How does Windows handle audio device sample rate and bit depth?

[JEntwistle]

Yes, my point was that "HD" is a pretty meaningless marketing term.

Perhaps, but if you use it at least make sure your solution delivers whatever it means.

gvl, are you talking semantics, i.e.: that Amazon is improperly using the term "HD"? They do relabel things to their own nomenclature, but they also do provide some music in true HD. Their labeling is "nothing" = compressed; "HD" = what is usually referred to as CD quality; and "Ultra HD" = what is usually referred to as Hi Res or HD.

They are confusing people with the terminology. There really is no logical reason to confuse the labeling; I don't disagree with that. But once you get past the marketing, they are actually providing quite a large selection of true high resolution content.
- Although I would add that I have found that while they may label an album as "Ultra HD", you will often find that some of the tracks on a given album will be at the lower "HD" quality.

 

[gvl]

Even if you set the sampling rate by hand and it matches your material you're still not protected from internal processing in the OS mixer..

 

[JEntwistle]

Even if you set the sampling rate by hand and it matches your material you're still not protected from internal processing in the OS mixer..

Hm.. That's what I was afraid of. It would be nice if Amazon provided a solid software solution.

Thank for the help!

 

[gvl]

gvl, are you talking semantics, i.e.: that Amazon is improperly using the term "HD"? They do relabel things to their own nomenclature, but they also do provide some music in true HD. Their labeling is "nothing" = compressed; "HD" = what is usually referred to as CD quality; and "Ultra HD" = what is usually referred to as Hi Res or HD.

They are confusing people with the terminology. There really is no logical reason to confuse the labeling; I don't disagree with that. But once you get past the marketing, they are actually providing quite a large selection of true high resolution content.
- Although I would add that I have found that while they may label an album as "Ultra HD", you will often find that some of the tracks on a given album will be at the lower "HD" quality.

No, I was terminology agnostic. HD, lossless, whatever, if it is marketed as such I want to be able to play it without bytes being butchered by the OS mixer.

 

[JEntwistle]

No, I was terminology agnostic. HD, lossless, whatever, if it is marketed as such I want to be able to play it without bytes being butchered by the OS mixer.

Ah, I get it now. So you are having the same issue as I am, essentially.

How would they be able to implement it, though? Would you then need another format decoder in your external DAC?

 

[gvl]

Hm.. That's what I was afraid of. It would be nice if Amazon provided a solid software solution.

Thank for the help!

Your best bet is is to try something like JRiver and its WDM driver, this is a virtual device driver that you set as the default OS device and it is able to bootleg music samples bypassing the mixer into JRiver, from there you can forward it onto the ASIO or WASAPI exclusive device.

 

[gvl]

Ah, I get it now. So you are having the same issue as I am, essentially.

How would they be able to implement it, though? Would you then need another format decoder in your external DAC?

Their app needs to be able to talk to an ASIO driver or WASAPI exclusive interface. It can't, maybe in the future it will be able to. They were either lazy, rushed, or plain ignorant in this regard.

 

[audimus]

Even if you set the sampling rate by hand and it matches your material you're still not protected from internal processing in the OS mixer..

This is not necessarily bad from a fidelity point of view, only increases latency. Not a problem for music streaming, only for gaming, recording, real time communications, etc. But you might get interrupted by other system sounds or other applications making sound.

Unless you have also turned on some effects processing that the mixer is able to do. There is no digital to analog conversion and if there is no sample rate conversion, the same PCM that comes in will come out. Since the audio engine can only handle PCM, this can be a problem for applications that want to pass through encoded formats like Dolby, DTS, etc as typical in HT applications. So they cannot do pass through unless they use direct mode. Music streamers send PCM anyway (as far as I know almost all do).

Your best bet is is to try something like JRiver and its WDM driver, this is a virtual device driver that you set as the default OS device and it is able to bootleg music samples bypassing the mixer into JRiver, from there you can forward it onto the ASIO or WASAPI exclusive device.

Voicemeeter is free unlike JRiver. But you would not avoid going through the Windows audio engine to get to the virtual sound devices if the app cannot do direct mode. But like I said above, this is not necessarily a problem. What Voicemeeter or JRiver will do is take any sample rate that the player provides and feed it bit perfect into the output or do a competent sample rate conversion if necessary.

Their app needs to be able to talk to an ASIO driver or WASAPI exclusive interface. It can't, maybe in the future it will be able to. They were either lazy, rushed, or plain ignorant in this regard.

They don’t need to go into those low level APIs unless latency was an issue which is not in music streaming. Just Windows Direct Mode is fine (which is not that different from the APIs they are using now in shared mode) but they will need to be prepared to do sample rate conversion if they have to go direct since no one in the middle will in direct mode.

 

[audimus]

Audimus, this is quite helpful. Many thanks. A few questions, if you don't mind:

- The TEAC DAC provided an ASIO driver for Windows, which I have installed. In simple terms, does this mean that if I am using the TEAC software, which uses these drivers, it is bypassing the Windows mixer and handling the data itself? This is what I use for files I have on my PC, and the DAC seems to indicate it is recognizing the correct format and sample rate.

I don’t know what the TEAC software (as opposed to its driver) does but if it is a music player meant for playing music files to its DAC, then it is most likely bypassing the mixer. Does it give an audio output device selection in its preference settings assuming it has one? If so, the choices will likely tell you what it is capable of and what the current chosen mode is.

- With regard to the Amazon player, exclusive mode would only apply if the application recognizes direct mode, and the consensus is that the Amazon software uses shared mode? So in this case, the DAC will see whatever is selected in the Windows properties; i.e.: I would need to change that manually based on the input format, since Amazon's software is likely not automatically changing it, for example when the source changes from CD to HD quality? (BTW, the "exclusive mode" selection is now on the Advanced tab in Windows 10 (see below)).

Yes, that is my understanding of the Amazon Music Player. The Exclusive mode selection has always been in the Advanced Tab as I said in the earlier post “below your sample rate selection”. The endpoint supported parameters are in the Supported Formats tab. What is enabled in that tab is what the software going through Direct Mode will see as supported formats to send as is or do sample rate conversion.

 

[JeffGB]

I have a Teac ud-301 dac which was supplied with the same software. In the last few months I have uninstalled the TEAC driver and use the Windows 10 native driver which has been working very well. In addition, you can use any of the standard players with WASAPI exclusive mode, which I believe runs bit perfect. I set the windows setting to 32bit 44.1khz and if I play a 24/96 file through Musicbee or whatever, the dac displays the correct sampling frequency, but if I use normal windows applications, it uses the 32/44.1 resampling for all the odd stuff found on the internet. This all works better now with the new Windows 10 usb2 audio driver than it did with the TEAC driver.

 

[JoachimStrobel]

I am not an expert too but have accumulated the following:
Amazon HD is using the Windows mixer. Even if you set the sample rate to the same as the source, it will not be bit-perfect, not sure why. You can disable “system sound” in the mixer and then at least do not get windows’ internal sounds messing around with your audio stream.
It may be tempting to sample everything to 96 or 192 kHz to be on the save side. But besides this not being bit perfect, there was a comment somewhere that these high sampling rates are very demanding for the converter/algorithm and amplitude and time errors might creep in cause by imperfection in the up-sampling process, at least with ordinary PCs.
I leave the setting at 44khz, the lowest HD rate.

 

[amirm]

Amazon HD is using the Windows mixer. Even if you set the sample rate to the same as the source, it will not be bit-perfect, not sure why.

All sound going through the kernel mixer will be converted to floating point. It will be processed (e.g. mixed with other sounds) and then converted back to integer PCM format. This involves addition of dither. As a result, bits will be modified even if the sample rate is not.

One way to minimize this effect is to set Windows bit depth to 24 bits. That way the dither will be buried in the noise of 24-bit samples.

Another reason there will be changes is that the mixer will attempt to build headroom in the digital samples so that they don't overflow when you mix two max amplitude PCM signals.

 

[JoachimStrobel]

So, a side question: My DAC only does 20 bit (a 20 bit D20400A in a Parasound DAC 2000 ), and potentially only with a HDCD bit set.
What happens if I feed it with 24 bit - does that have a negative impact?

 

[Julf]

So, a side question: My DAC only does 20 bit (a 20 bit D20400A in a Parasound DAC 2000 ), and potentially only with a HDCD bit set.
What happens if I feed it with 24 bit - does that have a negative impact?

No.

 

[gvl]

So, a side question: My DAC only does 20 bit (a 20 bit D20400A in a Parasound DAC 2000 ), and potentially only with a HDCD bit set.
What happens if I feed it with 24 bit - does that have a negative impact?

For 16 bit material no, for 24 bit maybe. Do some research If the DAC truncates to 16 bits internally, if so you may want to apply dither to 16 bits prior to feeding the data to the DAC.

 

[satrams]

OK- what about for TIDAL- and if you are not currently using a DAC but just the 'integrated audio' of the computer. Will the bit depth and sample rate automatically be changed to the propper values when in exclusive mode? Is there any way of seeing what the true current rates are?