Head-Fi.org › Forums › Equipment Forums › Sound Science › Gain vs. Volume
New Posts  All Forums:Forum Nav:

Gain vs. Volume

post #1 of 17
Thread Starter 

Hello head-fi,

 

Most amps have a gain setting that determines the voltage "gain," which usually take the form of a click-switch mechanism that selects resistors or occasionally potentiometer. Another knob (connected to a potentiometer) controls volume.

 

What is the difference between gain and volume? By my understanding, a change in the gain knob changes the voltage increase. If so, what does the volume knob change? Does it change line out resistance or does it change how the OP amp/discreet amp's voltage gain as well, or is it something completely different? Why does increasing gain tend to increase distortion whilst increasing volume does not significantly?

 

Also, is it correct that by increasing the voltage input to the headphones the volume changes and current is determined by the voltage (V) and resistance of the headphone + amp (Ohm) AND that the volume of the headphone is determined by the amount of power flowing though where power = V*I.

 

By Ohm's law V=IR, so higher resistance = less current = less power = lower volume (assuming same sensitivity).

 

Thanks :3

post #2 of 17

In simple amps, it looks like this: first, there's the volume potentiometer, then there's an opamp with fixed voltage gain resistors. If the volume is set to the max, 1 V of input results in 2 V of output if the voltage gain is set to 2x.

By reducing the volume using the potentiometer you also decrease the signal to noise ratio (the lower the volume, the closer the signal is to the noise floor). By setting the gain higher you don't just amplify the signal but also the noise.

 

See the upper part of the cmoy schematic:

CMOY_SCHEMATIC.png

 

A potentiometer is basically a variable voltage divider.


Edited by xnor - 6/22/12 at 7:40am
post #3 of 17
Sometime the volume control (pot) and gain are the same thing, sometime they are not. It depends how the volume control mechanism is implemented.

For cmoy style control, pot is placed between the input and the opamp, so the pot is controlling the level of input signal while the opamp is set to amplified on a certain gain (which is usually set by the ratio of two resistors). For PA2V2 style control, pot is placed in the gain control, using the pot to change the ratio to another resistor so the gain is changed as well. In this case, gain change is volume change. Gain is where opamp is actively trying to increase the signal strength, and thus it is where noise and distortion are introduced. So the higher the gain, the harder the opamp need to work and the more noise is introduced. Which is why PA2V2 style of gain/volume control is generally not in favored since turning up volume means more distortion.

There is also the rarer kind of volume control that is placed after the opamp (gain) stage, acting as an attenuator. This is more likely used as pre-amp rather than for driving headphone directly. Also, there is digital volume control which is usually more similar to PA2V2 style of gain/volume control, but done in the digital domain and has less side-effect.

These are just some of the concepts I learned, might not be that correct or completed (*No EE background whatsoever tongue.gif).
Edited by ClieOS - 6/22/12 at 7:46am
post #4 of 17
Thread Starter 

So in other words the amp produces a set level of voltage gain determined by the gain resistor+OP amp and the volume pot divides the voltage so only a fraction of the voltage is actually being applied to the jack?

post #5 of 17
Quote:
Originally Posted by iamthem View Post

So in other words the amp produces a set level of voltage gain determined by the gain resistor+OP amp and the volume pot divides the voltage so only a fraction of the voltage is actually being applied to the jack?

 

Bingo.

 

Basically, the amplifier applies gain to whatever is fed into it and the volume control attenuates whatever is fed into it. Though there have been some rare implementations where the volume control is used to change the gain of the amplifier.

 

se

post #6 of 17
Quote:
Originally Posted by iamthem View Post

So in other words the amp produces a set level of voltage gain determined by the gain resistor+OP amp and the volume pot divides the voltage so only a fraction of the voltage is actually being applied to the jack?

No, the volume pot divides the voltage so that only a fraction of the voltage is applied to the opamp, in the case of a cmoy anyway.

So it looks like this: input -> volume pot -> gain stage & buffer (a single opamp) -> output.

 

The O2 on the other hand looks like this: input -> gain stage (configurable with gain switch) -> volume pot -> buffer (unity gain) -> output.


Edited by xnor - 6/22/12 at 8:07am
post #7 of 17
Quote:
Originally Posted by xnor View Post

In simple amps, it looks like this: first, there's the volume potentiometer, then there's an opamp with fixed voltage gain resistors. If the volume is set to the max, 1 V of input results in 2 V of output if the voltage gain is set to 2x.

By reducing the volume using the potentiometer you also decrease the signal to noise ratio (the lower the volume, the closer the signal is to the noise floor). By setting the gain higher you don't just amplify the signal but also the noise.

 

See the upper part of the cmoy schematic:

CMOY_SCHEMATIC.png

 

A potentiometer is basically a variable voltage divider.

 

And Grado charges how much for a variation on this? Puh'leeeze!

post #8 of 17
Quote:
Originally Posted by ClieOS View Post

Gain is where opamp is actively trying to increase the signal strength, and thus it is where noise and distortion are introduced. So the higher the gain, the harder the opamp need to work and the more noise is introduced. Which is why PA2V2 style of gain/volume control is generally not in favored since turning up volume means more distortion.

 

That is not the real reason, in fact, the variable gain method would use lower gain than the fixed gain one whenever the volume is not at the maximum (i.e. almost always), therefore, if the potentiometer was an ideal resistor, it would normally have lower distortion and noise. However, potentiometers are in reality quite bad quality resistors, and having them in a feedback loop can degrade performance and stability. It also requires using the amplifier in an inverting configuration, since the non-inverting design can only have a gain of at least 1 (which is likely often too loud as a minimum volume). Additionally, it does not handle contact problems in the potentiometer very well, since they result in very high gain (i.e. loud artifacts) instead of the sound just cutting out temporarily.

 

Quote:

Originally Posted by ClieOS View Post

There is also the rarer kind of volume control that is placed after the opamp (gain) stage, acting as an attenuator. This is more likely used as pre-amp rather than for driving headphone directly.

 

This is the method used in the O2 amplifier. It has several advantages (significantly lower noise, stronger negative feedback in the output stage so reduced distortion, and output op amps with unity gain can be paralleled easily for higher current output and even lower output loading distortion), but it also requires a more careful selection of gain, since a too high gain will always clip, regardless of the volume setting.

 

Quote:
Originally Posted by ClieOS View Post

Also, there is digital volume control which is usually more similar to PA2V2 style of gain/volume control, but done in the digital domain and has less side-effect.

 

No, it is most similar to the "volume before gain" method, and it always outputs the maximum noise level, regardless of the volume. It is very common in PC sound hardware.

post #9 of 17
Quote:
Originally Posted by stv014 View Post

No, it is most similar to the "volume before gain" method, and it always outputs the maximum noise level, regardless of the volume. It is very common in PC sound hardware.

I don't think that is true. A digitally controlled volume control doesn't mean that the volume has to be attenuated in the digital domain. In fact, many PC audio chips use voltage controlled amplifiers.


Edited by xnor - 6/23/12 at 3:59am
post #10 of 17
Quote:
Originally Posted by xnor View Post

I don't think that is true. A digitally controlled volume control doesn't mean that the volume has to be attenuated in the digital domain. In fact, many PC audio chips use voltage controlled amplifiers.

 

By "digital volume control", I did mean digital attenuation, rather than digitally controlled analog volume. The former seems to become very common in PC audio hardware, as it is cheaper, and newer DACs have high enough dynamic range to make it an acceptable (even if often not ideal) solution.

post #11 of 17

What do you mean with common PC audio hardware? Cheap USB DACs?

post #12 of 17
Quote:
Originally Posted by xnor View Post

What do you mean with common PC audio hardware? Cheap USB DACs?

 

I meant onboard audio and sound cards.

post #13 of 17
Quote:
Originally Posted by stv014 View Post

 

This is the method used in the O2 amplifier. It has several advantages (significantly lower noise, stronger negative feedback in the output stage so reduced distortion, and output op amps with unity gain can be paralleled easily for higher current output and even lower output loading distortion), but it also requires a more careful selection of gain, since a too high gain will always clip, regardless of the volume setting.

 

 

No, it is most similar to the "volume before gain" method, and it always outputs the maximum noise level, regardless of the volume. It is very common in PC sound hardware.

 

 

Quote:
Originally Posted by xnor View Post

I don't think that is true. A digitally controlled volume control doesn't mean that the volume has to be attenuated in the digital domain. In fact, many PC audio chips use voltage controlled amplifiers.

 

you guys are talking about 2 different things; there is Digitally controlled analogue attenuation and there is digital attenuation. the former often uses a buffer opamp on the input, then a sort of internal resistor ladder to control the gain, which outputs through a further buffer opamp. the highest performance of these will drive a 600ohm or higher load directly, but for most headphone loads will require another buffer at a minimum so you cant use it at the output alone, this seems far too much extra components and feedback for a volume control in my book. dont get me wrong, i'm not a no feedback pundit, but there is such a thing as too much, for too little.

 

I'm not a huge fan of this method as it basically limits and filters the quality of whatever you have previous to this, which may actually be higher performance. above audibility? perhaps, but it just seems inelegant to me. the idea of going to all that work for a clean and linear signal only to do that just to control the volume. …

 

Digital volume, particularly modern digital volume that is either internal to the DAC (like with AKM, ESS etc which operate at ≥32bit (40 for ESS)) or prior to that at even higher resolution (up to 64bit) has a noise floor that is so low its really not worth worrying about; as should the DAC.

 

Analogue controls are OK for analogue sources, but for digital sources I dont get it. You may lower the noise from the dac, but then you add your own voltage noise from the resistors, add error due to them not matching and not insignificant expense to do it properly.


Edited by qusp - 6/23/12 at 6:32am
post #14 of 17
Quote:
Originally Posted by qusp View Post
Analogue controls are OK for analogue sources, but for digital sources I dont get it. You may lower the noise from the dac, but then you add your own voltage noise from the resistors...

 

Don't see that being a problem unless you've got Schiitload of gain after the attenuator, which for headphones, you're not going to have.

 

se

post #15 of 17
Quote:
Originally Posted by stv014 View Post

 

I meant onboard audio and sound cards.

 

Both my desktop and netbook's onboard sound gives increased noise when I crank up the master volume with nothing playing.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Gain vs. Volume