[korben_dallas]

Gordon,

Have you achieved bit-perfect playback with Vista? If so, what was your setup?

 

[sangel]

Hello Elias,

Thank you very much.

Best regards

sangel

 

[Wavelength]

Elias,

I talked to Apple yesterday and this morning had a conference call with the engineers.

They were not aware of this problem and it had not been seen in their problem reports.

They are going to try and reproduce it and if not I can send them out my DAC1 for them to see what the problem is.

I did send them your enumeration as they said it maybe linked to the way you handle something.

Thanks
Gordon

 

[Wavelength]

Quote:

Originally Posted by korben_dallas /img/forum/go_quote.gif Gordon, Have you achieved bit-perfect playback with Vista? If so, what was your setup?

KD,

Actually as Thomas has indicated in several posts now this is easy to do. Much easier and more assured than XP ever would and also applications where kernel streaming and ASIO were not supported via Direct Sound (iTunes, WMP, etc) can achieve this.

It does require a 24 bit dac and setting that option in the device panel. At that point 16 bit audio will cruise through uninterrupted.

I tried this out with J River, iTunes and Media Player. Results were much better.

Thanks
Gordon

 

[Wavelength]

Gang, Elias;

I had a long banter with Apple yesterday in regards to the 7.5 stuff.

Ok here it is and this has been the same for a long while:

Quote:

I don't believe there have been any changes in this regard, but the driver will try to select formats in the following order the first time a device is attached to a particular port on a system: 1. 2-channel 16-bit @ 44.1kHz 2. 1-channel 16-bit @ 44.1kHz 3. 2-channel 16-bit @ highest supported sample rate 4. 1-channel 16-bit @ highest supported sample rate 5. alternate setting 1 @ highest supported sample rate If the device doesn't do 16 bit, the correct behavior is probably 24-bit at the highest supported sample rate. After changing the sample rate, channel count or bit depth, the Mac should remember the last setting if you plug it back into the same USB port. One last thing: 10.4.11 and Leopard have the same version of the USB audio driver.

So really nothing has changed here and why this came up seems to be a question.

Elias, the guys in the sound group want to know who you are talking too as it may not be the correct people.

Thanks
Gordon

 

[thomaspf]

Quote:

Originally Posted by Wavelength /img/forum/go_quote.gif KD, Actually as Thomas has indicated in several posts now this is easy to do. Much easier and more assured than XP ever would and also applications where kernel streaming and ASIO were not supported via Direct Sound (iTunes, WMP, etc) can achieve this. It does require a 24 bit dac and setting that option in the device panel. At that point 16 bit audio will cruise through uninterrupted. I tried this out with J River, iTunes and Media Player. Results were much better. Thanks Gordon

And just to be extra clear about this to avoid any more drawn out discussions with the marketing folks from Benchmark. This will give you a 16 bit stream embedded in the top 16 bits of a 24 bit stream. If you use a true 24 bit stream the lower bits still get modified. Vista is not truly bit transparent except that you can pass through a 16 bit stream inside a 24 bit stream as described.

Cheers

Thomas

 

[FRANKe]

Quote:

Originally Posted by thomaspf /img/forum/go_quote.gif Vista is not truly bit transparent except that you can pass through a 16 bit stream inside a 24 bit stream as described. Thomas

Thomas, is this a general "blanket" statement, or are you specifically referring to the Benchmark DAC and the use of "shared mode" (with iTunes & WMP)?

With Vista's "exclusive mode" is it not "truly bit transparent"?

FWIW, I believe XXHighEnd is the only media player right now utilizing "exclusive mode".
XXHighEnd - Index

It would be interesting to see the results of these bit transparent tests done with XXHighEnd. And BTW, it has been reported that XXHighEnd will not lock onto "exclusive mode" with the Benchmark DAC presumably because the DAC is "reporting" as 24bit and at the moment XXHighEnd does not output 24bit.
Benchmark DAC1 USB

 

[tuffgong]

Hello again, just some questions for the jury before I go. I am considering buying a dac1 usb as well for dj purposes, I love the dac-1 classic alot. The DJ program that seems to be the best for me is 'disco xt'. It automixes flac files, and the company emailed me back and said the program uses directsound, so I think that means with vista, running at 24 bit/48hz, with volume at 100% on the dj program, I should have the best quality possible considering the automix feature? I also think i'm going to use my old platinum audio solo's to take on the road, and I was going to bi-amp each speaker with a pair of bryston powerpac 120 sst's, kind of a poor man's aml1

My amateur question is how can I split the xlr signal coming out the back of the dac1 so I can get four signals to amplify without messing up the line levels like a normal guitar center splitter. Thanks in advance.

 

[thomaspf]

Quote:

Originally Posted by FRANKe /img/forum/go_quote.gif Thomas, is this a general "blanket" statement, or are you specifically referring to the Benchmark DAC and the use of "shared mode" (with iTunes & WMP)? With Vista's "exclusive mode" is it not "truly bit transparent"? FWIW, I believe XXHighEnd is the only media player right now utilizing "exclusive mode". XXHighEnd - Index It would be interesting to see the results of these bit transparent tests done with XXHighEnd. And BTW, it has been reported that XXHighEnd will not lock onto "exclusive mode" with the Benchmark DAC presumably because the DAC is "reporting" as 24bit and at the moment XXHighEnd does not output 24bit. Benchmark DAC1 USB

No blanket statements from me in general...

Benchmark has been claiming for a while now that by changing the USB firmware in their DAC, they can get the USB audio driver in Windows to produce bit perfect results for applications like foobar as described on their WEB site. I have tried to correct that statement from the beginning since there is nothing any USB device can do to achieve that with applications using WAV or DirectSound APIs using the usbaudio driver in Windows.

This has always worked with kernel streaming. The wasapi in Vista is a new option for a direct sound API and the first commercial application I have seen a while back that supported exclusive wasapi was ntrack but that is not a conventional software player. I am pretty sure if your application supports wasapi in exclusive mode then this should work but I have not tested it.


Cheers

Thomas

 

[EliasGwinn]

Quote:

Originally Posted by Wavelength /img/forum/go_quote.gif Gang, Elias; I had a long banter with Apple yesterday in regards to the 7.5 stuff... "I don't believe there have been any changes in this regard, but the driver will try to select formats in the following order the first time a device is attached to a particular port on a system: 1. 2-channel 16-bit @ 44.1kHz 2. 1-channel 16-bit @ 44.1kHz 3. 2-channel 16-bit @ highest supported sample rate 4. 1-channel 16-bit @ highest supported sample rate 5. alternate setting 1 @ highest supported sample rate If the device doesn't do 16 bit, the correct behavior is probably 24-bit at the highest supported sample rate. After changing the sample rate, channel count or bit depth, the Mac should remember the last setting if you plug it back into the same USB port. One last thing: 10.4.11 and Leopard have the same version of the USB audio driver." ...So really nothing has changed here and why this came up seems to be a question.

Gordon,

This information correlates with what I understand about OS X 10.4. However, this information does not pertain to iTunes. It is iTunes 7 that has changed, not OS X.

iTunes 7 sets a default output sample rate, when it is launched, to match that set in AudioMIDI. After iTunes is launched, if the user changes the sample-rate in AudioMIDI, the sample rate of the final digital output will correspond with that change. In other words, if you monitor the USB activity, or the built-in digital audio output, the sample-rate will change according to the settings in AudioMIDI.

However, the important detail to understand is that the sample-rate of the (internal) digital audio, leaving iTunes and going to CoreAudio, will remain at the sample-rate that iTunes defaulted to upon launch. You will not be able to monitor this sample rate with a USB analyzer or any other external hardware. This is a difference between the two software ports inside the computer.

Here is a quote from my correspondence with Apple:

Quote:

On the Mac, the OS (i.e. CoreAudio) mixes all of the audio into one stream which it then hands to the hardware. In order for this to work, all of the audio streams must be mixed at the current sample rate of the hw. So, *somebody* has to do SRC and it can either be the app or CoreAudio doing it for you. Pre iTunes 7.0, we let CoreAudio do it. Post-7.0, we do it. One caveat is that iTunes doesn't notice if the user changes the hw sample rate *after* it launches -- It captures it at launch time and leaves it that way. It's certainly possible that these people didn't figure that out, so changing the hw rate after we launch means we'll sometimes be doing SRC *and* CoreAudio might be doing SRC at the same time. For example: 1) At launch, the hw sample rate is 48kHz. We set up our processing chain to run at 48kHz. 2) You play a 44.1kHz file. We do SRC to 48kHz and hand the data to CoreAudio who send it along to the device. 3) You change the hw sample rate to 44.1kHz while we're running. 4) You play the same 44.1kHz file. We now do SRC to 48kHz and give the data to CoreAudio who does SRC back to 44.1kHz!

.

Thanks,
Elias

 

[EliasGwinn]

Quote:

Originally Posted by thomaspf /img/forum/go_quote.gif And just to be extra clear about this to avoid any more drawn out discussions with the marketing folks from Benchmark... Cheers Thomas

Quote:

Originally Posted by thomaspf Benchmark has been claiming for a while now that by changing the USB firmware in their DAC, they can get the USB audio driver in Windows to produce bit perfect results for applications like foobar as described on their WEB site. I have tried to correct that statement from the beginning since there is nothing any USB device can do to achieve that with applications using WAV or DirectSound APIs using the usbaudio driver in Windows.

Hello Thomas,

I would like to recognize your contributions to this thread; they are appreciated. However, I would like to clarify a few things before we continue forward.

First, our marketing department (which consists of 1 person, a graphic designer) has no input and makes no claims pertaining to the performance or capabilities of our products. Those claims are solely made by the two engineers in the company: myself and John Siau, the designer of the DAC1 and most other current Benchmark products.

If you are confused and thought that I was involved in marketing, I would like to clarify that my title at Benchmark is "Applications Engineer", and I hold a degree in Electrical Engineering.

Benchmark has always had a mission to maintain objective, scientific truth and honest claims about our products' performance, and audio technology in general, in an effort to achieve within the audio community a clear understanding of audio technology. In fact, this forum is one example of our dedication to maintaining an honest discourse about our products and technology in general. The 'Audio Wiki', the 'Feedback - Informational Newsletter', and the white papers on our website are other examples.

We refuse to engage in the "Black Magic" or "Snake Oil" claims that linger like a bad smell among the audiophile community. We refuse to let marketing-spin interrupt objective fact. We pride ourselves in being as honest and transparent as possible. I know many folks in both the pro and audiophile community will attest to that, as it has become a large part of our reputation.

Now, as for XP transparency and the DAC1 USB...

I would like to clarify our claims: WE DO NOT CLAIM THAT OUR FIRMWARE MANIPULATES THE BEHAVIOR OF KMIXER or USBaudio.sys. Please do not propagate that misinformation. We claim that Windows XP can inherently operate bit transparently as long as the relevant software (app, driver and/or firmware) doesn't prevent transparent operation.

I understand that you and Gordon disagree with this claim. I respect your objections, and I would like to resolve this issue. Please understand: if Kmixer is not capable of transparent operation - we would like to know just as much as you! However, according to our tests, and communications with Microsoft, we have concluded that it is possible.

Please understand: we are not 'closing the case' on this issue. If you have proof that shows otherwise, I would like to see it and take it into consideration. I know that Thomas has claimed that he has developed his own software to check for bit-transparency. He claims that this software proves that XP is not bit-transparent. This intrigues me, and I am very interested in obtaining a copy of this software so that I can also see the results.

I also understand that Thomas and/or Gordon have also spoken with representatives at Microsoft. Perhaps the discrepancies in statements from Microsoft (and/or Apple) are occurring because the statements are represented here as 2nd hand communication. In other words, I'm sure if these representatives were speaking to each other, the discrepancies would be cleared up relatively quickly and succinctly.

To conclude, the main point I would like to stress is that I (and I'm sure many other Head-Fi folks) would not like to see slanted statements and spin in this thread. I represent Benchmark as a qualified engineer, and I do not work in marketing. All claims that I and/or Benchmark make are true to the extent of our knowledge and testing. We will respect and will try to resolve any discrepancies.

Thank you for your participation.

-Elias

 

[EliasGwinn]

Quote:

Originally Posted by thomaspf /img/forum/go_quote.gif No blanket statements from me in general...

...isn't that statement a 'blanket statement?

Just kidding...trying to lighten the mode a bit after my last saga of a post

 

[EliasGwinn]

Quote:

Originally Posted by tuffgong /img/forum/go_quote.gif ...how can I split the xlr signal coming out the back of the dac1 so I can get four signals to amplify without messing up the line levels like a normal guitar center splitter. Thanks in advance.

You can use simple 'Y' splitters, as long as the input impedances on the loads (active crossovers, I believe...) are sufficiently high (>1.5k). Also, you should be aware that if one of the devices are powered down, it may present an awkward load and distort the output. It won't damage the DAC1 USB, but the signal may become distorted.

Thanks,
Elias

 

[Wavelength]

Elias,

ThomasPF is with Microsoft! He tested your DAC1 USB at Microsoft.

~~~

Look neither of us are your enemies... I don't see you as a competitors and have been trying to help you out here as Thomas has.

The only thing we are saying as is most of the rest of the Windows team is that it is impossible to get bit perfect audio without bypassing the KMIXER.

That is an easy thing to do and I am sure your insrtuctions for doing that will make it capable of bit perfect resolution.

If you follow Thomas's instructions for Vista there too you will get audio nirvana.

~~~~~~~~~

You may want to pick up a Prism dScope III for testing then you can easily see what is and is not working. Since it has native mode it can stream kernel, DirectSound and via drivers and you can quickly see what's bit perfect or not.

~~~~~~~~~

Heck next time we are in your area I say let's have some dinner and drinks at my fav restaurant the Dinosuar (my wife is from Manilus).

Thanks
Gordon

 

[EliasGwinn]

Quote:

Originally Posted by Wavelength /img/forum/go_quote.gif Elias, ThomasPF is with Microsoft! He tested your DAC1 USB at Microsoft. ~~~ Look neither of us are your enemies... I don't see you as a competitors and have been trying to help you out here as Thomas has. The only thing we are saying as is most of the rest of the Windows team is that it is impossible to get bit perfect audio without bypassing the KMIXER. That is an easy thing to do and I am sure your insrtuctions for doing that will make it capable of bit perfect resolution. If you follow Thomas's instructions for Vista there too you will get audio nirvana. ~~~~~~~~~ You may want to pick up a Prism dScope III for testing then you can easily see what is and is not working. Since it has native mode it can stream kernel, DirectSound and via drivers and you can quickly see what's bit perfect or not. ~~~~~~~~~ Heck next time we are in your area I say let's have some dinner and drinks at my fav restaurant the Dinosuar (my wife is from Manilus). Thanks Gordon

Gordon,

Thank you for this kind note. And don't worry, I was not feeling attacked or insulted. I simply wanted to make sure this thread did not stray from the facts. Sometimes mis-quoting or mis-stating someone, even if unintentionally, can drive the conversation to undesirable places. But I feel we are all working towards the same goals, and I really do appreciate your correspondence.

Thomas and I are corresponding with Microsoft engineers to come to a consensus on this topic of bit-transparency through kmixer. Hopefully, we will be able to provide a resolution to the audio community once and for all.

I would be happy to join you for an evening at the Dino. Please let me know when you are in town next.

Thanks,
Elias