Hi.

Finally I got my Terratec Aureon 5.1 Sky.

I haven't tested much yet, but so far I like it a lot. In headphone mode it is capable of driving my Sennheiser HD 540 (300 Ohms) much louder than I'll ever want, no audible background noise, udial.wav sample sounds perfect without resampling ...

BUT: Some system sounds are very loud and terribly distorted, like extreme clipping. I found out that it seems only to happen on .wav files @ 11025Hz sampling rate. If I play the same files in fb2k and resample to higher sampling rate, everything is fine.

I have Win2KSP4 and the newest drivers from Terratec Homepage installed. Changing Windows' hardware sampling rate conversion quality settings doesn't change anything.

Does anyone have similar experiences or maybe even a sollution?

(Of course I could replace all 11025Hz .wav files with manually resampled versions, but I'd rather find a real sollution than a workarround).

Thanks in advance for any comments.

Cheers tigre

Because the Envy24 chip or newer versions natively support only samplerates higher than 22050Hz or 32kHz (can't remember now). The Sky uses such a chip.

PS. Set your wave slider to mute and listen through FB2k using ASIO to avoid ear-blasting events (or tweeter-blasting if you use speakers).

ASIO drivers are not available for the Aureon Sky series ...

Thanks for your answers so far.

From what I've tested, 16000 Hz and higher works, 11025 and lower doesn't.

There an "ASIO" button in control pannel, If I press it I can adjust buffer size, but I haven't found out so far if/how fb2k ASIO output works with this card. Probably it doesn't.

Does this mean all Envy24 based cards do have these problems - or are some cards/drivers that use some kind of workarround?

As "hotfix" I've used fb2k to find and resample all .wav files with 11025Hz sampling rate and lower.

working playback frequencies should then be 16000, 22050, 24000, 32000, 44100, 48000, 64000, 88200, 96000, 128000, 176400, 192000 (perhaps this depends on the master clock pll and possible divisors) ... and the terratec info page does not state anything about ASIO capability ...

Try the IP-voice switch in the "Digital" tab. This toggles a resampling mode (allowing playback of different sample rates at the same time).

ASIO works here with Aureon Sky. Both in Winamp and Foobar.

FB-plugin: http://www3.cypress.ne.jp/otachan/

(Make sure you selected the correct output data format in the Playback options, e.g. 24bit padded to 32.)

Continuum, you're da MAN! IP-voice switch in the "Digital" tab worked. I replaced all resampled system sounds by the originals again and everything sounds perfect so far. Thanks alot!

Now that you've mentioned ASIO output: What are the advantages/disadvantages of using it?

Actually, I tried it for the first time just an hour ago. As lucpes said you can mute the wave slider, which is especially nice when you just play music in the background.

OTOH it seems to be more susceptible to CPU-load, blockades other sound and doesn't have the nice crossfading features of DSv2.

The two advantages of ASIO are that it avoids kmixer and allows for bit-perfect output in WinXP and Win2K (similar to kernel streaming in some cards), and that it makes possible to get very low latency input and output. The later is useful only for musicians, not for just playing back music.

So this means waveout/directsound are not bit perfect in Win2K. If I use fb2k with 24 bit output resolution, isn't this a non-issue anyway?

Most probably it isn't in your case. With M-Audio cards, for example, kmixer only changes last bit (16th or 24th bit). On other cards, it may perform some actual resampling of the data, that would degrade the signal somewhat more. This depends on the driver implementation, but the first behaviour should be the right one, on non-resampling cards such as yours.

This can be checked running an asychonous RMAA test, both using ASIO or KS and directsound. If the dynamic range graph shows obviously higher low-level distortion on the directsound test, then kmixer is actually resampling. Not that it would be very audible in practice, but, as I said, should not happen in non-resampling cards such as yours.

Some cards with special non-standard WDM drivers such as RME cards are bit-perfect in WinXP or Win2K even when using directsound output.

Continuum, do you have a link to the Winamp ASIO plug-in? I've been dying to try ASIO with my Echo sound card since I got it, but have yet to find a W/A plug-in to support it. I know my soundcard supports it (based on the documentation, anyway).

winamp 2/3/5 plugin: http://www3.cypress.ne.jp/otachan/

Thanks!

foobar2000's kernel streaming mode bypasses kmixer as well.

If this tab enables resampling, udial.wav should not sound correct anymore, now.

Can the WinAmp 5 ASIO PlugIn be used for WinAmp 2.9x as well ?

Yep. (but NOT Winamp3)

I can choose the sampling frequency and as I've set it to 44100Hz udial sounds alright. (If I set Windows' sampling rate conversion quality to high it sounds still OK if (re)sampling rate is set to 48kHz.)

Thanks for that hint anyway, I think I'll look into ASIO because of that (sometimes I play back 48 or 96kHz sampled files with fb2k and I don't want to change sampling rate in control panel everytime.)

It must be for small sample rates only. For instance, M-Audio Revolution too resamples everything under 32kHz to this value.

No. I've just tested this and other things using fb2k (v 0.7.3a + asio output plugin), udial.wav and a 10 -> 20kHz sine sweep. Windows' sampling rate conversion quality was set to "low". Results:

Direct sound (normal as well as 2.0):

"IP Voice" enabled -> Downsampling to Control Panel's sampling rate is performed (Everything > 1/2*fs is removed); awful aliasing is audible

"IP Voice" disabled -> Everything is normal, but the "system sounds" problem reappears.


Wave out:

"IP Voice" enabled -> Downsampling to Control Panel's sampling rate is performed (Everything > 1/2*fs is removed); aliasing hardly audible

"IP Voice" disabled -> Everything is normal, but the "system sounds" problem reappears.


Kernel streaming:

"IP Voice" enabled -> Playback only works if Control Panel's sampling rate is identical to fb2k's output sampling rate, otherwise fb2k reports an error.

"IP Voice" disabled -> Everything is normal, but the "system sounds" problem reappears.


ASIO:

"IP Voice" enabled -> If Control Panels's sampling rate is different from fb2k's, playback pitch and speed is changed!

"IP Voice" disabled -> Everything is normal, but the "system sounds" problem reappears.


It appears to me more and more like replacing system beeps by upsampled versions is the best compromise.

HA ! HA ! HA ! That's what I call bit exact playback !

OT : same thing here when I slave the Marian Marc to a 48000 Hz digital input signal and ask SoundForge to record at 44100 Hz