What are the best settings for digital sources and DACS with volume controls?

[GGroch]

Sorry for the lengthy post but I have a long question.

I listen to several sources in which there are multiple components with volume control in the chain. My question, does where I control the volume impact measured or audible fidelity?

For example:

I Use the Optical Digital Out on my Chromecast Audio. The Google Home app provides volume (and tone) controls...so I can control volume with it, fully in the digital domain.

The Chromecast feeds DACS (a Topping D50 or Loxjie D20) with volume controls for the preamp outs. It is not clear to me whether the adjustment happens in the digital realm, or in the analog preamp stage...or if that matters.

The DAC may feed a headphone amp like the JDS Atom with volume control; variable pre-amp outs

The headphone amp feeds an integrated amp with its own control.

Is there an issue with Digital volume control. In the early days of Digital I was taught that with PCM audio...the higher the volume the greater the accuracy (uses more bits). Plus...if my Chromecast lowers the volume before the DAC....then the DAC is not seeing a bit perfect copy of the original stream right?

Then all of the other analog options. I know that if the amp has pots for control it may not be linear at very low levels. And of course over-driving is bad too.

Any rules of thumb on how I can maximize fidelity?

 

[amirm]

I plan to study this topic at some point using measurements. I suspect the answers will be situation specific.

 

[LF78]

I was assuming that keeping the digital signal @0dB until DAC, and controlling volume digitally at DAC would be the best I can do: do I miss something?

 

[Krunok]

Plus...if my Chromecast lowers the volume before the DAC....then the DAC is not seeing a bit perfect copy of the original stream right?

If your preamp lowers the volume your amp is not seeing perfect analog copy of the original signal either.

 

[andreasmaaan]

I was assuming that keeping the digital signal @0dB until DAC, and controlling volume digitally at DAC would be the best I can do: do I miss something?

This is likely to be the optimal way to do it, yeh. The D50 is a 32 bit DAC so it should be capable of a lot of attenuation before even approaching any audible loss of fidelity.

 

[LF78]

I'm thinking: what about intersample clipping? Until you stay digital no issues, but could the DAC suffer from it in some cases? In my opinion the digital volume control, being digital, should attenuate the digital signal before D/A conversion, hence preventing any issue, am I correct?

 

[andreasmaaan]

I'm thinking: what about intersample clipping? Until you stay digital no issues, but could the DAC suffer from it in some cases? In my opinion the digital volume control, being digital, should attenuate the digital signal before D/A conversion, hence preventing any issue, am I correct?

The maximum level will never be higher than 0dBfs, and unless you have a strange gain structure and very low sensitivity speakers you'll never even get that high. All the digital volume control does is attenuation. If there's intersample clipping in the recording, this will be present in the output at 0dBfs, but the same would be true without digital volume control (NB: I'm not a technical expert in digital processing).

 

[helloworld]

I plan to study this topic at some point using measurements. I suspect the answers will be situation specific.

Please include the volume control in windows and volumio!

 

[bennetng]

The maximum level will never be higher than 0dBfs, and unless you have a strange gain structure and very low sensitivity speakers you'll never even get that high. All the digital volume control does is attenuation. If there's intersample clipping in the recording, this will be present in the output at 0dBfs, but the same would be true without digital volume control (NB: I'm not a technical expert in digital processing).

If the signal chain and/or audio format involve floating point, the maximum level can exceed 0dBFS. That's why people should use the internal software volume control and/or ReplayGain to avoid clipping.

Please read every link in the post below.

https://www.audiosciencereview.com/...lity-in-windows-using-wasapi.5272/post-116878

 

[LF78]

people should use the internal software volume control and/or ReplayGain to avoid clipping

Do you think that using the digital volume control of the DAC (if available) is the same?

 

[bennetng]

Do you think that using the digital volume control of the DAC (if available) is the same?

Mathematically, no. In practice (analog output and potential audible difference), it depends on several things.
(1) Does the DAC have internal headroom? (e.g. Benchmark)
(2) Does the signal chain and/or audio format involve floating point data?
These questions are answered in the link I posted earlier. There are some test files and a calculator to try out.

Archimago helped me to test some of his DACs for clipping and volume control, my method can eliminate clipping in all of his DACs.
http://archimago.blogspot.com/2018/...howComment=1514806847604#c6918600883178190233

In short, volume control on a DAC, digital or not, is not always safe to use in the sense of avoid clipping. An hybrid approach (volume/ReplayGain on playback software + analog volume control at or after DAC output) can avoid clipping and retain analog dynamic range, so you get the best from digital and analog technology.

 

[LF78]

Interesting discussion.

Mathematically, no. In practice (analog output and potential audible difference), it depends on several things.
(1) Does the DAC have internal headroom? (e.g. Benchmark)
(2) Does the signal chain and/or audio format involve floating point data?

I would like to understand better these points. Let's say that the digital signal is bit-perfect from source file to DAC, i.e. no attenuation (0 dB). Let's also assume the DAC is a miniDSP, working internally with floating point 32bit precision.

Can we have intersample clipping in this scenario? In my opinion no: miniDSP should digitally attenuate the source signal as per volume setting, and send it to D/A conversion already attenuated. Do you agree?

 

[Deleted member 65]

The maximum level will never be higher than 0dBfs, and unless you have a strange gain structure and very low sensitivity speakers you'll never even get that high. All the digital volume control does is attenuation. If there's intersample clipping in the recording, this will be present in the output at 0dBfs, but the same would be true without digital volume control (NB: I'm not a technical expert in digital processing).

I think it's a bit more complicated then this, see Intersample Peaks and the possibility of overloading the DAC. Some DAC's eg Benchmark DAC3 & RME ADI-2 FS are able to handle Intersample Peaks between +2-3dB.

Archimago has a lengthy discussion here - - > http://archimago.blogspot.com/2018/09/musings-measurements-look-at-dacs.html

Edit: @bennetng beat me to it ... ;-)

 

[bennetng]

Interesting discussion.



I would like to understand better these points. Let's say that the digital signal is bit-perfect from source file to DAC, i.e. no attenuation (0 dB). Let's also assume the DAC is a miniDSP, working internally with floating point 32bit precision.

Can we have intersample clipping in this scenario? In my opinion no: miniDSP should digitally attenuate the source signal as per volume setting, and send it to D/A conversion already attenuated. Do you agree?

If the digital source is floating point based and it has >1.0 values, then clipping can happen at the software output (ASIO/WASAPI/whatever) before entering the DAC. What the DAC received is some clipped data. Then the DAC is just using floating point to process some clipped data, and distortion still happens.

Source (float) --> software output and DAC input (fixed and clipped) --> DAC internal processing (float) --> analog output.

My test files can demonstrate the effect.

 

[LF78]

If the digital source is floating point based and it has >1.0 values, then clipping can happen at the software output (ASIO/WASAPI/whatever) before entering the DAC

Very interesting, I didn't know the possibility for clipping in the digital realm before D/A. Does this happen only for floating point files/sources, and not for 16/24 bit material?

 

[bennetng]

Very interesting, I didn't know the possibility for clipping in the digital realm before D/A. Does this happen only for floating point files/sources, and not for 16/24 bit material?

These two post addressed this question. You are, and will be affected even if you don't use float formats.
https://www.audiosciencereview.com/...lity-in-windows-using-wasapi.5272/post-117218
https://www.audiosciencereview.com/...lity-in-windows-using-wasapi.5272/post-117704

 

[Krunok]

Please include the volume control in windows and volumio!

With mixer type set to "hardware" Volumio will do volume control via USB chip with USB DACs if such control is supported. With XU208 based USB DACs it is.

 

[bennetng]

rchimago has a lengthy discussion here - - > http://archimago.blogspot.com/2018/09/musings-measurements-look-at-dacs.html

Edit: @bennetng beat me to it ... ;-)

Actually I left a comment in that blog entry to vent my frustration.
http://archimago.blogspot.com/2018/...howComment=1538291070249#c3235820390853013432

 

[LF78]

These two post addressed this question. You are, and will be affected even if you don't use float formats

Thanks for the links, I think I'm starting to understand this topic much better. So:

- MP3 (and compressed files in general) may have peaks > 0dBfs.
- Uncompressed FLAC should be different, but any FLAC could be the result of MP3 decompression or any other mastering process, so it's the same.

I still don't understand the concept of "headroom" at DAC level in the digital domain: unless you set the digital volume to maximum (0 dB), any DAC with digital volume control should have plenty of headroom at normal volume levels.

For example I usually listen music with my miniDSP 2x4HD with master volume at -30/-35 dB: does it mean I have 30/35 dB of headroom (from a digital perspective - let's ignore the amplifier later in the chain)?

 

[bennetng]

- Uncompressed FLAC should be different, but any FLAC could be the result of MP3 decompression or any other mastering process, so it's the same.

FLAC does not support float, if the input contains > 1.0 float data, those data will be clipped upon conversion. The float data must be normalized to within 1.0 before converting to FLAC. There are other lossless formats with float support, for example Wavpack:
http://www.wavpack.com/

For example I usually listen music with my miniDSP 2x4HD with master volume at -30/-35 dB: does it mean I have 30/35 dB of headroom (from a digital perspective - let's ignore the amplifier later in the chain)?

Right, but reducing too much volume in digital domain can harm dynamic range. Let's say the DAC reviews in this site usually show something like 110 to 120dB dynamic range, if you reduce a 110dB device by 30dB, it will become 80dB. Some audiophiles believe it is an issue, so some people may just decrease for example 3 to 6dB digitally to avoid clipping, and use an analog volume control to do the rest. For example, ESS themselves still agree with some of the advantages of analog volume control.

http://www.esstech.com/files/3014/4095/4308/digital-vs-analog-volume-control.pdf

Of course, a high quality analog volume control is required, otherwise it will only make things worse.

Also, an analog volume control can ensure safety. Sometimes software/firmware bugs or user mistakes can reset the digital volume to 0dB, which is quite dangerous or at least annoying. A physical analog volume control like a fader or knob does not have these issues.