How to test for bit perfection?

[bex]

This is a really n00by question.......

i have just bought an av710 and posted info on another thread. its nice to do tweaking again.

anyway, i did all the tweaks and now would like to know how to test for bit perfection.

i downloaded some random dts.wav file, and ran in fubar but got static!?!

what can i do?

 

[Cyrilix]

Some random dts.wav file? How about you play a wav file that you actually know?

 

[clarke68]

Two ways that I know of:

1. This way, where you hook your soundcard to an a/v receiver that decodes DTS and play a DTS file. If you hear music, it's bit perfect. If you hear static, it's not.
2. Or this way, where you download a file, play it, and listen for artifacts. Be careful not to turn this one up too loud, it has lots of HF stuff you can't hear, but can still damage your hearing (plenty of warnings once you click on the link)

 

[sgrossklass]

If DTS via the digital out (!) doesn't play correctly, check the settings again...
Normal 2-channel (non high sampling) mode selected?
Digital out is enabled and set to PCM only?
Automatic sample rate is enabled?

 

[bex]

i played a dts_44k_diatonis_soal.wav(can find on google) and got static but this is with my normal ext dac.

is this not bit perfect?

i cant see dts on it, but i see dolby digital on it. also i did the 96k slow mode 24bit padded to 32bit,pcm only set and automatic sample rate done, what am i doing wrong, i hate being a n00b....

 

[bex]

Quote:

Originally Posted by bex i played a dts_44k_diatonis_soal.wav(can find on google) and got static but this is with my normal ext dac. is this not bit perfect? i cant see dts on it, but i see dolby digital on it. also i did the 96k slow mode 24bit padded to 32bit, what am i doing wrong, i hate being a n00b....

clarke68, the second test ou mentioned, i played the clip and it sounds ok, i can hear numbers being dialed or something and some high pitch stuff in the middle. other than that, what am i supposed to hear?

with the first test, i used the last link, and the very last file at the bottom that is 14mb. i get static in fubar, but wmp gets fatal error and shuts down. (wmp has ks plugin)

i just dont know, sorry for all the questions

im just unhappy i cant get good sound.

Quote:

Originally Posted by sgrossklass If DTS via the digital out (!) doesn't play correctly, check the settings again... Normal 2-channel (non high sampling) mode selected? Digital out is enabled and set to PCM only? Automatic sample rate is enabled?

all is the right setting

 

[bex]

funny thing i found out, on the drivers vinyl control panel, "sample rate state" it says "current = 48k" automatic box is checked.

however, on wmp Kernel streaming plugin, it says 44100hz sample rate, and 16bit,

which should i trust?, foobar doesnt tell me much, on the bottom bar, when playing music it says, PCM 1411kbps 44100 hz stereo.

 

[sgrossklass]

If in doubt, the card's control panel should know the actual sample rate. That you can hear the dial tones (which in fact are aliasing artifacts) proves that the card indeed is running at 48 kHz rather than 44.1. The question that remains is why this would be. (Since you apparently don't have a device capable of DTS decoding, a test with such a file is useless.)
Which driver version are you using? Does QSound happen to be enabled? Also, try out ASIO4All with the ASIO output plugins for Foobar2k and Winamp. If ASIO4All reports a "beyond logic" error, then the sample rate is still locked for some reason.

 

[bex]

I just removed the v310 drivers i was using and changed to v473b

and now it says 44.1k in the current state.

and when playing, with all the dsps enabled, i get just the dial tone sounds and nothing else, the two wav, and dts files i mention now have no sound instead of static, all other music seem fine.

with all this it still says 44.1k in the driver control panel.

could have have solved my problem? so all i need is a dts dac ?

if so THANK YOU SOOO MUCH

also, what is qsound?

 

[sgrossklass]

Seems I should have been reading the udial.wav instructions again - apparently getting the dial tones is normal, but there shouldn't be anything else, like siren sounds. So it looks like the 4.73b driver has solved this problem.

QSound is the software engine used for rendering 3D sound (EAX2 etc.). I think 3D sound on the AV-710 requires that you use an old driver implementing the Sensaura library and you won't get DSound3D support with newer drivers employing QSound. But hey, the card is more interesting for plain music playback anyway...

 

[bex]

Quote:

Originally Posted by sgrossklass Seems I should have been reading the udial.wav instructions again - apparently getting the dial tones is normal, but there shouldn't be anything else, like siren sounds. So it looks like the 4.73b driver has solved this problem. QSound is the software engine used for rendering 3D sound (EAX2 etc.). I think 3D sound on the AV-710 requires that you use an old driver implementing the Sensaura library and you won't get DSound3D support with newer drivers employing QSound. But hey, the card is more interesting for plain music playback anyway...

yeah, the drivers seem to be the trick. it automatically chooses 48 or 44.1k

some of my music is encoded in 48!?!?!? its weird. i guess im getting good sound. its ok, cs is only part time for me in terms of gamin, i just need to setup my analog rear out for my new pair of Audio Technica ATH AD700.

my first expensive cans, AUD230, or USD160, i know i paid alot more, but its australia, lack of stuff. i saw it in us for USD100

bl for me.

anyway, my next pair will be ms-1 alessandro.

but probalby need an amp first with this card, or maybe it might work on the ext dac?

 

[bex]

THANKS AGAIN MATE for all the help

one more thing, with fubar, i cant seem to run without at least some of the DSPs, i wonder why this is?

 

[Andrew_WOT]

dts file method is used for testing bit-perect *digital* output by spdifing PCM to the receiver with DTS decoding capabilities.
Udial method is used to test the quality of the 44.1 to 48 conversion of the DAC. (can be mistaken but they all upsample to 48khz internally, please correct me if I am). If you hear sirens or high pitched noises then this is a problem. Make sure that no upsampling is used in foobar when you do the testing.

 

[sgrossklass]

Quote:

Originally Posted by bex one more thing, with fubar, i cant seem to run without at least some of the DSPs, i wonder why this is?

That shouldn't be. I still have 0.8.3 here, but using no DSPs at all is no problem.

As for whether you need an amp, if you external DAC (model?) has a headphone out of decent quality you may be able to do without. The 7/8 out on the AV-710 basically is not suited for cans.

 

[sgrossklass]

Quote:

Originally Posted by Andrew_WOT Udial method is used to test the quality of the 44.1 to 48 conversion of the DAC. (can be mistaken but they all upsample to 48khz internally, please correct me if I am).

Not quite. Basically you can detect whether there's resampling of crappy (anti-aliasing lowpass) quality going on somewhere in the signal path. (The resampling done by a CS46xx already cannot be detected this way IIRC.) Typical targets are EMU10k(2) based SB cards (working at 48 kHz from DSP to DAC/ADC) with their resampling tuned for speed rather than quality. Why one would want a sound card to resample everything has one reason: mixing of multiple audio streams with different sample rates. You get along with only one master clock crystal, too. "Decent" sound cards allow being run at a multitude of different sample rates (the ProDigy 7.1 driver allows 16, 22.05, 24, 32, 44.1, 48, 88.2, 96, 176.4 and 192 kHz, for example), which includes the D/A and A/D converters. This requires two main clock crystals (typically 22.5792 MHz for 44.1 kHz and such and 24.576 MHz for 48 kHz and such, as you cannot generate all sample rates from just one crystal using clock division - this would work with a PLL, but this always is a potential jitter source), and the driver has to do some kind of sample rate management if audio streams of different rates are to be mixed; usually the one with the highest sample rate wins. With a single stream, an audio stream can be spit out as-is. The upside to this is that you don't have to throw massive amounts of computing power at resampling (this is not at all trivial to do in high quality, the critical aspect being the anti-alias lowpass filtering).