Did I just got bit-perfect output from Audigy NX?

[Sycraft]

Quote:

Originally Posted by AdamWill maybe I should rephrase and say that with the correct configuration you should be able to pass through DTS / DD correctly on _anything_ with a digital output: being able to do this does not imply that you are able to pass through a raw PCM stream at the correct sampling rate, that's not the same kettle of fish at all.

Except that's what DTS is. DTS is just encoded in a PCM stream. If you feed a DTS signal in to a non DTS decoder, you get static (same with Dolby Digital). It's a valid PCM stream, just not audio data. That's how DTS CDs work. You encode a DTS file at CD rate, which is slightly different form DVD full rate (full rate DVD is 1.536mb/s half rate is 768kb/s which is by far the most common) at 1.234mb/s. It is then just output as PCM form the CD player to the reciever. The reciver recognises it as valid DTS (if it knows aobut DTS that is) and decodes it.

So if you get DTS out digitally form a card, that's an unmodified PCM stream right there. Any modification, including resampling, would hose the signal.

However, I think DVD DTS is output at 48kHz, which is what the full rate is different, and I think the cards know to not use their DSPs when it's a DTS signal they are passing.

 

[maarek99]

There were some hacked drivers floating around that could give bitperfect out to certain audigy models (not 2 zs). So it's not impossible.

 

[maarek99]

Quote:

Originally Posted by Sycraft However, I think DVD DTS is output at 48kHz, which is what the full rate is different, and I think the cards know to not use their DSPs when it's a DTS signal they are passing.

It can't be 48 khz because cd's are 44khz. And there are dts cd's.

 

[Sycraft]

Quote:

Originally Posted by maarek99 It can't be 48 khz because cd's are 44khz. And there are dts cd's.

That isn't relivant. DTS CDs are 44.1kHz, no question. However that doesn't mean DTS DVDs are. I don't have a copy of the DTS spec, but it would be perfectly possible for it to be specified to use two different sample rates as transport. In fact I'm almost certian DTS DVDs are 48kHz as there's no way you can fit full rate DTS on a CD. CDs are 1.411mb/s (using the SI definition of mega meaning 10^6 not the computer definition of 2^20). That's 44,100 samples by 2 channels by 16 bits. As I noted, full rate DTS is 1.536mb/s which is not conincedentally the value given by 48,000 * 2 * 16.

If you like I can check the sample rates of the wave files generated by my DTS encoder in the different modes, but I'm pretty sure I'm correct, DTS CDs are 44.1kHz, DTS DVDs are 48kHz.

 

[maarek99]

The card should re-quantizize anyway so the signal should be screwed if there is any resampling going on. Many people think that if you output 48khz with a Live or Audigy it won't resample. That is wrong. It will still re-quantize it.

 

[Eagle_Driver]

Quote:

Originally Posted by maarek99 The card should re-quantizize anyway so the signal should be screwed if there is any resampling going on. Many people think that if you output 48khz with a Live or Audigy it won't resample. That is wrong. It will still re-quantize it.

Yep. The only way it won't re-quantize at 48kHz is if the audio file is already at the native quantization rate for the DSP. That means on an Audigy, if you feed the Audigy 48kHz/16bit audio, it will be re-quantized to 48kHz/32bit. And then, even if the card does not resample/re-quantize on its own, there is the Windows Kmixer coming into play: By default, Creative's drivers send everything to the Windows Kmixer (notorious for resampling audio to the same sampling rate as the original, but changing the bits as it resamples).

 

[jiiteepee]

Quote:

Originally Posted by Eagle_Driver Yep. The only way it won't re-quantize at 48kHz is if the audio file is already at the native quantization rate for the DSP. That means on an Audigy, if you feed the Audigy 48kHz/16bit audio, it will be re-quantized to 48kHz/32bit. And then, even if the card does not resample/re-quantize on its own, there is the Windows Kmixer coming into play: By default, Creative's drivers send everything to the Windows Kmixer (notorious for resampling audio to the same sampling rate as the original, but changing the bits as it resamples).

Do you mean it does this with ASIO or WDM/KS too ??

jiitee

 

[needgoodphones]

Quote:

Originally Posted by Sycraft Except that's what DTS is. DTS is just encoded in a PCM stream. If you feed a DTS signal in to a non DTS decoder, you get static (same with Dolby Digital). It's a valid PCM stream, just not audio data. That's how DTS CDs work. You encode a DTS file at CD rate, which is slightly different form DVD full rate (full rate DVD is 1.536mb/s half rate is 768kb/s which is by far the most common) at 1.234mb/s. It is then just output as PCM form the CD player to the reciever. The reciver recognises it as valid DTS (if it knows aobut DTS that is) and decodes it. So if you get DTS out digitally form a card, that's an unmodified PCM stream right there. Any modification, including resampling, would hose the signal. However, I think DVD DTS is output at 48kHz, which is what the full rate is different, and I think the cards know to not use their DSPs when it's a DTS signal they are passing.

DTS is not at all encoded in a PCM streamhttp://www.dts.com/media/uploads/pdfs/whitepaper.pdf. It is a compression algorithm simmilar to mp3 (except with much better psychoacoustics), which means that it will not be bit perfect, just like most soundcards aren't bit perfect. DTS may sound better or worse than resampling, I guess I'll just keep listening to my 192 kbps mp3s and not worry about resampling error. If you think that DTS is a PCM stream, then you don't understand what pulse channel modulation is.

 

[chris719]

I think you don't understand what Pulse CODE Modulation is. In any case, a DTS encoded file in a .wav container will only arrive at the decoder intact if the transmission is bit perfect. It is pretty clear you have no idea how any of this works if you think that MP3's or any other kind of compressed data is not subject to the same processing as uncompressed PCM. Bit perfect transmission is the topic here, not bit perfect (i.e. lossless) reproduction of data that has been encoded. There is no device on earth that can create audio directly from DTS or MP3, it must be decoded to PCM first. Resampling has absolutely nothing to do with any of what you said either... comparing resampling to DTS is like comparing filters in photoshop to jpeg files. If DTS is not encoded in a PCM stream, then please I beg you to explain to me how it gets through the SPDIF tranceiver ICs, through the I2S bus and into a decoder, when it is the case that SPDIF and I2S only transport serialized PCM.

 

[thomaspf]

Maybe we can accelerate this discussion by introducing HDCD as another example of an encoded PCM stream.

Only if all the bits get delivered intact a suitable DAC with an HDCD decoder will be able to even detect that the data is actually carrying an HDCD encoding.

Cheers

Thomas

 

[Andrew_WOT]

I am evaluating option on using Audigy 2NX as USB to SPDIF converter. It uses asyncronous USB mode which reduces jitter.
How can I check if the digital output is not get upsampled? Or if anyone has actuall hard proves that it get passed through intact.

Thanks

 

[tschanrm]

Andrew,

This sure is an old thread. Anyway, I would advise against buying the Audigy 2NX if you think Asynch mode reduces jitter in windows. Asynchronous audio mode in all current versions of windows (95,98,ME,2000,XP,Etc.) is broken, if it uses USBaudio.sys. With the release of Vista, Microsoft plans to fix the asynch audio mode bug. To prove it, here is a quote from "Windows Platform Design Notes, Draft 0.3, 2003:"

Quote:

Starting with Windows 98, Usbaudio.sys supported the adaptive and synchronous endpoints, but it did not implement the asynchronous endpoint correctly. Full support for asynchronous endpoints in Usbaudio.sys is planned for Windows Longhorn.

Now, the one thing I do not know is if the Audigy 2NX has to use usbaudio.sys as part of it's driver structure. If it does, that's bad. If Creative somehow designed a custom usb driver on top of the NX driver, then asynch mode would be fine. But considering Creative's track record of poor drivers, I'm guessing they use USBaudio.sys and don't care about the consequences.

And to check if a digital output is not getting resampled, try and play a 44.1khz DTS file (go here for 44.1 DTS/AC3: http://www.kellyindustries.com/sounds.html ). Other people said to try this earlier in the thread, then other's tried to discredit it, but it does work. Think of it this way: When your soundcard sends DTS/AC3 streams, the S/pdif is like a USB cable from your computer to your home theater receiver. Just like a USB hard drive transfer, if the file gets altered in anyway it becomes corrupt. So if the DTS/AC3 stream gets altered in anyway, it also becomes corrupt. Once a DTS/AC3 file is corrupt, all you hear is pink noise and/or nothing.

EDIT: Ok, just found this piece of information, taken from Microsoft's Universal Audio Architecture Guidlines, October, 2005:

Quote:

In Windows Server 2003 and earlier versions of Windows, usbaudio.sys supports adaptive and synchronous endpoints, but not asynchronous endpoints.

So the Audigy 2 NX uses custom drivers, since USBaudio.sys does not even support asynch audio mode. I always though asynch audio mode was supported, just flawed. So the NX is a good choice if you want to use it as a straight usb to s/pdif box.

 

[Andrew_WOT]

Accordingly to this
http://www.audioasylum.com/forums/pc...ages/3200.html
Audigy 2NX does not use USB audio chip but rather generic Philips, designed for things like disk drives. So I doubt it utilizes usbaudio.sys for communication with computer bus.

 

[tschanrm]

Andrew, read above, I edited my post while you were replying....

But as a side note, the fact that it uses a Philips generic USB chip does not mean it could not use the USBaudio.sys. A good example is the Yamaha DP-U50. It uses an Agere USS-820DT USB chip, a generic device controller, with the USBaudio.sys driver.

 

[Andrew_WOT]

Thanks, for the link. I'll make sure to post the results of the test.