Confused about volume

[my_ears_are_up_here]

Hello! First time poster, long time objectivist...

I've spent most of my life appreciating flat response, spent most of my time, money, and energy towards that in as thrifty way as reasonable.

Recently I stumbled onto these fora, Amir's reviews and videos, and was transformed from "flat is gold" to "distortion free is gold". (I had tolerated "it's okay" distortion, previously, thinking flat was most important.)

Using the guidance of all the resources here, I built my first very low distortion setup for the bedroom:

- Elac B6.2
- Aiyima A07
- Lavaudio DS200Pro LDAC BT receiver

I'll grab a proper DAC, EQ, etc shortly, but wanted to see, first, if I enjoyed the low distortion optimization target.

Herein lies the rub: historically, I've just gotten "good enough" integrated amps that handle everything. Fine for convenience, but obviously not ideal. So my experience with separate components is totally zilch.

I understand the digital side of audio quite intimately, but the analog side is very foreign to me. Which brings me to my question: (sorry for all the preamble)

One thing I cannot seem to unpack is the role and positioning of volume control in the chain. Maybe I missed it, but a search of these fora didn't reveal anything, and advice on the internet at large is all over the map.

I understand that a power amp is not *technically* a power amp if it has volume control (at which point it's an "integrated amp"), but where is that volume control in the chain, exactly?

In my setup, for example, my amp has a volume control. The (mediocre) preamp in my BT receiver *also* has volume control. (By way of changing the BT volume, which I *assume* either raises/lowers the gain of the preamp, or more likely, changes the shape of the digital signal before it's amplified.)

So which do I use for actual volume control? (In respect to quality.) Do I keep the amp volume up at whatever maximum level I'd ever want, then control the volume with the Bluetooth controls? Or do I keep the Bluetooth at/near maximum and adjust the listening level with the amp volume knob?

In short:

- What is the specific role of the volume knob in a simple integrated amp such as mine?
- Is it better to attenuate the input signal digitally (before preamp) or after/analog?

And I guess, fundamentally, my question is:

- Is amplification fixed/continuous and we're just attenuating the input signal? Or does it attenuate the output power? (I guess that wouldn't make sense, would it? Be very hot and wasteful.)

(I may have answered my own last question, just thinking through the practicality of it, but still confused where the best place to attenuate the signal is.)

Sorry this is so long and verbose! I guess (over) clarity is a feature of objectivism! Glad to join the community!

 

[my_ears_are_up_here]

I should note: subjectively, it seems to sound better/cleaner with the BT receiver at full volume vs lower with the amp at mid/high volume, but I have no way to measure it.

I would guess amplifying the weak signal — possibly closer to the distortion floor — is worse than higher preamp level + lower amp level, but since I don't quite understand what's happening in the chain, it's hard to be sure.

Does that sound like a reasonable idea what's going on? Or is this all in my head? It's subtle, but not imperceptible.

 

[Chazz6]

I have no theoretical knowledge here, just my experience with a preamp feeding into another preamp or into an integrated amplifier. I just tried different combinations of volume levels that brought the volume to where I wanted it. Generally, one combination was better than another.

I think that some components have better volume controls than others. And that most controls do better in the broad middle rather than at the softest or loudest. A weighted combination favoring one component produced the best result, the least background noise during quiet passages, etc.

 

[Doodski]

Hello! First time poster, long time objectivist...

Welcome to ASR.

I understand that a power amp is not *technically* a power amp if it has volume control (at which point it's an "integrated amp"), but where is that volume control in the chain, exactly?

Generally it is at the front panel in a convenient location ready for immediate use. If it is placed at the rear panel or behind a small door then that is probably a power amp and not a integrated amp. The location in the circuitry is after the input selection and bass and treble controls and before the amplifier circuitry.

So which do I use for actual volume control? (In respect to quality.) Do I keep the amp volume up at whatever maximum level I'd ever want, then control the volume with the Bluetooth controls? Or do I keep the Bluetooth at/near maximum and adjust the listening level with the amp volume knob?

Good question. One we get fairly often here at ASR but it never gets old. There are generally two different kinds of volume controls. There is the analogue potentiometer type and then there is a digital control. The digital control usually has better accuracy and that means the variation/tolerances between the left and right channels are smaller. With the analogue control they are prone to tolerance variations in the carbon resistive element used in the volume control assembly and that causes the left and right channels to have a imbalance where one channel is a bit louder than the other. Kind of like moving the balance control to one side or the other and one channel gets louder.

If my gear then I would turn up the analogue control on the integrated amp to ~maybe 11 o'clock to 1pm and then use the bluetooth volume control to adjust the volume level. I would do this because the digital control in the bluetooth stuff has better channel to channel accuracy and is less prone to one channel being slightly louder than the other. The volume setting on the amp does not have to be at 11am to 1pm but those are good starting points. If that is too sensitive then turn/dial down the knob a bit or put it a bit higher if you need more power output. This is not a exact science. It's a combination of what volume control to use (digital or analogue) and at what levels to set them at and that combination can be varied considerably to get them where you are comfortable and things make sense. The long and short of it is digital is better but analogue is pretty good. Analogue can be very accurate and have tight tolerances although the manufacturing of such super accurate analogue volume control assemblies is expensive and they cost muchO.

- Is amplification fixed/continuous and we're just attenuating the input signal? Or does it attenuate the output power? (I guess that wouldn't make sense, would it? Be very hot and wasteful.)

The amplifier circuitry is set at a fixed gain. It is basically sitting there ready to amplify whatever comes in and to do that at the gain level that the designer made it capable of. The amplifier circuitry is not variable in the gain as that is done in the active pre-amp section that amplifies the voltage that comes into the pre-amp section.

 

[DVDdoug]

Generally the main volume control should be last in the chain because of noise, but if you can't hear the background noise it's not an issue.

All analog electronics generate some noise and anything with gain will amplify any existing noise. If you have a strong signal into the power amp and then you turn down the volume at the amp's input, you are reducing the signal and noise into the power amp. But any noise generated inside the power amp remains. You are tuning-down the signal and noise together. Any time the volume is reduced before the noise is generated, you're not turning down the noise and the signal-to-noise gets worse... Simple!

But most of the time, a "preamp" isn't really amplifying (unless you have a phono preamp) so the preamp (or the preamp part of an integrated amp) isn't generating much noise anyway.

I don't claim to have "golden ears" but I've NEVER heard distortion from anything that wasn't either broken or over-driven into clipping.
Preamps usually have plenty of headroom, so in most setups they never clip. Of course you can clip your power amplifier by trying to get too much power out of it. And you can clip the digital data (by boosting with EQ, etc.).

When you reduce the volume digitally, you do loose resolution. You can't hear the loss of quality because the resulting quantization noise remains low, where you can't normally hear it. But you might hear it if you re-amplify, similar to how you might hear noise if you reduce the analog volume and then re-amplify. It's not something I worry about but if you want "bit perfect" digital audio, of course you can't adjust the volume digitally.

I understand that a power amp is not *technically* a power amp if it has volume control (at which point it's an "integrated amp"),

A power amp can have a volume/gain control. Some pro power amps have DSP EQ so you can use them for bi-amplification or tri-amplification, etc., without an external active crossover. An integrated amp (or preamp) has multiple-selectable inputs, and usually tone controls and other "switches & knobs".

 

[ADU]

I am not familiar with any of the gear you are using. But I think I'd leave the digital volume on your source (presumably a phone or computer?) at maximum though.

Set the volume on the BT receiver to its equivalent of unity gain. And adjust the power amp to the max volume you'd generally need. Typically about 85 dB for a -20 dBFS pink noise test signal, or 75 dB for a -30 dBFS pink noise test signal, which is more common on alot of home receivers. This is if your components can handle that volume. It can certainly be set lower (or a bit higher) if you prefer though.

Then use the volume control on the BT receiver as your master volume control for any on-the-fly adjustments. -10 dB below the unity setting is a common volume setting for alot of home systems. That effectively gives you 75 dB with -20 dBFS pink noise, and 65 dB with -30 dBFS pink noise. If you find that too loud for most listening, then you might want to lower the volume on your power amp a bit, so that unity on the receiver is 5 or 10 dB lower than the above reference values.

 

[ADU]

If you don't have a way to actually measure the volumes, then just set the volume on the BT receiver to approximately -10 dB below its normal or nominal volume setting. And adjust the volume on the power amp to a comfortable listening level, and leave it there.

Use the volume control on the BT receiver for on-the-fly adjustments.

 

[my_ears_are_up_here]

Welcome to ASR.


Generally it is at the front panel in a convenient location ready for immediate use. If it is placed at the rear panel or behind a small door then that is probably a power amp and not a integrated amp. The location in the circuitry is after the input selection and bass and treble controls and before the amplifier circuitry.


Good question. One we get fairly often here at ASR but it never gets old. There are generally two different kinds of volume controls. There is the analogue potentiometer type and then there is a digital control. The digital control usually has better accuracy and that means the variation/tolerances between the left and right channels are smaller. With the analogue control they are prone to tolerance variations in the carbon resistive element used in the volume control assembly and that causes the left and right channels to have a imbalance where one channel is a bit louder than the other. Kind of like moving the balance control to one side or the other and one channel gets louder.

If my gear then I would turn up the analogue control on the integrated amp to ~maybe 11 o'clock to 1pm and then use the bluetooth volume control to adjust the volume level. I would do this because the digital control in the bluetooth stuff has better channel to channel accuracy and is less prone to one channel being slightly louder than the other. The volume setting on the amp does not have to be at 11am to 1pm but those are good starting points. If that is too sensitive then turn/dial down the knob a bit or put it a bit higher if you need more power output. This is not a exact science. It's a combination of what volume control to use (digital or analogue) and at what levels to set them at and that combination can be varied considerably to get them where you are comfortable and things make sense. The long and short of it is digital is better but analogue is pretty good. Analogue can be very accurate and have tight tolerances although the manufacturing of such super accurate analogue volume control assemblies is expensive and they cost muchO.

The amplifier circuitry is set at a fixed gain. It is basically sitting there ready to amplify whatever comes in and to do that at the gain level that the designer made it capable of. The amplifier circuitry is not variable in the gain as that is done in the active pre-amp section that amplifies the voltage that comes into the pre-amp section.

Thank you for clarifying the pitfalls of both attenuation methods. I will experiment with the levels and see if I can find a happy middle ground. The analog control seems to do fine; there's no perceptible L/R balance issue. (Though that may be hidden by the fact that ideal speaker placement is impossible in this room, though it'll be more valuable in rooms in which I have more control.)

I'm not sure I trust the DAC on the BT receiver, so when I get a dedicated DAC, I'll play around more with it.

Generally the main volume control should be last in the chain because of noise, but if you can't hear the background noise it's not an issue.

All analog electronics generate some noise and anything with gain will amplify any existing noise. If you have a strong signal into the power amp and then you turn down the volume at the amp's input, you are reducing the signal and noise into the power amp. But any noise generated inside the power amp remains. You are tuning-down the signal and noise together. Any time the volume is reduced before the noise is generated, you're not turning down the noise and the signal-to-noise gets worse... Simple!

But most of the time, a "preamp" isn't really amplifying (unless you have a phono preamp) so the preamp (or the preamp part of an integrated amp) isn't generating much noise anyway.

I don't claim to have "golden ears" but I've NEVER heard distortion from anything that wasn't either broken or over-driven into clipping.
Preamps usually have plenty of headroom, so in most setups they never clip. Of course you can clip your power amplifier by trying to get too much power out of it. And you can clip the digital data (by boosting with EQ, etc.).

When you reduce the volume digitally, you do loose resolution. You can't hear the loss of quality because the resulting quantization noise remains low, where you can't normally hear it. But you might hear it if you re-amplify, similar to how you might hear noise if you reduce the analog volume and then re-amplify. It's not something I worry about but if you want "bit perfect" digital audio, of course you can't adjust the volume digitally.


A power amp can have a volume/gain control. Some pro power amps have DSP EQ so you can use them for bi-amplification or tri-amplification, etc., without an external active crossover. An integrated amp (or preamp) has multiple-selectable inputs, and usually tone controls and other "switches & knobs".

It's possible, then, what I'm hearing when I turn down the BT and turn up the amp is just a bad DAC maybe exacerbating distortion via aliasing or something. I'll get a proper DAC and see if that strengthens my options.

If you don't have a way to actually measure the volumes, then just set the volume on the BT receiver to approximately -10 dB below its normal or nominal volume setting. And adjust the volume on the power amp to a comfortable listening level, and leave it there.

Use the volume control on the BT receiver for on-the-fly adjustments.

I do have a way to measure — a very nice SPL meter with switchable modes — but what I've gathered from Amir is that this is not an appropriate way to measure. I'm not opposed to investing in a reference mic, so maybe I'll try that to find a reference level.

Thanks, everyone for your replies!

 

[ADU]

There is no reason that you have to set your gear to reference levels. Alot of lower end gear can't really handle that much volume without clipping or distorting.

The 85 dB unity level that I initially mentioned would be for a -20 dBFS pink noise signal btw, rather than the -30 dBFS test signals that alot of home receivers normally use for their internal setup. I have re-edited my previous post to try to explain this a bit better.

If your receiver includes either one of the above pink noise test signals, then I'd just follow its directions on how to correctly measure and set them up. You can always use a lower target SPL measurement though, if you feel that the above reference levels are too loud for your gear or ears.

 

[Deleted member 46664]

I understand that a power amp is not *technically* a power amp if it has volume control (at which point it's an "integrated amp"), but where is that volume control in the chain, exactly?

The volume control (or gain control) is generally the first thing in the chain inside a power amplifier. That way everything between it and the speakers can run "flat out" doing it's best job.

Strategically the best bet is to max out your BT volume on your source device then set the power amp gain to the highest acceptable level in your listening area, that does not produce distortion. (i.e. just short of noise complaints) Then, from there, control the actual listening level from your BT sending device, where it's nice and convenient.

 

[ADU]

I think I asked this question in another similar ASR topic. But don't think there was ever a reply. But maybe DB or someone else will have a better idea on this....

I'm not sure if this would apply to the OP's particular situation, but I'm sort of curious to know which volume or gain setting might have the most potential for introducing or increasing the noise in a home system if it's lowered... the volume on the preamp, or the gain/volume on the power amp (or gain/trim control on an integrated amp)?

I agree btw that the digital volume control on the source device should generally be left at maximum.

 

[BJL]

I think I asked this question in another similar ASR topic. But don't think there was ever a reply. But maybe DB or someone else will have a better idea on this....

I'm not sure if this would apply to the OP's particular situation, but I'm sort of curious to know which volume or gain setting might have the most potential for introducing or increasing the noise in a home system if it's lowered... the volume on the preamp, or the gain/volume on the power amp (or gain/trim control on an integrated amp)?

I agree btw that the digital volume control on the source device should generally be left at maximum.

Just in my experience, this is dependent on the particular equipment. For example, if a power amplifier uses a cheap potentiometer for attenuation, I would definitely take it out of the circuit by setting the volume to max. The problem is if both the pre-amp and power amp use potentiometers. If so, I'd figure out which one works better (by ear), and use that one to control levels. I've had much much better results with ladder type stepped analog attenuators or 32 bit digital attenuators, and if that is an option with either the amplifier or pre-amp, then that is what I would use to control the volume. Actually, I haven't owned any equipment with a potentiometer for decades. I dislike them.

My opinion, informed by listening, is that the amplifier section should be set to zero attenuation, with the levels controlled by the pre-amp, this assuming that the pre-amp has a good quality digital or analog stepped attenuator. I agree with ADU and others that digital sources should generally be left at 100%; exception, if DSP is used at by a source and it becomes necessary to back off -5dB or so to prevent clipping. Phono pre-amps should be set for the maximum gain possible without clipping, with no attenuation at the output (if that is offered).

 

[RayDunzl]

When you reduce the volume digitally, you do loose resolution.

Technically, yes.

Practically? I'm not so sure.

Visual representation of -60db (that's a lot of attenuation)...

Anyway, I quit worrying about digital attenuation, for sure when the DAC uses 24 or 32 bits to do it.

 

[Deleted member 46664]

I think I asked this question in another similar ASR topic. But don't think there was ever a reply. But maybe DB or someone else will have a better idea on this....

I'm not sure if this would apply to the OP's particular situation, but I'm sort of curious to know which volume or gain setting might have the most potential for introducing or increasing the noise in a home system if it's lowered... the volume on the preamp, or the gain/volume on the power amp (or gain/trim control on an integrated amp)?

In a properly functioning circuit noise is a more or less fixed quality, That is, since volume/gain controls are usually very early in the device's internal paths, most of the noise creating components will come after it.

It is possible that reducing the volume at any point prior to a noisy circuit could cause an apparent increase in noise because you are effectively reducing the dynamic range of everything that comes after it. I say "apparent" because it is most likely the noise level is the same but appears greater due to the reduced level of the audio signal.

I agree btw that the digital volume control on the source device should generally be left at maximum.

This isn't universally true, either. Take the example of a Windows sound card that pumps out 2.5 volts into the input of an integrated amp that is designed for 1 volt ... the only way to avoid over-driving in that case is to reduce the volume controls in Windows, to lower the sound card's output level.

There is a lot of misunderstanding about volume and levels in digital audio. You will hear people talk about losing bits and other horror stories, none of which are actually the case.

A digital sample is not on or off. It is a binary number representing the level (voltage) of the audio waveform at a specific moment in time. It's range varies with the size of the sample ... and 8 bit sample can represent 256 levels, a 16 bit sample ups that 65,536, 24bits gives us 16,777,216 levels, 32bits ends up with 4,294,967,296 possible levels. So by the time we are playing with 24 bits for a line level 1v output we are talking about each step representing 0.05 microvolts per step. Now that is some pretty fine control!

Except for the most disastrous implementations digital volume controls usually add or subtract some value to or from each sample. For example on a 24 bit sample a 1db change would represent adding or subtracting 1.12% from the sample's binary value. That is not going to strip away bits or change the sample size. The sample is still 24bits, it will simply change the numerical value it carries.

So, don't be afraid to use digital volume controls ... often it's possible to make smaller and more refined adjustments than with analog controls, with very little effect on sound quality.

 

[ADU]

Thank you for the replies, BJL and DB.

Re the digital volume controls on the source... I think the only time I'd generally lower or change this on my source would be if I was using a pair of headphones plugged directly into the source's analog audio output. Or to make some small adjustments in the left or right channels to compensate for imbalances in the headphones, when using them with a separate amp or DAC that lacks a stereo balance control.

The only DSP I currently use in my setup is Equalizer APO, to add some EQ to my headphones' frequency response. And this would usually also include a digital preamp filter or gain control, with a negative gain (generally somewhere around -3 to -5 dB) to ensure that the digital audio content is not clipped at any frequency due to positive adjustments in the frequency response via the other EQ filters... even when the source's basic volume setting is maxed out.

My source at the moment is a laptop btw.

 

[ADU]

This isn't universally true, either. Take the example of a Windows sound card that pumps out 2.5 volts into the input of an integrated amp that is designed for 1 volt ... the only way to avoid over-driving in that case is to reduce the volume controls in Windows, to lower the sound card's output level.

There is a lot of misunderstanding about volume and levels in digital audio. You will hear people talk about losing bits and other horror stories, none of which are actually the case.

A digital sample is not on or off. It is a binary number representing the level (voltage) of the audio waveform at a specific moment in time. It's range varies with the size of the sample ... and 8 bit sample can represent 256 levels, a 16 bit sample ups that 65,536, 24bits gives us 16,777,216 levels, 32bits ends up with 4,294,967,296 possible levels. So by the time we are playing with 24 bits for a line level 1v output we are talking about each step representing 0.05 microvolts per step. Now that is some pretty fine control!

Except for the most disastrous implementations digital volume controls usually add or subtract some value to or from each sample. For example on a 24 bit sample a 1db change would represent adding or subtracting 1.12% from the sample's binary value. That is not going to strip away bits or change the sample size. The sample is still 24bits, it will simply change the numerical value it carries.

So, don't be afraid to use digital volume controls ... often it's possible to make smaller and more refined adjustments than with analog controls, with very little effect on sound quality.

I think any reduction in the source's digital volume control can potentially increase the noise levels, and also cause resampling errors. And this would probably apply not only to the basic volume control on the source, but also to any other digital filters (like the EQ filters and preamp control mentioned in my last post) that can also change or manipulate the audio content's amplitude in the digital domain.

Increasing the bit depth on my laptop's audio device (HDMI in this case) from 16 to 24 bits reduced the resampling/quantization errors in my setup to where they seem inaudible though. So I think, as you say above DB, that as long as your audio device is set to 24 bits or more, the resampling due to level changes on the source shouldn't be that noticeable. It might start to become an issue with lower bit depths though (ie 16-bits or less). And for the best resampling, I suspect most audiophiles will prefer bit depths a bit higher than 24-bits.

I think noise could still be a potential issue though if you reduce the digital volume on the source significantly, even if you are using a higher bit depth on your audio device.

 

[Deleted member 46664]

I think any reduction in the digital volume control can potentially increase both the noise levels, and also resampling errors.

I think we're going to disagree on that one.
I manipulate the digital controls in my HTPC all the time... only difference I ever hear is the music gets louder or quieter.

 

[ADU]

I think we're going to disagree on that one.

Perhaps, Douglas.

The gear that's involved probably makes some difference as well, in terms of what you can or cannot get away with for gain staging and attenuation of the signal.

All of my gear is very inexpensive. And probably more prone to producing noise and/or distortion in some cases than the type of gear you or others here may be using. So it probably makes sense for me to pay a little more attention to how and where I could potentially be adding or injectiing any other unnecessary noise, distortion or other degradation of the signal into the system than for other folks with better gear.

Since I'm using mostly cheaper gear, there are some areas in my setup where I have more control over this than others. And one of those areas where I seem to have a pretty good degree of control is at the sound source, where I can easily keep any degradation to a bare minimum by simply ensuring that both the output levels and also the bit depth (which maxes out at only 24-bits in my setup) are as high as possible, without introducing any clipping into the audio content.

I have somewhat less control over what's happening to the signal after that point. And with the exception of the volume control on my headphone amp, most of that control is simply in the form of the components that I've chosen, and how they are connected together.

Most of the attenuation in my setup actually occurs at the headphones themselves, since they are higher in impedance and lower in sensitivity than most other consumer gear. And this is in part what enables me to keep the signal running as hot (and clean imo) as possible, with as little loss as possible, in all of my other gear. This effectively takes the pressure off of some of that other gear to perform at a higher level.

Lowering the volume of your sound source with a passive digital control like the one on my laptop is, imo, inherently lossy though, because it necessitates resampling of the audio data. And it reduces dynamic range, signal-to-noise ratio, and potentially also the signal level at the analog output stage. So the chances are that you'll usually get better results and better signal integrity if you can do your attenuation somewhere else than at that particular gain stage.

I can understand that there are issues of convenience involved in this as well though. And that it may be just as convenient, if not more so, to do your attenuation at your sound source, if, for example, most of your gear is of better quality than mine. And if you do not have as ready or easy access to the gain or attenuation controls at the other stages in your audio setup as I do.

 

[ADU]

If you don't have a way to actually measure the volumes, then just set the volume on the BT receiver to approximately -10 dB below its normal or nominal volume setting. And adjust the volume on the power amp to a comfortable listening level, and leave it there.

Use the volume control on the BT receiver for on-the-fly adjustments.

The -10 dB level that I mentioned in the above post was just a conservative ballpark figure, btw. And some listeners may prefer a nominal level that is somewhat lower than this. Perhaps even as low as -20 dB below reference in some cases.

So if you feel that you need a bit more headroom in your volume adjustments for some of the content you listen to, then you could set the receiver's master volume a bit lower, perhaps more in the -15 or -20 dB range, below unity. And use that as your reference point for adjusting the trim or volume of your power amp to a more comfortable listening level by ear instead.

 

[ADU]

I think we're going to disagree on that one.
I manipulate the digital controls in my HTPC all the time... only difference I ever hear is the music gets louder or quieter.

Alot of audiophiles also prefer to send their audio content at its native bitrate to an external DAC (where the upsampling and D/A conversion is presumably better), which means that the sound source could be restricted to only 16-bits, or possibly even less in some cases. Which can presumably also make it more vulnerable to noise and resampling errors, if the volume is reduced at that stage.