Digital and DAC Volume control

[Subtly_In_Class]

Hello everyone,

I have been wondering about something for a while regarding the volume control done by windows vs the volume control done by most DACs. To preface this, I'm not looking for a theoretically perfect maximum bit depth, I'm only concerned about what can and cannot be heard by human ears.

To my knowledge, windows uses 32 bit floating point for volume control, and what I was curious about is if this means that you could turn a signal down by a full 98 dB (16 bits) and still preserve the full dynamic range of a 16 bit PCM playback? To further this, I was also curious about how most dacs that had a volume adjustment feature worked. The way I see it, it could be done both on the analog side and the digital side, either by a series of resistors after the AD conversion or by reducing the bit depth before the AD conversion, and I'd be curious to know the approach that most DACs take (I'd especially be curious to know what approach Topping DACs take).

The reason I'm asking about all this is that I have quite sensitive headphones that I'm using a pretty powerful amp to power (if anyone's wondering I'm using an iFi Zen Can to power Beyer T5 gen 3's) and as a result I have to turn my volume down on my dac by 18-24 dB, depending on what I'm listening to. I don't like doing this on the amplifier as the volume control only has good channel balance if it is turned up at least half way. So I was wondering if it was better to turn the volume down on my Topping E30 or on Windows itself. I know Topping advertises the E30 as 32 bit capable but I also know that delta-sigma DACs don't truly decode the bits they claim, rather using interpolation. So I'm not sure how the E30 is actually reducing the volume. For what it's worth I can't hear a noticeable difference between turning the volume down in Windows or on the DAC, but I would be curious to know.

 

[Mike123]

After a series of experiments, I settled on a passive resistive volume control. It reduces not only the signal volume, but also the DAC noise volume. With a low power amplifier noise level and high requirements for the output noise level, this option is more promising.

 

[somebodyelse]

To my knowledge, windows uses 32 bit floating point for volume control, and what I was curious about is if this means that you could turn a signal down by a full 98 dB (16 bits) and still preserve the full dynamic range of a 16 bit PCM playback?

In theory maybe. In practice No.

We haven't seen any DACs measured here that reach 24 bits worth of dynamic range, although we are getting closer. IIRC the best is about 22 bits worth. Usually this doesn't matter because that's still more than enough to cover the range between the noise floor in the listening space and the loudest noise you want to hear.

If you can't hear the DAC's noise in your headphones then even with the excessive gain of your headphone amp for those headphones you still don't have a problem. Make sure it's the DAC's noise you hear, not the headphone amp's - is the noise the same for the DAC playing silence and the headphone amp input shorted? If you can hear the DAC's noise then you can either use some inline attenuation before the headphone amp input, or get an amp with gain more suited to your headphones.