This is a general question, not related to any specific sound card or Windows version.
If you drive an USB sound card into overload on its input, you can reduce the record audio level in the Control Panel - Sound - Record - Properties - Signal level
This control goes from 0% - 100%. What does it actually do? How does it interact with the sound card? Does it do some kind of digital attenuation? How could I get the best S/N? By increasing the analog input level and apply the Windows attenuation, or set the level back and push the control to 100%?
Also it is a bit disturbing there is no VU or dBu scale, neither VoltsRMS or any common level reference. You never know what is the actual level at the input of the soundcard, and how is it scaled to the digital domain, referenced to digital full scale.
I suppose Linux is not different either.
Any ideas, comments?
If you drive an USB sound card into overload on its input, you can reduce the record audio level in the Control Panel - Sound - Record - Properties - Signal level
This control goes from 0% - 100%. What does it actually do? How does it interact with the sound card? Does it do some kind of digital attenuation? How could I get the best S/N? By increasing the analog input level and apply the Windows attenuation, or set the level back and push the control to 100%?
Also it is a bit disturbing there is no VU or dBu scale, neither VoltsRMS or any common level reference. You never know what is the actual level at the input of the soundcard, and how is it scaled to the digital domain, referenced to digital full scale.
I suppose Linux is not different either.
Any ideas, comments?
Some possible links of interest:
https://www.diyaudio.com/forums/dig...z-music-files-cd-standard-11.html#post5993897
https://www.diyaudio.com/forums/digital-line-level/314935-es9038q2m-board-410.html#post5731319
In addition to what is said there, sometimes turning down the Windows volume control or the ASIO playback app volume control down a few dB from maximum not only can prevent intersample overs, its also true that some dac chips have lower HD if peak digital levels are kept perhaps no higher than -6dBFS (depends on the particular dac chip). As explained at the above links, probably best not to turn it down more than a few dB in software.
https://www.diyaudio.com/forums/dig...z-music-files-cd-standard-11.html#post5993897
https://www.diyaudio.com/forums/digital-line-level/314935-es9038q2m-board-410.html#post5731319
In addition to what is said there, sometimes turning down the Windows volume control or the ASIO playback app volume control down a few dB from maximum not only can prevent intersample overs, its also true that some dac chips have lower HD if peak digital levels are kept perhaps no higher than -6dBFS (depends on the particular dac chip). As explained at the above links, probably best not to turn it down more than a few dB in software.
Turning down by 6 dBfs means shift the slider to 50%? Also, what input level on the sound card will give full scale digital output? 1Vrms, 0dBu or there is no such convention?
Edit: I am not interested in Windows sound quality. My question relates to sound card audio processing (ADC). Windows receives 16 bit 48 kHz audio stream through USB in my application (FFT analyzis), so it should not alter it. That said, I don't understand what the level slider does on the Record tab.
Edit: I am not interested in Windows sound quality. My question relates to sound card audio processing (ADC). Windows receives 16 bit 48 kHz audio stream through USB in my application (FFT analyzis), so it should not alter it. That said, I don't understand what the level slider does on the Record tab.
Last edited:
Might depend on the device driver and ADC hardware. IIUC Windows can control volume levels in software, or send commands for the hardware to do it.
Thanks, but I don't understand a single word of it.
So far I followed the experimental method, and adjusted the Level setting in the Audio Devices - Record tab until the soundcard output started distorting. I found the optimal setting around 50...60%. This setting somewhat interacts with the analog Level setting on the soundcard, but it is possible to find the optimum for a given input level, 2V in my case. For higher level I have to use an external attenuator.
So far I followed the experimental method, and adjusted the Level setting in the Audio Devices - Record tab until the soundcard output started distorting. I found the optimal setting around 50...60%. This setting somewhat interacts with the analog Level setting on the soundcard, but it is possible to find the optimum for a given input level, 2V in my case. For higher level I have to use an external attenuator.
might depends on the material if you're only Red-Book 16 bits. I would putt the windows level at 100%, not being sure of where the attenuation is made in order not no loose bits. The harder thing for me is more the computer is a ground problem due to its internal smps and complex electric scheme but its off topic of course.
It very much depends on the device in question. The Windows sliders can either alter the gain digitally in Windows itself or it can issue commands out to the USB device itself.
If it sends commands to the USB device itself then Windows does absolutely nothing. In this case the gain/volume is then controlled by the USB hardware in some way. This could either be digital or analogue, or a combination of both, depending on how the hardware is configured.
Not useful I know but that's the way it is.
It's the same with PCI/PCI-E cards.
If it sends commands to the USB device itself then Windows does absolutely nothing. In this case the gain/volume is then controlled by the USB hardware in some way. This could either be digital or analogue, or a combination of both, depending on how the hardware is configured.
Not useful I know but that's the way it is.
It's the same with PCI/PCI-E cards.
Vast majority of external soundcards (USB, PCI(e)) use digital volume control provided by their DAC chips. Actually the highest chance of using the analog volume control is in Intel HDA codecs of internal soundcards on motherboards. The codecs often implement complex analog routing and volume controls of their individual routes/outputs. However that does not mean a better quality of output signal than digital volume control, just a technical nuance.
Windows always resamples any audio that goes through its audio engine that didn't come from a source file at the exact same sample rate and bit-depth as the sound device default setting. The only way around that is to use WASAPI Exclusive Mode, or else ASIO. There are many devices available with support for the latter option.
The resampling is audible to me in double-blind tests. That's the biggest problem I have with it.
The resampling is audible to me in double-blind tests. That's the biggest problem I have with it.