24 bit 96 k FLAC, what am I doing wrong?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
When I play 24 bit 96 kHz FLAC files from my laptop, using Foobar2000, via USB, to my DAC (Pro-Ject Pre Box S2 Digital), the display on the DAC says 48 kHz. The Foobar application screen says 96 kHz. I suspect that somewhere along the line the 96 kHz is being down sampled to 48 k.

Do I need a special driver to support USB bit rates at 96k? Do I need ASIO drivers ?
 
Re-sampler plugin active in Foobar2k ...
Actually, not all devices have automatic sample rate switching (or not all driver modes supports it) so, you might need to set the device sample rate manually (or try other drivers as like ASIO ) then (where from you set it depends on your OS).
 
Check what options are listed in the device options (sound settings --> device properties --> additional device properties for Windows 10). Under advanced, you can select a default format and press a test button to see if sound pops out at the other end.
 
Problem solved... I was (stupidly) trying to troubleshoot the device settings with the device (DAC) not connected to the USB port. When the DAC is connected to the laptop, many more options are available. I noticed that the output was fixed at 32 bit 48 kHz. Ah ha! The source of the problem. I was able to set it to 32 bit 96 kHz.

In the process of troubleshooting this, I tried loading Audirvana and using it instead of Foobar2000. Audirvana got the sample rate correct right out of the box. Both applications work well now, but I think I like Audirvana better. I have 30 days of free trial usage before I have to decide, but for $USD 75, I think Audirvana is something I will purchase.

Jim
 
Windows does all kinds of upsampling and downsampling of Audio before handing off to the DAC in whatever bit rate you specify in audio device settings. Not to mention the "enhancements" it applies of Loudness Equalisation, Bass Boost etc.


To get an unmolested audio playback install the WASAPI Output Plugin for Foobar, then the DAC will play in whatever sample rate and bit depth the source material is recorded in.



foobar2000: Components Repository - WASAPI output support
 
mcandmar, annoyingly, I just lent out my good PCM recorder but, because I wanted to check my audio path, I recently spat out some ultrasonics through Winamp via a 192 and 96/24 FLAC as well as using a few tracks with those present and it certainly made it through on the frequency front. I mention lending it because I could check the bit depth by adding a frequency that's 1/217 of peak, meaning it would be gobbled up by compression to 16 bit or below at some stage but I won't be getting it back until Christmas. All my 'enhancements' are disabled.

Has anyone done any measurements of WASAPI vs others? I couldn't find anything concrete.

edit: full disclosure: I'm just using budget headphones (BD DT990 Pro, Fostex T50RP, Grado SR125e).
 
Last edited:
In a previous life, I spent 25 years on a team of engineers whose main product was a sophisticated engineering software application which simulated aircraft structural loading and metal fatigue. For the last 10 years, I was the leader of this team. The software was both a unix and windows application, with ~ 100,000 lines of code.

Even with all of that, I am still not confident I understand how audio drivers work in windows... I don't understand the differences between ASIO and WASAPI, and I am not confident I would be able to get "bit perfect" PCM data from my hard drive to my DAC... There is nothing quite like Windows to make a person feel stupid.

That is why Audirvana appeals to me. I am stupid, and I admit it...
 
Even with all of that, I am still not confident I understand how audio drivers work in windows... I don't understand the differences between ASIO and WASAPI, and I am not confident I would be able to get "bit perfect" PCM data from my hard drive to my DAC... There is nothing quite like Windows to make a person feel stupid.
Windows Audio Architecture - Windows drivers | Microsoft Docs

What's of interest to us is that both potentially enable fairly low-level access to the sound hardware. Windows (Vista and up) sound drivers MUST provide a WASAPI (exclusive mode) interface to be of any use in everyday life. They MAY also provide an ASIO interface (and in equipment geared towards music production you may well see the main focus on that, with only rudimentary functionality available via WASAPI), and on some older soundcards you would also have seen some native DirectSound access with WASAPI internally piled on top of that.

IMHO it's not so much Windows that complicates things, but rather having to deal with additional abstraction layers in legacy sound hardware and generally quirky drivers. For example, the CMI878x based Asus Xonar cards as well as X-Fi and older Creative Sound Blaster cards have internal sample rate and format controls that have to be adjusted manually, as the hardware generally features some resampling capabilities of its own. This potentially leaves you with having to keep sample rate in sync in up to 3 places:
* playback software
* Windows audio interface
* driver control panel

WASAPI exclusive mode would sync the first two but not the third. Unfortunately there are cases when you best off with shared mode, as I found in my Xonar D1/D2 adventures. I've been dabbling in PC sound for something like 20 years now, and somehow good hardware hobbled by bad driver support keeps on being a recurring theme.
 
For example, the CMI878x based Asus Xonar cards as well as X-Fi and older Creative Sound Blaster cards have internal sample rate and format controls that have to be adjusted manually, as the hardware generally features some resampling capabilities of its own.

Every soundcard must be told what rate, format (if more supported), and channel count (if more variants supported) the incoming stream of bytes represents. It's up to the driver authors what level of control they leave to the user and what will be set automatically from parameters of the audio stream sent by the audio layer of an OS.

Most soundcard "features" are delivered by SW DSP bundled in the windows support package (the actual driver is only tiny part of most windows HW-enablement bundles). Only a handful of soundcards have HW DSP onboard, e.g. the Creative/EMU complicated audio interfaces. CMI8788 is a standard "dumb" PCI audio controller like ICE1724 http://www.hardwaresecrets.com/datasheets/OxygenHD CMI8788 Datasheet Rev0.6.pdf
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.