IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
Maximize volume/gain: another case against loudness war?
user1
post Apr 9 2014, 03:19
Post #1





Group: Members
Posts: 39
Joined: 18-January 14
Member No.: 114096



Loud CDs and music files can force listeners to reduce their volume control.

A comment by DIYer linux-works at EEVBlog notes:
QUOTE
ideally, your dac should output 100% signal (no atten at all) and the rest of the chain should allow a fullscale vol control knob (at max) to JUST play about as loud as you can take it. ... if you're not using most of the vol control range, then your gain structure is not optimal or its trying to support too many load Z levels.

Not sure how applicable this is to the case against loudness'd music.

This post has been edited by user1: Apr 9 2014, 03:20
Go to the top of the page
+Quote Post
saratoga
post Apr 9 2014, 04:18
Post #2





Group: Members
Posts: 5161
Joined: 2-September 02
Member No.: 3264



Hes saying that you should be using the full range of your DAC, and then adjusting gain in analog after the DAC. In practice this is how virtually all sound cards and portable devices work, so its not really a big deal.

I don't think this relates at all to how things are mastered.
Go to the top of the page
+Quote Post
user1
post Apr 9 2014, 05:43
Post #3





Group: Members
Posts: 39
Joined: 18-January 14
Member No.: 114096



QUOTE (saratoga @ Apr 9 2014, 05:18) *
Hes saying that you should be using the full range of your DAC, and then adjusting gain in analog after the DAC. In practice this is how virtually all sound cards and portable devices work, so its not really a big deal.

I don't think this relates at all to how things are mastered.
That whole thread on EEV is about using the full range of an analog volume control. If the program material is loud, then this discourages listeners from using the volume control over its full range -- one ends up selecting below 0dbFS, which is attenuation (slightly bit compromised, IIRC).
If one turns down the SC gain (digital volume), then this -- according to linux-works -- is a compromise. Ditto argument for not being able to use the full range of an analog pot (that was the last part in the quote above).
Not sure how much actual diff these tweaks make in final SQ, tho'.

This post has been edited by user1: Apr 9 2014, 05:48
Go to the top of the page
+Quote Post
saratoga
post Apr 9 2014, 07:13
Post #4





Group: Members
Posts: 5161
Joined: 2-September 02
Member No.: 3264



QUOTE (user1 @ Apr 8 2014, 23:43) *
That whole thread on EEV is about using the full range of an analog volume control. If the program material is loud, then this discourages listeners from using the volume control over its full range -- one ends up selecting below 0dbFS, which is attenuation (slightly bit compromised, IIRC).


I don't understand what you are trying to say so I can't really reply to this. 0dBFS is digital gain, whereas volume control is generally analog. Which do you mean?
Go to the top of the page
+Quote Post
user1
post Apr 9 2014, 08:21
Post #5





Group: Members
Posts: 39
Joined: 18-January 14
Member No.: 114096



QUOTE (saratoga @ Apr 9 2014, 08:13) *
QUOTE (user1 @ Apr 8 2014, 23:43) *
That whole thread on EEV is about using the full range of an analog volume control. If the program material is loud, then this discourages listeners from using the volume control over its full range -- one ends up selecting below 0dbFS, which is attenuation (slightly bit compromised, IIRC).


I don't understand what you are trying to say so I can't really reply to this. 0dBFS is digital gain, whereas volume control is generally analog. Which do you mean?
Many (if not most) DAPs and sources, like CDPs, have non-adjustable line output level.

For the digital volume noted in the EEV thread, the digital volume discussed is controlled in the CS4398 DAC. The Colorfly CK4 DAP that uses that DAC has a line out that has the same volume control as the headphone out. Most users of that DAP (e.g., on the Head-Fi forum, etc.) suggest that max volume (0 dbFS) for that DAP is optimum line-out level.

The line out of that DAP is fed to, e.g., a headphone amp that has the std. rotary volume knob. As l-w suggests (and as per most of my own admittedly subjective evals.), subjective SQ is best if the entire range of the rotary pot is used. So max (5 o'clock) should be as high as you can stand. For DIY amp building, a personal hobby, I tweak my amps gain for this. In fact, better preamps also work similarly -- unlike common consumer-grade equip where things get very loud above, say, 12 o'clock.

With loudness'd music, one cannot utilize more of the volume pot's range, w/o output getting uncomfortably loud.

The questionable and controversial issue of the EEV thread is whether the UPPER "half" of the analog pot is more "linear" because (a) there is less resistance; and/or (b) the resistance in the upper half of the pot is better impedance-matched for the entire volume topology. If this latter part is correct (my belief as well), then one way to utilize the upper half of the pot is to lower the source level.

If the source is non-loudess'd, you're okay. If not, you can: (a) turn down the pot [possibly removing it from impedance matching]; or (b) use (or install) a digital gain (atten.) control for the source. But ,as noted above, digital atten., is a slight compromise (as noted in the EEV thread). I think digital volume was at most a 1-bit loss in rez.

This post has been edited by user1: Apr 9 2014, 08:27
Go to the top of the page
+Quote Post
saratoga
post Apr 9 2014, 19:53
Post #6





Group: Members
Posts: 5161
Joined: 2-September 02
Member No.: 3264



The vast majority of line out ports have analog volume controls, so if you need a lower than normal level its very easy to find a part which can accommodate you.
Go to the top of the page
+Quote Post
drewfx
post Apr 9 2014, 20:38
Post #7





Group: Members
Posts: 98
Joined: 17-October 09
Member No.: 74078



QUOTE (user1 @ Apr 9 2014, 03:21) *
or (b) use (or install) a digital gain (atten.) control for the source. But ,as noted above, digital atten., is a slight compromise (as noted in the EEV thread). I think digital volume was at most a 1-bit loss in rez.


I don't think turning down the digital volume is a compromise in this case, as all you're likely doing is reducing the average (dBFS) level to near where it was originally (before the peaks were cut off).
Go to the top of the page
+Quote Post
DVDdoug
post Apr 9 2014, 20:54
Post #8





Group: Members
Posts: 2713
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE
The questionable and controversial issue of the EEV thread is whether the UPPER "half" of the analog pot is more "linear" because (a) there is less resistance; and/or (b) the resistance in the upper half of the pot is better impedance-matched for the entire volume topology. If this latter part is correct (my belief as well), then one way to utilize the upper half of the pot is to lower the source level.
The only issue I've run across is the matching of dual (stereo) pots. I've had pots where the left-right balance was off near the center and I've had dual pots with more "play" in the rear pot, so that a very-small adjustment only affects one channel. As you approach maximum volume (minimum attenuation) the "linearity" of the pot makes little difference. (Audio volume-pots aren't actually linear.)

Depending on the design, the actual resistance/impedance generally isn't an issue... In most circuits you could change the value of the pot by a factor of 2 (or more) and it would make zero difference.

I think the solution is to use digital pots, which can be controlled with a single analog pot or a rotary encoder.

And, there's nothing wrong with digital volume control! ...Let's say you reduce the volume of a 16-bit file to -80dB. Sure, the resolution is gone and there is LOTS of distortion (relative quantization noise) but the quantization noise is still more than 90 dB down and you can't hear it. In fact, you probably can't hear the signal at -80dB either! Now yes, if you re-boost the signal (with digital or analog gain) it's going to sound terrible. But, the sound quality will also be destroyed if you reduce the analog signal by 80dB and then re-amplify it.

A similar thing happens with photos... If you shrink a photo by reducing the number of pixels, you loose resolution. But, you don't notice it unless you blow-up the reduced picture. If you shrink and then enlarge a picture digitally it gets "pixilated". If you shrink then enlarge a picture with "analog" film, the picture gets fuzzy or "grainy".

This post has been edited by DVDdoug: Apr 9 2014, 20:59
Go to the top of the page
+Quote Post
saratoga
post Apr 9 2014, 23:26
Post #9





Group: Members
Posts: 5161
Joined: 2-September 02
Member No.: 3264



Even ignoring that, there is really no reason to use digital volume control. Almost modern DAC will have multiple mixers post DAC, each of which can apply a variable gain. If this is DIY, just set the line output's amps to match the gain your custom device needs, or better yet, just buy an integrated device like a DAP where this is taken care of for you.
Go to the top of the page
+Quote Post
user1
post Apr 10 2014, 04:59
Post #10





Group: Members
Posts: 39
Joined: 18-January 14
Member No.: 114096



Please note: there is distinction between volume level (topological circuit, part of which may be a VR) and volume pot (single device VR).

If/when I find some time, I may do this experiment:
Level match with 'scope or DMM.
Low source level / High analog (preamp) level
High source level / Low analog (preamp) level [as with loudness'd music]
Run test signals; measure distortion at outputs. Follow up with listening tests.

Today, as I was on my commute, and had my playlist running in my DAP+headphone amp, I noticed how many times I found myself reaching for the rotary pot on the headphone amp. (Playlist consists of tracks from myriad CD rips, some were loudness'd ) . The non-loudness'd tracks sounded better, but I wonder if a smidgen of that was because the high-setting on the pot was operating in a more "linear" range???
Volume (analog pots) can be crazy and complex. As noted earlier, their innate and variable resistance can interact complexly with the impedance of preceding and following stages, affecting load Z. And as the signal encounters more resistance (goes thru more resistive material), Johnson-Nyquist noise issues may also be at work.

So maybe a high-end volume pot is the answer (!!!) ...


This post has been edited by user1: Apr 10 2014, 05:02
Go to the top of the page
+Quote Post
saratoga
post Apr 10 2014, 06:04
Post #11





Group: Members
Posts: 5161
Joined: 2-September 02
Member No.: 3264



QUOTE (user1 @ Apr 9 2014, 22:59) *
Today, as I was on my commute, and had my playlist running in my DAP+headphone amp, I noticed how many times I found myself reaching for the rotary pot on the headphone amp. (Playlist consists of tracks from myriad CD rips, some were loudness'd ) . The non-loudness'd tracks sounded better, but I wonder if a smidgen of that was because the high-setting on the pot was operating in a more "linear" range???


If your headphone amp actually puts out worse audio at slightly lower volume, its not a great design. There will inevitably be some loss of SNR if you cut out 25 or 30 dB of volume, but at reasonable differences, the output should be identical aside from amplitude in a competently designed part.

QUOTE (user1 @ Apr 9 2014, 22:59) *
Volume (analog pots) can be crazy and complex. As noted earlier, their innate and variable resistance can interact complexly with the impedance of preceding and following stages, affecting load Z.


Only if the device is terrible.
Go to the top of the page
+Quote Post
Martel
post Apr 10 2014, 08:00
Post #12





Group: Members
Posts: 568
Joined: 31-May 04
From: Czech Rep.
Member No.: 14430



If I recall correctly, it should be perfectly fine to place a variable voltage divider between a DAC and an amplifier since an amplifier usually has input impedance in range of (tens of) kilo ohms and draws very little current (is a very light load).

Placing it after an amplifier (just before the load) is terrible. smile.gif


--------------------
IE4 Rockbox Clip+ AAC@192; HD 668B/HD 518 Xonar DX FB2k FLAC;
Go to the top of the page
+Quote Post
cliveb
post Apr 10 2014, 08:23
Post #13


WaveRepair developer


Group: Developer
Posts: 855
Joined: 28-July 04
Member No.: 15845



QUOTE (user1 @ Apr 10 2014, 04:59) *
Today, as I was on my commute, and had my playlist running in my DAP+headphone amp, I noticed how many times I found myself reaching for the rotary pot on the headphone amp. (Playlist consists of tracks from myriad CD rips, some were loudness'd ) .

Try using ReplayGain.

QUOTE (user1 @ Apr 10 2014, 04:59) *
The non-loudness'd tracks sounded better, but I wonder if a smidgen of that was because the high-setting on the pot was operating in a more "linear" range???

Far more likely the reason the "non-loudness'd" tracks sounded better is because they had better dynamic range - unlikely to be anything to do with where the volume pot was.
Go to the top of the page
+Quote Post
user1
post Apr 10 2014, 08:33
Post #14





Group: Members
Posts: 39
Joined: 18-January 14
Member No.: 114096



QUOTE (DVDdoug @ Apr 9 2014, 20:54) *
And, there's nothing wrong with digital volume control! ...Let's say you reduce the volume of a 16-bit file to -80dB. Sure, the resolution is gone and there is LOTS of distortion (relative quantization noise) but the quantization noise is still more than 90 dB down and you can't hear it. In fact, you probably can't hear the signal at -80dB either! Now yes, if you re-boost the signal (with digital or analog gain) it's going to sound terrible. But, the sound quality will also be destroyed if you reduce the analog signal by 80dB and then re-amplify it.
I like true digital volume controls, too. By "true", I mean in digital-domain via "scaling"/non-decimating method (as in a HQ DAC, like my DAPs CS4398). As you note, dig. vol. does suffer from bit quant. noise, etc.
However, analog volume pots (even HQ ones like the descrete type shown in my prev. post) come with their own share of problems ... Johnson-Nyquist noise from the resistive material used, complex/varying impedances, parasitics, ground-loop issues, etc., etc.!
For level-setting and variation I currently use a combo of both digital (source) and analog (amp). My DAP's volume, I assume, only utilizes the DAC chips innate digital-volume feature. About 2/3 level seems to sound best for its Line Out level, when used with the outboard amp.

All that said, one thing that puzzles me is where 0dbFS (true 100%) is on the UI display? E.g., on my DAP, volume level is displayed as a segmented horiz. bar scale. Full-scale on the UI display may be more than 100%. I suppose I should get my meter out!
Go to the top of the page
+Quote Post
user1
post Apr 10 2014, 13:05
Post #15





Group: Members
Posts: 39
Joined: 18-January 14
Member No.: 114096



QUOTE (Martel @ Apr 10 2014, 08:00) *
If I recall correctly, it should be perfectly fine to place a variable voltage divider between a DAC and an amplifier since an amplifier usually has input impedance in range of (tens of) kilo ohms and draws very little current (is a very light load).

Placing it after an amplifier (just before the load) is terrible. smile.gif
Yes, and maybe that's why the Colorfly CK-4 DAP I'm referring to above uses digital-domain (in-the-DAC) volume scaling.

I have another DAP, a Teclast T-51, that has fixed LO. ...
DAC --> LPF --> Out jack
For its HO ckt ...
DAC --> LPF --> potentiometer--> Headphone Amp (TDA1308) --> Out jack
The T-51 is about iPhone-sized ... it's my main "music server" ...

Go to the top of the page
+Quote Post
saratoga
post Apr 10 2014, 19:24
Post #16





Group: Members
Posts: 5161
Joined: 2-September 02
Member No.: 3264



QUOTE (user1 @ Apr 10 2014, 02:33) *
QUOTE (DVDdoug @ Apr 9 2014, 20:54) *
And, there's nothing wrong with digital volume control! ...Let's say you reduce the volume of a 16-bit file to -80dB. Sure, the resolution is gone and there is LOTS of distortion (relative quantization noise) but the quantization noise is still more than 90 dB down and you can't hear it. In fact, you probably can't hear the signal at -80dB either! Now yes, if you re-boost the signal (with digital or analog gain) it's going to sound terrible. But, the sound quality will also be destroyed if you reduce the analog signal by 80dB and then re-amplify it.
I like true digital volume controls, too. By "true", I mean in digital-domain via "scaling"/non-decimating method (as in a HQ DAC, like my DAPs CS4398). As you note, dig. vol. does suffer from bit quant. noise, etc.
However, analog volume pots (even HQ ones like the descrete type shown in my prev. post) come with their own share of problems ... Johnson-Nyquist noise from the resistive material used, complex/varying impedances, parasitics, ground-loop issues, etc., etc.!


None of these should be a real problem, and in practice analog control is always preferable. Have you actually observed any of these problems, or are they just theoretical? IMO you're worrying about nothing.

QUOTE (user1 @ Apr 10 2014, 02:33) *
For level-setting and variation I currently use a combo of both digital (source) and analog (amp). My DAP's volume, I assume, only utilizes the DAC chips innate digital-volume feature. About 2/3 level seems to sound best for its Line Out level, when used with the outboard amp.


If its actually digital (which would be a terrible design), you should run it at 100% volume. Anything else is throwing away dynamic range. However, if in this case I'd probably just get a better player. Theres no excuse for such poor design on modern devices. Proper analog volume control is available on parts costing less than $1.

QUOTE (user1 @ Apr 10 2014, 02:33) *
All that said, one thing that puzzles me is where 0dbFS (true 100%) is on the UI display? E.g., on my DAP, volume level is displayed as a segmented horiz. bar scale. Full-scale on the UI display may be more than 100%.


If you really have digital gain control, then 100% is 0dBFS. You only have positive gain with analog control generally.
Go to the top of the page
+Quote Post
KMD
post Apr 18 2014, 15:50
Post #17





Group: Members
Posts: 123
Joined: 21-January 09
From: UK
Member No.: 65825



Surely analogue attenuation is prefferable and digital attenuation needs re-dithering at each adjustment.

Also noteworthy is where gain make up disipline applys to laptops and PC applications as they have their own volume controls in series with the windows volume output control.
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 25th December 2014 - 05:38