• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

RME Adi-2 dac - stepped attenuators - what is the point?

nscrivener

Member
Joined
May 6, 2019
Messages
76
Likes
117
Location
Hamilton, New Zealand
I was just thinking about the volume control function on my ADI-2 FS dac. It uses a combined analog/digital volume control, it switches between the analog steps and uses digital volume attenuation (the manual says using 48 bits) in between the steps in order to make the volume transitions smooth. It works perfectly and other than an audible click from the DAC when it switches over to a different output you wouldn't know it's there.

I understand the theoretical benefit of this is that it maximises the SNR over a wider range of volumes, combining the benefits of both digital and analog attenuation.

However, is there really any point for this, when the noise level is already below audible thresholds? Yes you might be technically reducing the SNR when you digitally attenuate, but the noise is already below your hearing threshold anyway so how does this provide any advantage from a straight digital attenuation with no analog steps? At 48 bits any quantization noise introduced from the volume control would be miniscule...
 

Rja4000

Major Contributor
Forum Donor
Joined
May 31, 2019
Messages
2,682
Likes
4,220
Location
Liège, Belgium
At 48 bits any quantization noise introduced from the volume control would be miniscule...
It's not about the noise of the volume control itself.
It's about the analog noise of the DAC.

The RME allows to attenuate analog stage by as much as 20dB (from +24dBu @0dBFS down to +4dBu), by switching analog output ranges.
That's not minuscule.

The difference in actual noise level is not that big, but we still speak of around 10dB.
To be compared to the around 120dB of dynamic range of the DAC's analog stage.
That's roughly 2 bits out of 20.

If you listen at, say, -20dBFS, SNR is around 100dB at best.
10dB difference is the difference between 100dB and 90dB. That may be significant.

000228-THD+N_CompareLevels.png

As an example, in the attached measured graph, measured with RME ADI-2 Pro fs, at a listening level of 0dBu (-13dBFS of the input, on the X scale), SNR is around 105dB for +4dBu range and 98dB for +19dBu range.
That's 7dB improvement between +19dBu and +4dBu range.

https://www.audiosciencereview.com/...essary.9327/page-3#lg=attachment35891&slide=0
 
Last edited:
OP
nscrivener

nscrivener

Member
Joined
May 6, 2019
Messages
76
Likes
117
Location
Hamilton, New Zealand
It's not about the noise of the volume control itself.
It's about the analog noise of the DAC.

Thanks for your reply.

Well actually all digital volume control will introduce some noise because of quantization. I was just ruling that out as insignificant due to internal processing algorithms.

Your answer seems to be focused in terms of absolute terms.

But what seems to me is that you have a noise level that is most likely imperceptible to the listener regardless of how much signal there is over that noise level. As long as the dac has good linearity, which the ADI-2 does. A SNR of say 50db at a high volume is clearly worse than at a quiet volume as the noise is amplified within the range at which your ear can hear the noise. But if you have a SNR of say 120db and you attenuate the signal but not the noise, then you have the same absolute noise level as before. You might get down to a SNR of 50db but that will just mean your playback is very much quieter than before. The noise from the dac was at an imperceptible level before and remains so after the attenuation.

Also the graph you put there shows SINAD. Distortion isn't really relevant to this conversation as the distortion products will scale with the volume control (as will any noise floor on the recording itself). It's only the noise floor of the DAC we are concerned with.

Maybe I'm missing something.
 

MC_RME

Addicted to Fun and Learning
Technical Expert
Audio Company
Joined
May 15, 2019
Messages
854
Likes
3,564
People will use the unit very differently, in many different applications and setups. The combined volume control makes sure that the unit delivers the highest SNR at all times. Example: if you set the phones output to Hi-Power and connect a very sensitive headphone you can hear the noise floor. Lo-Power decreases the noise floor by a whopping 15 dB - no longer audible, something not possible with digital volume control alone.

At the XLR outputs it is exactly the same. It depends on what you connect. A high level mismatch that requires a strong digital attenuation might not only make the DAC's noise become audible, it also opens the door for other kinds of unwanted noise. Plus the combined level setting can easily prevent you from blowing the speakers by accidentally raising output level to the max.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,522
Likes
37,053
I think the reason RME provides variable analog output levels is simply due to various use cases. Home use, a lower output will work best with your amp input. A pro studio use with active monitors that require higher input levels than consumer gear, you dial up the output voltage. As you say the available dynamic range is very good in all cases. It is just a matter of being able to match output levels of the RME with various devices it may feed whether consumer, studio or what have you. It is part of gain staging. Even with quiet gear like the RME, you don't want to output a high voltage which you then have to reduce just to keep from clipping an amp it feeds. You are throwing away dynamic range to some extent when it is better not to do so.

EDIT: I see RME beat me to it by a few seconds. At least I am in agreement with them.
 
OP
nscrivener

nscrivener

Member
Joined
May 6, 2019
Messages
76
Likes
117
Location
Hamilton, New Zealand
People will use the unit very differently, in many different applications and setups. The combined volume control makes sure that the unit delivers the highest SNR at all times. Example: if you set the phones output to Hi-Power and connect a very sensitive headphone you can hear the noise floor. Lo-Power decreases the noise floor by a whopping 15 dB - no longer audible, something not possible with digital volume control alone.

At the XLR outputs it is exactly the same. It depends on what you connect. A high level mismatch that requires a strong digital attenuation might not only make the DAC's noise become audible, it also opens the door for other kinds of unwanted noise. Plus the combined level setting can easily prevent you from blowing the speakers by accidentally raising output level to the max.

Ok, that makes more sense now, in the context of headphones at least. In terms of DAC use, if you had a very high gain amp, such that a lot of digital attenuation was necessary, I would have thought some simple XLR attenuators on the output could resolve that issue. But given it's already in the unit, that is one less thing to worry about.

Thanks for your reply!
 

Rja4000

Major Contributor
Forum Donor
Joined
May 31, 2019
Messages
2,682
Likes
4,220
Location
Liège, Belgium
the graph you put there shows SINAD. Distortion isn't really relevant to this conversation as the distortion products will scale with the volume control (as will any noise floor on the recording itself). It's only the noise floor of the DAC we are concerned with
I published this one because I had it at hand.

But at those levels, on the RME, SINAD figure is pretty much identical to SNR figure.
Distortion is so low that it doesn't make any significant difference.
(around 0.1dB difference, as far as I remember)
 

Mnyb

Major Contributor
Forum Donor
Joined
Aug 14, 2019
Messages
2,640
Likes
3,608
Location
Sweden, Västerås
Yes you can use attenuators if you got them , but this is quite nifty many pros may swap gear and work at different places with with different gear.
And gain and input sensitivity is not as standard as we like it to be espacially on consumer gear , so it's cool with product that adjust automatically for this .

Consumer gear , I had a mixed an match HT systems once with power amps from 3 diffrent brands one amp had the sensible function of adjustible gain (which for mysterius reasons are ommitted on very expensive audiophile gear ? pay $$$$ and flush 10-20dB of sn ratio down the drain ).
I opted to use attenuators on one amp .

Good active speakers have also the option to adjust gain but not all of them .

So RME have "fixed" that weird case you will stumble upon soner or later :) nice of them .

mini rant :
Gain structure is very neglected in home audio and high end everything has to much gain and is to senstive .
The signal has to be "born" at the DAC at a reasonably high level , because the bottom is not moving the laws of physics makes it so that you get some thermal noise in even the best equipment that dont go away.
 
Top Bottom