I have a question, hoping that some experienced user can give me the answer.
When using WaveIO in combination with Foobar, one has to install the Wasabi driver and tell Foobar what to do, including which Fs and bitdepth has to be used.
Suppose 192Khz/24bit is entered here as the default.
Does that mean that all content offered to Foobar is converted to 192/24 regardless of the input sample rate and bitdepth ?
Hans
When using WaveIO in combination with Foobar, one has to install the Wasabi driver and tell Foobar what to do, including which Fs and bitdepth has to be used.
Suppose 192Khz/24bit is entered here as the default.
Does that mean that all content offered to Foobar is converted to 192/24 regardless of the input sample rate and bitdepth ?
Hans
Do you mean Wasabi or WASAPI? Asking because there is DJ software called Wasabi.
If you mean WASAPI, Foobar to this day required to install WASAPI add-on output extension (not a system driver) to use this Windows interface. The alternative add-ons required for bit-perfect playback are ASIO and KS (Kernel Streaming - now obsolete). Since version 1.6 WASAPI Shared access is built-in, still required to install WASAPI Exclusive extension.
foobar2000: Change Log
If you download Foobar now v1.6.7, it has WASAPI Exclusive support built in. You only need to follow guidelines to use bit-perfect playback like this: Foobar 2000 for Dummies (Part 1) – General Setup | DIY-Audio-Heaven
When using WASAPI Exclusive output device, Foobar will not resample by default, but keep a volume at 100% for bit-perfect transfers and disable replay gain, as shown in the guide linked above
If you mean WASAPI, Foobar to this day required to install WASAPI add-on output extension (not a system driver) to use this Windows interface. The alternative add-ons required for bit-perfect playback are ASIO and KS (Kernel Streaming - now obsolete). Since version 1.6 WASAPI Shared access is built-in, still required to install WASAPI Exclusive extension.
foobar2000: Change Log
If you download Foobar now v1.6.7, it has WASAPI Exclusive support built in. You only need to follow guidelines to use bit-perfect playback like this: Foobar 2000 for Dummies (Part 1) – General Setup | DIY-Audio-Heaven
When using WASAPI Exclusive output device, Foobar will not resample by default, but keep a volume at 100% for bit-perfect transfers and disable replay gain, as shown in the guide linked above
Last edited:
Of course a silly typo 😀
I meant Wasapi and not the Japanese Horseradish, but this in combination with Foobar and WaveIO.
DDDAC 1794 NOS DAC - Non Oversampling DAC with PCM1794 - no digital filter - modular design DIY DAC for high resolution audio 192/24 192kHz 24bit
Hans
I meant Wasapi and not the Japanese Horseradish, but this in combination with Foobar and WaveIO.
DDDAC 1794 NOS DAC - Non Oversampling DAC with PCM1794 - no digital filter - modular design DIY DAC for high resolution audio 192/24 192kHz 24bit
Hans
Im no expert but Foobar usually allows bit depth to be selected with or without WASAPI. WASAPI is brilliant for Windows. Your know its working as the Windows volume control is usually bypassed.
Some USB devices can require specific bit depth and will not work with the wrong selection. All files streamed out will be at the selected bit depth regardless of the original bit depth. So with an Amanero USB that needs 32bits streamed to it, all 16bit files will be padded out to 32 bits but will only have 16bits of actual signal. Other USB devices are more flexible and you can select the bit depth you want. Just play back at low volume initially in case the DAC puts out white noise with bit mismatch.
As for sample rate that's usually even more configurable as most USB devices allow multiple rates. You can play back NOS by default most of the time or upsample as you like. Resampler-V allow you to select which sample rate files to upsample, the upsampled rate, the filer eg SOX, the impulse ringing you want, phase, roll off, etc.
Foobar is a DIY gift.
Some USB devices can require specific bit depth and will not work with the wrong selection. All files streamed out will be at the selected bit depth regardless of the original bit depth. So with an Amanero USB that needs 32bits streamed to it, all 16bit files will be padded out to 32 bits but will only have 16bits of actual signal. Other USB devices are more flexible and you can select the bit depth you want. Just play back at low volume initially in case the DAC puts out white noise with bit mismatch.
As for sample rate that's usually even more configurable as most USB devices allow multiple rates. You can play back NOS by default most of the time or upsample as you like. Resampler-V allow you to select which sample rate files to upsample, the upsampled rate, the filer eg SOX, the impulse ringing you want, phase, roll off, etc.
Foobar is a DIY gift.
I used WaveIO before, Lucian will send you appropriate drivers - you should be able to use Lucian's ASIO driver pack with Foobar. Ask him for the latest firmware as well.... see how that goes.
I sent one of my WaveIO cards back to him, for NDK SD to SDA upgrade, which was a noticeable improvement in areas that... can't be improved by any other means. I was able to easily pick the card with lower phase noise oscillators.
I sent one of my WaveIO cards back to him, for NDK SD to SDA upgrade, which was a noticeable improvement in areas that... can't be improved by any other means. I was able to easily pick the card with lower phase noise oscillators.
Hi, Hans,
This is a good start. It's more reaction reports than we recieved from the prior 176.4 file set 😀.
Given the interesting combinations of file set comparisons which are possible, it's probably irresistible for participants not to audition most of them. I had expected the reaction reports to initially flow in a bit slowly for that reason, but then to pick up, so things are about where I thought they would be. 🙂.
Last edited:
As an interesting fact to know for the contenders to this test:
My Dac has a SRC4392 that is restricted to 24 bits and a PCM1792A restricted to the same 24 bits.
That's why I didn't even try the 32 bit versions.
But to my big surprise the files are accepted as such but the last 8 bits are simply discarded!!.
So be careful when thinking to hear a difference between 24 and 32 bits.
Hans
My Dac has a SRC4392 that is restricted to 24 bits and a PCM1792A restricted to the same 24 bits.
That's why I didn't even try the 32 bit versions.
But to my big surprise the files are accepted as such but the last 8 bits are simply discarded!!.
So be careful when thinking to hear a difference between 24 and 32 bits.
Hans
.LOL. Please don't remind us that we are so incredibly stupid. 🙂
There is still a difference between these 176k files, as after truncating the file remain undithered. However I don't think it will make any difference at 24-bit level. I would rush myself to confirm this, but I am now recovering on pain killers. Not a good time for testing...
It is worth however to compare it with a previous 176k file which was noise shaped.
There is still a difference between these 176k files, as after truncating the file remain undithered. However I don't think it will make any difference at 24-bit level. I would rush myself to confirm this, but I am now recovering on pain killers. Not a good time for testing...
It is worth however to compare it with a previous 176k file which was noise shaped.
Not knowing the facts Hans shared does not mean someone is stupid, so it is ok to take it a notch down and concentrate on getting healthy again. Hope nothing serious
Stupid is defined as "having or showing a great lack of intelligence or common sense."
Ignorant is defined as "lacking knowledge, information, or awareness about a particular thing"
Hence those that aren't familiar with such differences could be considered ignorant.
Ignorant is defined as "lacking knowledge, information, or awareness about a particular thing"
Hence those that aren't familiar with such differences could be considered ignorant.
It completely escapes my mind me how the word “stupid” applies to sharing information on 32 bit words being truncated to 24 bits.
But in fact when this applies to your situation, you will compare a dithered 24 bit to a non dithered 24 bit truncated file, so there is a technical difference between the two that might possibly be audible.
Hans
But in fact when this applies to your situation, you will compare a dithered 24 bit to a non dithered 24 bit truncated file, so there is a technical difference between the two that might possibly be audible.
Hans
One think for sure is my ignorance on this subject... I'm using WaveIO-PCM1794. WaveIO can do 384 kHz/32bit but PCM1794 can do only up to 192 kHz/24bit. So I had set Foobar output to 24bit and forgot about that. Today I went back and set it to 32bit and guess what, it plays just fine. Earlier I had verified that it plays sampling rates higher than 192 kHz too. I can't tell if the conversion is happening at the WaveIO or the PCM1794.
I need to add that I use the digital volume control in my system via the keyboard. When I'm making a change, the only slider that moves is at the WaveIO control panel, not the Windows nor Foobar's. I can't tell if this could make any difference between 24 and 32bit either.
I need to add that I use the digital volume control in my system via the keyboard. When I'm making a change, the only slider that moves is at the WaveIO control panel, not the Windows nor Foobar's. I can't tell if this could make any difference between 24 and 32bit either.
Hi Kostas,
I’m almost sure we have the same situation while having the same Dac.
That means that when playing 32 bits from Foobar it will be received by the 1794 but truncated to 24 bits.
The other thing you mention about playing 384Khz was confirmed also by Doede. When having the 1794 configured as NOS Dac, the switched off 8* upsampler gives room for higher sampling frequencies.
This 384Khz will thus not be possible in OS mode.
Hans
I’m almost sure we have the same situation while having the same Dac.
That means that when playing 32 bits from Foobar it will be received by the 1794 but truncated to 24 bits.
The other thing you mention about playing 384Khz was confirmed also by Doede. When having the 1794 configured as NOS Dac, the switched off 8* upsampler gives room for higher sampling frequencies.
This 384Khz will thus not be possible in OS mode.
Hans
Hi Hans,
As I don't have any real music at 384 kHz, I'm generating test files with audacity. But they are recognized as such from Foobar and the DAC. And they are reproduced both in NOS and OS @ 384 kHz. I don't know how...
Kostas
As I don't have any real music at 384 kHz, I'm generating test files with audacity. But they are recognized as such from Foobar and the DAC. And they are reproduced both in NOS and OS @ 384 kHz. I don't know how...
Kostas
The limiting factor is the system clock. When you use BCK as SCK in the no filter mode, there is lots of headroom, as the max SCK is like 73MHz
Hi, Kostas,
A typical transmitted 2-channel digital audio data frame is serial, and consists of a total of 64-bits per 2-channel data frame. That's 32-bits (called a, half-frame) for each of the two channels. Together, the 64-bit frame represents a single matching pair of stereo samples, transmitted serially one after the other in a time-multiplexed fashion. For example, serially, 1Left/1Right, 2Left/2Right, 3Left/3Right, etc. The three samples above represent a total of 3X64-bits = 192 serial audio bits.
This 64-bit frame per 2-channel sample pair format is the same for 16-bit CD samples, as it is for up to 32-bit high-rate samples. When the received sample data resolution exceeds that of a given DAC chip's ability to quantize, the extraneous bits are either dithered (the proper way) to fit the quantizer's capability, or are simply truncated (ignored) to fit. So, for example, should the maximum 32-bit sample be sent to a 16-bit capable DAC, the sample will typically be accepted by the DAC, but will be reduced to 16-bits either by dithering or by truncation before arriving at the DAC chip's 16-bit quantizer unit.
A typical transmitted 2-channel digital audio data frame is serial, and consists of a total of 64-bits per 2-channel data frame. That's 32-bits (called a, half-frame) for each of the two channels. Together, the 64-bit frame represents a single matching pair of stereo samples, transmitted serially one after the other in a time-multiplexed fashion. For example, serially, 1Left/1Right, 2Left/2Right, 3Left/3Right, etc. The three samples above represent a total of 3X64-bits = 192 serial audio bits.
This 64-bit frame per 2-channel sample pair format is the same for 16-bit CD samples, as it is for up to 32-bit high-rate samples. When the received sample data resolution exceeds that of a given DAC chip's ability to quantize, the extraneous bits are either dithered (the proper way) to fit the quantizer's capability, or are simply truncated (ignored) to fit. So, for example, should the maximum 32-bit sample be sent to a 16-bit capable DAC, the sample will typically be accepted by the DAC, but will be reduced to 16-bits either by dithering or by truncation before arriving at the DAC chip's 16-bit quantizer unit.