximum gain it can provide at a given frequency), a stage may produce distortion if called upon to produce more gain than it can successfully deliver in clean fashion.

Of course, that can't-do-any-more point is partly dictated by the input signal itself.  A circuit running off a 9v battery can certainly amplify a 1-microvolt signal 10,000 times without running out of headroom, but it can't amplify a 100 millivolt signal that much.

Similarly, the diodes often found in distortion circuits (and transistors used as diodes) have an absolute threshold at which they conduct, and their clipping properties become manifest.  Should the threshold be particularly high and the signal level be particularly low, then the signal requires that more gain be applied to it to reach that threshold.  If the threshold is low-ish, and the signal itself hot enough, then it doesn't take much gain to produce distortion.

So, gain can result in distortion, but isn't always necessary for it, and sometimes isn't even enough for it.  Like I say, they are two inter-related-but-separate things.

Often, when people say they "want more gain" from a circuit, they really mean that they want a more obvious and intense distortion.  They assume that more gain will produce more distortion, but that may not necessarily be true.  If a given circuit has a high threshold of clipping, a more distorted ("gainier") sound can be achieved without touching the gain a single bit, but lowering the threshold instead.

Does gain, and/or distortion, always result in greater output?  Not necessarily.  For example, a circuit may have wads of gain, and produce considerable distortion, but lose enough of the signal via a passive tone-shaping circuit, that the output is likely to be fairly modest.  The EHX Big Muff Pi and the Shin-Ei FY-2 Companion Fuzz are examples of this.  Both lose a great deal of signal via the tone/filter circuit that comes after the fuzz portion.  In the case of the BMP, EHX added a "gain recovery" stage after the tone circuit to bring the level back up.  In the case of the FY-2, Shin-Ei did not, so the output, while generally louder than bypass, is still quite modest.

Remember that, even though gain IS multiplicative, the multiplying factor applied to whatever is acquired from the previous stage, is applied to the signal level.  If something between stages A and B knocks the level down, or constrains it in any fashion (e.g., by clipping), then the gain of stage B is applied to a smaller output from stage A than you might otherwise expect.

Let's say I feed stage A with a 100mv signal and apply a gain of 20x to it.  I then apply a gain of 30x in stage B.  Is the output necessarily going to be a total of 500 x 100mv =  50v?  Well, certainly  if you use a 9v supply, it will clip like crazy because you can't produce 50v from a 9v supply.  But let us say that there was a tone circuit between A and B, that reduced the signal coming from A  from a hypothetical 2V down to 200mv.  Applying a gain of 30x to that (= 6v) still leaves us pretty much within the capabilities of what a 9v battery can do.  So, the overall gain applied by the circuit is high (500x), but the distortion can be quite modest, by virtue of how the gain is distributed.

Of course, if that second stage does nothing to clip or otherwise reduce the signal, the output level will be VERY hot.  Again, gain, distortion, and output level are NOT one in the same.
Because distortion is the addition of harmonic content that was not previously there, or there in such large proportions, greater clipping generates more treble.  Conversely, if much of that treble is removed, there is the illusion that less gain has been applied, or rather than the amount applied has not been enough to produce serious distortion.  Again, we call upon the Big Muff as an example.  Set the tone control to full bass, and it doesn't sound quite as distorted or "gain-ey", even though the gain is not altered in any way by the position of the tone control, nor has the amount of distortion been altered.  What HAS been altered is what segments of the resulting signal you are allowed to hear.

So, to cut this off before it gets too ponderous.....(though feel free to add to it)

GAIN can produce more distortion and greater output levels when the circumstances permit.

DISTORTION is more likely to occur when circumstances permit, but does not necessarily require huge amounts of gain, nor does it necessarily result in humungous output levels unless circumstances permit.

OUTPUT/LOUDNESS can be achieved with fairly modest gain and minimal distortion.

If you want more loudness, do not ask for how to achieve more gain.  If you want more distortion, do not ask for how to achieve more gain.  Always ask for the one you want.

[Johan]

just to another thing I've noticed lately.. a"buffer" is not a voltage amplifier. meaning it doesnt ad volume( or in Mark's words, Loudness ) and doesnt "boost". . it does ad strength in the form of current amplification meaning you can drive much lower(heavier) loads, like a 10k input instead of a 1M input, without loosing your signal. but it doesnt make it any louder.
J

[R.G.]

I wish you luck, Mark, but I have this feeling that you're shoveling sand against the tide. 

[anchovie]

Blame Marshall for the "gain" definition problem -  I know non-musicians that are aware that the Gain knob on a Marshall amp alters the level of distortion. I guess they wanted a label associated with a more direct effect than "preamp volume".

[Mark Hammer]

I wish you luck, Mark, but I have this feeling that you're shoveling sand against the tide. 

If I was trying to stop the ocean, that'd be one thing.  But I'm just trying to help our little cove communicate internally in a more effective manner.  I figure a little shovelling is good exercise!

[ayayay!]

Hmm... purely based on this post, I'm thinking of now labeling my gain pots "Volume", and my volume pots "Master."  I mean really, why not? 

[Mark Hammer]

Label them for what they mean to you, ought to mean to others, and for what they do.

As for me, I'm just tryng to make it so that when some kid pops in and asks how to make their pedal have "more gain", the rest of us are not scratching our heads wondering if they really mean they want the distortion to sound more intense, if they like the sound but want it to be louder at max volume, or simply want to improve the signal to noise ratio.

As I learned in my problem-solving and reasoning class so many years ago (thank you Prof. Marley), the hardest and most important part of solving a problem is identifying and defining it.  Once you know what the problem is, the rest is usually pretty easy.

[Gurner]

just to another thing I've noticed lately.. a"buffer" is not a voltage amplifier. meaning it doesnt ad volume( or in Mark's words, Loudness ) and doesnt "boost". . it does ad strength in the form of current amplification

IMHO it is a voltage amplifier  ...current magnification you speak of, comes from the increased signal voltage (ie gain applied by the buffer)  at the opamp output (assuming the same load)

For example if your buffer 'load' is 1K & your AC signal output is 1V RMS, then 1mA will flow out of your buffer. If your signal is 2V RMS & your load remains 1k, then 2mA will flow out of the buffer  - you've doubled the current, but only as a result of doubling the RMS signal voltage....and the voltage doubling was the result of voltage amplification.

edit: I guess we need to iron out what buffer means! To me, a buffer is the use of a tranny/opamp etc to stop a signal being loaded (&(it will often be unity gain ...but can have gain too) .... now re-reading your post, I'm taking it that by 'buffer' you likely mean a unity gain opamp arrangement device? Ok, true.... without a buffer, you'd have less current delivery, with the buffer you have more on tap ...therefore you've upped the ability to deliver current (so it's a terminology thing - do you call more current on tap - current magnification. More current maintening -no?)

[Fender3D]

Nice try Mark,
...but, you should agree that a more technical language requires a more technical acknowledgement, I guess the average forum user simply copies and pastes from every layout he finds, then, usually asks when some trouble occours. (CTRL+C, CTRL+?)  
... and this all without asking whether Ohm is a multimeter builder or someone else  

Edit:
BTW my "gain" is spelled "drive" usually...

[DougH]

What I want to know is what does "mix" and "ability" mean?...

[Mark Hammer]

just to another thing I've noticed lately.. a"buffer" is not a voltage amplifier. meaning it doesnt ad volume( or in Mark's words, Loudness ) and doesnt "boost". . it does ad strength in the form of current amplification

IMHO it is a voltage amplifier  ...current magnification you speak of, comes from the increased signal voltage (ie gain applied by the buffer)  at the opamp output (assuming the same load)

For example if your buffer 'load' is 1K & your AC signal output is 1V RMS, then 1mA will flow out of your buffer. If your signal is 2V RMS & your load remains 1k, then 2mA will flow out of the buffer  - you've doubled the current, but only as a result of doubling the RMS signal voltage....and the voltage doubling was the result of voltage amplification.

edit: I guess we need to iron out what buffer means! To me, a buffer is the use of a tranny/opamp etc to stop a signal being loaded (&(it will often be unity gain ...but can have gain too) .... now re-reading your post, I'm taking it that by 'buffer' you likely mean a unity gain opamp arrangement device? Ok, true.... without a buffer, you'd have less current delivery, with the buffer you have more on tap ...therefore you've upped the ability to deliver current (so it's a terminology thing - do you call more current on tap - current magnification?)

I don't want to get embroiled in a discussion addressing details that mean something to only a few people, but I will concur that a guitar can sound louder if it conserves bandwidth.  Higher amplitude can make things stand out more, but then so can better definition through high frequency content.  So, even if a buffer stage does not add gain, it can still make something seem louder.

[R.G.]

IMHO it is a voltage amplifier  ...current magnification you speak of, comes from the increased signal voltage (ie gain applied by the buffer)  at the opamp output (assuming the same load)

For example if your buffer 'load' is 1K & your AC signal output is 1V RMS, then 1mA will flow out of your buffer. If your signal is 2V RMS & your load remains 1k, then 2mA will flow out of the buffer  - you've doubled the current, but only as a result of doubling the RMS signal voltage....and the voltage doubling was the result of voltage amplification.

The industry thinks that "buffer" means "has a high input impedance to not load a signal source, and has a lower output impedance than the source so it can drive more loading than the original source without being loaded down". It's all about source/load impedance and "protecting" the signal source from overloading. Voltage gain is not an issue, and may be more or less than unity, although most people in the industry think of a buffer as having a nominal gain of one or very near it. In particular, buffers are not conceived of in most cases as having more output current because they have more voltage output.

I guess we need to iron out what buffer means! To me, a buffer is the use of a tranny/opamp etc to stop a signal being loaded (&(it will often be unity gain ...but can have gain too)

This is a correct concept, as long as the semantics to you mean "more current available, shielded from overloading" as the main concept, not "more voltage".

.... now re-reading your post, I'm taking it that by 'buffer' you likely mean a unity gain opamp arrangement device? Ok, true.... without a buffer, you'd have less current delivery, with the buffer you have more on tap ...therefore you've upped the ability to deliver current (so it's a terminology thing - do you call more current on tap - current magnification. More current maintening -no?)

You call "more current available"  'lower impedance'. This is, by definition, the ability to deliver more current to a load without the voltage sagging.

Buffers have a high(ish) input impedance and a lower (or MUCH lower) output impedance than what drives them. Voltage gain is often unity, but does not have to be.

[BubbaFet]

 Perhaps the audience you are trying to educate would
be better reached if you created a Facebook page
called 'Gain, Distortion, and Output'.... oh, and a companion
YouTube video as well, with a funny cat or three.

And what exactly do you mean by the terms 'mark' and 'hammer'? 

[Johan]

And what exactly do you mean by the terms 'mark' and 'hammer'? 

.."Mark of the Hammer"...

[victor nery]

Hi, I'm a begginer and this post made things a lot clearer. Yet I'm not able to perform changes in a circuit to achieve a better performance in each of these aspects. Mark would you mind giving an example? In practice, how to obtain more gain, distortion and output from let's say an MXR+? Thanks to you all from DIY stompboxes. You make a big difference in the DIY world.

[Mark Hammer]

Well thanks.  That's very kind of you to say, and certainly appreciated.

In the case of the MXR Distortion+, gain distortion, and output are related up to a point.  Certainly, increasing the gain of the single puny op-amp stage WILL result in more distortion.  But since the germanium diodes place a severe limit on the maximum output by "clamping" the signal t a low amplitude, attempting to increase the gain of the stage will get you more distortion, but not a whole lot more output volume.

Can you get more output volume independent of the gain?  Yes.  For the D+, the original issues used a 10k volume pot.  In combination with the 10k resistor just ahead of the output, that behaved like a 20k pot that could never be turned up more than halfway.  By replacing the 10k pot with a 50k or even 100k, more of the signal was allowed to leave the circuit, because the 10k + pot was not attenuating as much.

You can get still more output level from the unit by raisig the clipping threshold.  replacing the germanium diodes with silicon, or perhaps even using a 1N914+1N34 in series (in each direction) will raise the clipping threshold and allow for a much more robust output signal without doing anything whatsoever to the gain of the circuit.  Of course, by having a higher clippipng threshold, it means the distortion will not be as intense.

If you wanted that output level but still wanted to reclaim the distortion intensity, you could still increase the maximum gain of the stage.  Stock, the gain ranges between a little over 2x to a maximum  of 213x.  The gain goes up as the resitance of the "ground leg" gets smaller.  With a minimum resuistance of 4k7 the gain is 213x.  Drop that minimum resistance to 2k2, and that max gain is increased to 455x.  naturally, values between 2k2 and 4k7 get you max gains between those.

make sense?

[petemoore]

 Distortion = copy not the same as original.
 Distortion is presented in all instances where aural sensation is involved.
  At a fundamental level, eardrums 'distort' from the neutral position when encountering a sound pressure impulse.
Gain = A copy that is larger than the original, {for wire users like us.
 Stronger or with greater potential, pertains to voltage.
Output = what is put out.
 Output variables can be quantified in various ways such as 'more'.
Gain invariably adds ''distortion'' and increased 'output', though this may be percieved as distortion or even reduced output under certain circumstances.  

[victor nery]

Yes. Now I can get what I need of this circuit. The only thing I still don't understand are voltage dividers and their influence in the circuit. Are they also related to gain, distortion, output and frequency response? What's the difference between a pair of 1M and a pair of 22k? Thanks

[derevaun]

Another thanks from a relative newcomer. It's funny, I was just yesterday web searching to better understand the differences between overdrive, distortion, and fuzz. So this thread has been helpful with that, too.

I share your concern for correct terminology. But among words, meaning and usage, it's really usage that has the biggest effect on the other two. Words become what they're used for, and their meanings change as usage changes. My pet peeve is "sea change;" it's been misused so consistently that everyone assumes it to mean that the momentum of seas have themselves has turned around, which is totally not the correct, original meaning of the phrase. Nevertheless, the socially constructed meaning is the one that you have to assume people will "get" when you say the word.

But, as they say, you can't fly a plane on a social construction, and we need to know what is meant by technical terms. Coming from a non-technical background, I'm always glad to see someone just come out and explain stuff.

[R.G.]

Yes. Now I can get what I need of this circuit. The only thing I still don't understand are voltage dividers and their influence in the circuit. Are they also related to gain, distortion, output and frequency response? What's the difference between a pair of 1M and a pair of 22k? Thanks

Impedance.

!?!!?

A voltage divider not only divides the voltage, it does so by using resistors (these are what 'impedance' is at DC) to do it. That means that the resistors not only divide the voltage, they "eat" some of the current into the divider as a price for doing the division, and let less of the current out at the "divided" output than would otherwise be available. A pair of 1M resistors making a 1/2-voltage divider can provide much less current than a pair of 22K resistors without its voltage sagging badly. Some of these concepts are explained in the article at GEOFEX which talks about choosing resistors for a Vref divider.

Voltage dividers - obviously - can be used as a kind of anti-gain, a reduction in signal level instead of a gain in level. If made of resistors (and to some extent, inductors and capacitors) they have *zero* distortion in the divider, and where distortion is caused by high signal level, reduce distortion of other things by reducing signal below the distortion onset. They reduce "output" because they reduce the amount of signal. Resistors, to a first level of approximation, cause zero effect on frequency response. Because of stray capacitance, there may be some frequency response change, but it can be compensated for by deliberately adding capacitance in certain ways to make a "frequency-neutral" divider.

That is - it's not simple to say what voltage dividers do in the terms you asked.