Digital Volume Control

Thanks for the video.



I am impressed the maker of Sabre DAC chips (one of the best) so readily acknowledges the limitations of the technology.



Interesting that he notes that high quality resistors can be included on a chip. A DAC chip with a resistor ladder volume control would be very cool.



I am surprised he indicates that a volume control with thermal noise of -132db could be built. My understanding is that the thermal noise threshold is -120dB and is a physical characteristic of discrete components.



stereophilus said: It would seem that the SN ratio suffers progressively with each dB of digital domain attenuation . . .


This is true with a 16 bit data path as the noise is within the 96dB dynamic range of 16 bits. That is, as you turn the volume down there is no “lower” to which the noise can be pushed.



With a 24-bit data path there is there is 144dB theoretical S/N available (120dB or so real world). Thus, with a 16 bit signal, the signal can be decreased by at least 24dB (120dB minus 96dB) before running into the noise floor of the system.



As always, let your ears be your guide. :slight_smile:


Really good video and for the most part something I could understand to varying degrees. However as the volume control is the point here . It was the jitter that really I found fascinating in it’s explanation. To whom found this gem thanks.



Al. D

@johnjen that video was fascinating. Very useful to the topic here. I am left feeling somewhat pessimistic about keeping my DAC volume at 50 despite the internal chip architecture being 32-bit. It would seem that the SN ratio suffers progressively with each dB of digital domain attenuation... Gonna have to review those figures a few times! :-B


The wolfson chip does not have a 32 bit architecture. It merely can take a 32 bit word length on input but immediately truncated to 24 bit. Therefore the second you attenuate a 24 bit file you are losing something.

There are two questions that have not been answered:

Does the wolfson apply dither when the volume is applied to a bit reduced signal?

Why does the pwd show highest s/n ratio when volume is at 67 as opposed to 100? This makes no sense as 100 is supposed to be with no volume applied. These measurements came from two different sources...the hifi news review and a member here.

Seems like a better compromise for jriver users is to use it's internal 64 bit volume.


The Wolfson dithers the truncation of 32-bit to 24-bit, but I do not believe it dithers volume attenuation.



I do not think dither matters for volume attenuation. First, the dither of the original signal is maintained as the bits are shifted toward the LSB. There is no reason to re-dither.



Second, as the LSBs are truncated, they fall off into the noise floor. Thus, the signal is inherently dithered by the random thermal noise floor.



crabby said: Why does the pwd show highest s/n ratio when volume is at 67 as opposed to 100?


My guess is this is a function of the gain structure of the PWD, rather than a function of the DAC chip’s volume attenuation.

The wolfson chip does not have a 32 bit architecture. It merely can take a 32 bit word length on input but immediately truncated to 24 bit. Therefore the second you attenuate a 24 bit file you are losing something.


Thanks for clarifying that.

@elk yes, listening is the only way to compare definitively.

I also acknowledge the posts about a gain attenuation switch.

So, moving forward, I see 3 choices for Paul and PS Audio:
1) Pursue a future DAC with greater internal bit architecture, eg 64-bit when it is available.
2) implement a high quality analog volume control.
3) implement a gain attenuation switch.

I suspect 1 and 2 are quite expensive options. Option 1 is probably the way of the future. Option 2 probably gives the best real world performance currently available. Option 3 is probably the cheapest and easiest to implement, despite possible signal quality degradation.

Seeing as its Christmas, I'll have all 3 please, in 1 PW box, so I can choose for myself!
8->

It is strange that there exist no standard for a separate control line between source and renderer that regulate the volume without loss. The challenge is how to implement this on the renderer side. I guess there has to be a way of commanding digitally an analog attenuator.

@johnjen that video was fascinating. Very useful to the topic here. I am left feeling somewhat pessimistic about keeping my DAC volume at 50 despite the internal chip architecture being 32-bit. It would seem that the SN ratio suffers progressively with each dB of digital domain attenuation... Gonna have to review those figures a few times! :-B
Yea that video and its brother are quite revealing, on many issues.

The subject that has attracted my attention of late is the temporal aspect in relation to the (re)creation of the musical presentation.
Another way of looking at this is, our musical playback systems are really dealing with just two primary factors. #1 is amplitude of the signal and #2 is the timing of that same signal.
This applies to both analog and digital signals.
Granted these 2 factors become 'significant' in different portions of the signal path, but basically 'how much' and 'when' are really what we are dealing with hear. ( :-" )

Just a stray thought or 2…

Merry & Happy seasonal experiences for all.

JJ
crabby said: hy does the pwd show highest s/n ratio when volume is at 67 as opposed to 100?


@crabby - can you remind us where this s/n stat is from? Also, do you therefore keep your PWD at 67?

Several companies produce DACs with an integrated analog volume control. Anyone tried Lampizator or MSB DACs with analog volume control? I’m interested to hear any opinions on how these sound.

I have the Msb one. I posed that question earlier in this thread. I even have anolog inputs as to compare both ways to adjust the volume.



Al. D

@alrainbow And what do you think of the MSB compared to the PWD?

My email. BALDINO472003@msn.com

I would be happy to give you details , but in all fairness to ps audio not in the public forum.



Al. D



@crabby - can you remind us where this s/n stat is from? Also, do you therefore keep your PWD at 67?


http://www.psaudio.com/vanilla/discussion/comment/53182#Comment_53182

http://www.milleraudioresearch.com/download2009/reports/aug09/psaudio_perfectwave_dac.html#sthash.8mU9zxKG.dpuf


I do keep mine at 67 though i doubt i can tell a difference.
stereophilus said: Anyone tried Lampizator . . .

This is such an odd piece of equipment that its volume control would be of little comparative interest. :) I would really like to hear one of these.

There are many DACs with attenuation switches and with full range analog volume controls. There are also a number of DAC/preamps, such as the Benchmark and Grace M903. All sound excellent.

I have the mytek. It has both digital and analog . The digital and preamp is horrible and I use it as a line out to pre amp . When I first bought I did not know why it just did not sound very well. After some more reading and trying it in bypass it is much better.



Al. D


There are many DACs with attenuation switches and with full range analog volume controls. There are also a number of DAC/preamps, such as the Benchmark and Grace M903. All sound excellent.


I should have been more specific. The PWD volume control is very good, but *maybe* it can be bettered by a really good analog volume control. Lampizator and MSB come to mind as companies that are implementing analog volume controls of very high quality. I am wondering if I can get the improvements of a great preamp while keeping the DAC/preamp in one box. Probably I am wrong, but I would like opinions on this.

Today I spent about 6 hours trying different dacs , great day to be me. And one of the tests I did was the volume test , both digital and anolog and some had none. And something funny happened some of what I have come learn here was used. How cool is that and it turns out the Pwd mkii has a dam good volume control too. I also got to listen how important equipment volume matching was too. I also had a listen to a msb dac. And it’s anolog volume control and it also had two anolog inputs , one balanced and the other SE.



Paul I wish I had even one with the pwdmkii , please add on. The hell with the wave stream I want a anolog input.



I will list the equipment and do a nice post in the appropriate section. I want to thank all here for making me a much better Audiophille. If I hang out here long enough I might even learn to spell and better over all linguistic skills LOL. gotta love this spell check here.

I found little difference between using MSB analog VC and playing it at 0db attenuation and using upstream digital VC in trinnov processor. However, inserting arc ref5se preamp into the chain blew both out of the water.

stereophilus said: I am wondering if I can get the improvements of a great preamp while keeping the DAC/preamp in one box.

I like the thought, but unfortunately an analog volume control will not do the work of an active preamp. As Paul and others have found, a preamp does more than attenuate gain.
I found little difference between using MSB analog VC and playing it at 0db attenuation and using upstream digital VC in trinnov processor. However, inserting arc ref5se preamp into the chain blew both out of the water.


I would love to try the ref5se as well... I have some saving to do first!

I like the thought, but unfortunately an analog volume control will not do the work of an active preamp. As Paul and others have found, a preamp does more than attenuate gain.


Even if the solution is an active preamp and DAC in 1 box?