[Solude]

Bit Perfect stops below 40/100.

 

[valer_valer]

Why ?
It depends on the resolution of the input signal.
The WM8741's digital volume work on 24bit domain.

 

[Solude]

16bit material needs 19bits to allow for noise floor leaving 5bits which is 30dB and 40/100 is -30dB.
 
If you are playing high rez material... you're SOL.

 

[valer_valer]

OK, I found it:
Steps 40 - 100 are 0.5dB each ( 0dB to -30dB Attenuation)
Seps 10 - 39 are 1.0dB each ( -31dB to -60dB Attenuation)
Steps 0 - 9 are 2.0dB each ( -62dB to -80dB Attenuation)
Now, it's perfectly fits ...

 

[Wapiti]

The PWD volume control is 32 bit.  Thus, even with a true 24 bit source there is ~ 50dB of attenuation available before the potential loss of resolution.  As a practical matter, the noise floor of any system is, at maximum, 20 bit (~-120dB, which would be spectacular) so there is another four or five bits available before any loss can be discerned - a grand total of 75dB or more attenuation without loss.
 
Paul is well aware that many would like a selectable lower output range for running the unit directly into an amp.  He is exceedingly uncomfortable in adding an analog circuit as this only tosses a portion of a wonderfully clean single to ground to achieve attenuation by adding circuitry to the signal path (this is how analog volume controls work) as well as the concurrent degredation of sound.  I expect PSA will incorporate a digital gain selector at some point.
 
I am certain PSA will never add an analog volume control as they necessarily compromise the sound.  A well implemented digital volume control sounds better as there is nothing added to the signal path and no resolution is thrown away.  The prejudice many of us share against digital attenuation is well founded - early and less sophisticated digital controls immediately through away meaningful bits.  For example, a -12dB change accomplished by going from 16-bit to 14 bit is easily heard.

 

[jcx]

Quote:

For example, a -12dB change accomplished by going from 16-bit to 14 bit is easily heard

with a good dither? - not likely at all for any reasonable music listening scenerio - can you even ABX the -60 dB added Sousa march? http://www.libinst.com/Audio%20DiffMaker.htm > http://libinst.com/diffmaker_example_files.htm

 

[preproman]

Just get this and put the volume to 100
 
http://www.goldpt.com/sa1x.html

 

[valer_valer]

Quote:

The PWD volume control is 32 bit.

 
 
I think You are wrong !
 
The PWD volume control based on WM8741's Volume implementation.
 
And according to Wolfson's answer it's 24bit.
 
 
Quote:

 

Originally Posted by http://hifiduino.blogspot.com

I also got a response from Wolfson about how the volume control is implemented: Basically input less than 24 bit is padded to 24 bit, input that is 25 to 32 bit is reduced to to 24 bit with dither and the volume control is applied at the next stage...

 

 

[jcx]

so what? - doing a few multiplies isn't like the layers on layers of processing in a DAW - even 24 bits will keep round-off quantization error way below the DAC's own output thermal noise limit

 

[Wapiti]

Quote:

I think You are wrong !
 
The PWD volume control based on WM8741's Volume implementation.
 
And according to Wolfson's answer it's 24bit.

I could well be wrong.  Unfortunately the link you provided does not definitely answer the question and Wolfson's datasheet does not answer the question.
Quote:

Quote:
with a good dither? - not likely at all for any reasonable music listening scenerio -

Yes.  44.1/14 and 44.1/16 is readily distinguishable, at least on my own recordings of chamber orchestra, full orchestra and ensemble work. Timbre is impacted, as well as ambiance retrieval and decay envelopes.  I have not tried on pop/rock recordings and frankly do not care as I do not listen to this genre.  Your experience may well differ.

 

[jcx]

so you make, listen to recordings in an anechoic chamber? or don't use dither?
 
can't see how you're not playing games here, using false analogy - no ones talking about recording/mastering - we are talking about volume control - ie the at least 16 bit mastered, dithered, music was already too loud in the system and needs turning down, the gain structure isn't too screwed up to start with, using typical "audiophile" open back headphones in a domestic room - NC20 is a stretch goal for people building studio monitoring rooms
 
under typical domestic "audiophile" listening conditions the "lost bits" of a 12 dB digital turndown is not "clearly audible" - nowdays in most softwware it will be redithered if stuck with a 16 bit DAC
 
truncate, turn the gain up - to where the normal volume passages will blow out of your seat and yes, you can hear some things going on at the 16 bit lsb level -  not relevant to this digital volume control discussion

 

[valer_valer]

Quote:

I could well be wrong.  Unfortunately the link you provided does not definitely answer the question and Wolfson's datasheet does not answer the question.

 
Here:
 
 

Originally Posted by Paul McGowan
 
The PWD will accept 32 bit data but the DAC is only 24 bit. If we misspoke somewhere....... 

 
Does this answer Your question ?

 

[Wapiti]

Quote:

 
can't see how you're not playing games here, using false analogy - no ones talking about recording/mastering . . .
 
truncate, turn the gain up - to where the normal volume passages will blow out of your seat and yes, you can hear some things going on at the 16 bit lsb level -  not relevant to this digital volume control discussion

Sorry, no game playing and no false analogy; I am specifically describing my experience as a recording engineer.  Your experience obviously differs.  I never argue with another's description of his/her subjective experience.  If you do not hear a difference, you do not hear a difference.
 
I agree, listening to the noise floor alone at extreme gain is not meaningful.  I, for one, would never bother with the experiment.

 

[Wapiti]

Quote:

Here:
 
Does this answer Your question ?

Sure.  Paul is reliable source. 
 
24-bit obviously substantially reduces the available headroom for attenuation v. 32-bit

 

[preproman]

Quote:

Sure.  Paul is reliable source. 
 
24-bit obviously substantially reduces the available headroom for attenuation v. 32-bit

 
 
Trying to decide on one of these to fix that..  Which one you think?
 
http://spl.info/produkte/monitor-controller/volume2/beschreibung.html
 
or
 
http://www.goldpt.com/sa1x.html