Can you use digital volume control and skip a pre-amp?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Yes, you can with a few risks:
(1) Without manual control of the pot, unintentional maximum input signal can damage the speaker.
(2) Some sources have weak output so you can not reach the amplifier maximum output. Turn table is a special source that require special preamp. But modern sources have sufficient output level.
(3) Some amplifiers require voltage gain stage more than others to reach acceptable output power level. Special case is class-A amplifiers consisting of only a buffer stage (Pass' F4, Power Follower, etc).
(4) Some sources have low damping and some amplifiers are hard to drive (low input impedance). In this case you need a "buffer" which usually exist in a preamp so that distortion level is "under control".

The good thing is, without the preamp you have removed one source of noise generator. With a preamp, if you can control the noise to acceptable level, the sound quality (sound details) is usually better.
 
I do this very thing when I stream audio to my loudspeakers. I control the volume at the source. Indeed you do lose resolution and dynamic range. It's up to you to determine if this is undesirable and decide what to do about it (implement volume control in a different way, use more bits, etc.).

It's possible to get a very loud surprise if/when the software decides to reset the volume to maximum when you don't expect it, e.g. when the program is upgraded or whatever. I now check the volume indicator each time before hitting play...

You should check the soundcard output maximum level (sometimes around 1Vrms) compared to your amp's input sensitivity.
 
Going from a DAC directly into a power amp (with no voltage trimming of its own) may reveal the need for surprising amounts of practical dynamic range if you value your peace and quiet. Nothing unmanageable, mind you, but about 103 dB at least (assuming you want at least 100 of these with output level ~2 Vrms and input sensitivity ~1.5 Vrms), maybe 110 dB - so no way around 24 bit operation there! The specifics would depend on power amp gain and how sensitive your speakers are.
 
Another relevant issue is the amount of background noise (in the signal chain, not your listening room :) ) is there? Especially if you get into active EQ or other multi-connections to your speakers, it is more convenient to control the master volume, whether from a PC's slider or a preamp. The overall answer is, of course, "it depends." This is all part of gain staging: to get the maximum signal to noise ratio through the system. 60 Hz belongs in the wall and hiss on your old cassettes, not on the background of everything you listen to :D

Sent from my NV570P using Tapatalk
 
I do not understand how volume control in 32/48/64/... bits can eliminate the reduction of bit resolution of the attenuated signal going to 16/24bit soundcard/dac. Either it attenuates (thus eliminates the lower bits of 16/24 width), or it does not attenuate (keeps the bits, but does not do the job :) )
 
I do not understand how volume control in 32/48/64/... bits can eliminate the reduction of bit resolution of the attenuated signal going to 16/24bit soundcard/dac. Either it attenuates (thus eliminates the lower bits of 16/24 width), or it does not attenuate (keeps the bits, but does not do the job :) )

in pseudo binary :)

We have two sample voltages 02 and 03 and attenuate to half volume, now we have 01 and 01 (or 02 if we round up). Either way we have lost the 02:03 ratio between successive samples so we have lost resolution.

Now lets pad these with a 'bit' to 020 and 030. Attenuate to half to 010 and 015 still maintains the original ratio so we have lost no resolution. If we pad with quite a few bits we can do quite a bit of digital attenuation :)
 
I do not understand how volume control in 32/48/64/... bits can eliminate the reduction of bit resolution of the attenuated signal going to 16/24bit soundcard/dac. Either it attenuates (thus eliminates the lower bits of 16/24 width), or it does not attenuate (keeps the bits, but does not do the job :) )
Indeed.

And while the typical listening volume is rather low, you typically listen to pretty low resolution anyway.
At least, attenuate the signal out of your PC as much as possible with a resistive divider, or reduce amp gain.
Your goal should be that your desired max listening level corresponds to the max volume of your playback software.
 
Now lets pad these with a 'bit' to 020 and 030. Attenuate to half to 010 and 015 still maintains the original ratio so we have lost no resolution. If we pad with quite a few bits we can do quite a bit of digital attenuation :)

Yes, but in the end your 015 will get truncated to 01 or 02 anyway (depending on actual algorithm) since the DAC accepts only 16 or 24 bits. In the end all the precisely calculated lower bits will be truncated = dropped/lost.
 
I don't think that using 32bits at this level is totally pointless. It gives a better precision to the calculations, though not a better resolution.

I just wanted to point out that if the musical program is 16bits, it will be 16bits with a 16bit dac, 16bits on a 24bit dac, 16bits on a 32bits dac, etc...

Maybe a state of the art 24bit or 32bit dac works better than a 16bits oldie, but with a 16bits program, only 16bits will be used... and attenuated by software if the case.

Imho, the potential of these dac will be under exploited, unless resampling :)eek:) is applied.:cool:

But using the software volume control where it is usually implemented (in MPD for instance) will change nothing to the loss of the resolution issue. Software volume control should be applied after resampling, not before, as is usually the case.
 
Last edited:
Yes, but in the end your 015 will get truncated to 01 or 02 anyway (depending on actual algorithm) since the DAC accepts only 16 or 24 bits. In the end all the precisely calculated lower bits will be truncated = dropped/lost.

OK, I missed the qualification in your question about going to a 16/24bit soundcard :)

If its a 16bit soundcard and the music is 'properly' recorded to full scale 16 bits at max signal then yes, any attenuation in software has to be lost bits.

If its 16bit source going to 24bit soundcard we do have 8 bits to attenuate with though (perhaps if the source recording is made a few db down we might also have a bit or two to play with even on a 16bit card). Ideally of course any digital attenuation would be inside the DAC working at a higher upsampled resolution even if it is accepting 16/24 bit material, if you are fortunate enough to have such a beast.
 
JRIVER website about 64 bits volume control

:eek: paste & copy from JRiver site :

Why JRiver?

For the purist
Because the sound engineered into the CD is the sound you will hear when you play the audio. No changes are made to it by JRiver. Nothing will come between you and the recording. The promise we deliver is fidelity and ease of use.

For the person who likes to experiment
If you need to manipulate the sound, you can use JRiver to modify playback. Upmixing, downmixing, bitdepth, speaker control. Because the audio path is fully 64 bit, any adjustments to volume, bitdepth and sample-rate are mathematically lossless.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.