sgrossklass
Headphoneus Supremus
- Joined
- Dec 3, 2004
- Posts
- 2,803
- Likes
- 22
Got my (used) Live! 24-bit today and promptly dropped it into my secondary comp, not without noticing that the card is in fact smaller than I had imagined. (Luckily I bought the 6+1 P3B-F back then, so after ripping out the old AWE64 Value I had a free PCI/ISA combo slot.) Install under Win2k was a breeze with the supplied CD. (This, however, promptly reminded me that I should do something about this $&!(/% jet engine of a CD-ROM drive.) ASIO4All shows 2 ins and 8 outs with up to 96 kHz. I chose to install some of the supplied software as well, but all you /really/ need is the "device control" control panel applet and the manual. The rest is nice to have (like the "surround mixer" mixer app) but not really needed.
This card operates in one of four fixed output modes selectable by choosing the sample rate of either 48 or 96 kHz and the bit depth of either 16 or 24 bits in the "device control" applet. Make sure the settings match the output of the resampler used. (And of course, record in 48 and not 44.1 kHz.)
Using RMAA, loopback measurements were taken with 48 and 96 kHz sample rate and 24 bits. Best results were obtained in 48 kHz, with a THD of 0.0015% and IMD+N of 0.0063%; dynamic range came out at 93.7 dB. Stereo crosstalk levels proved to be excellent, with up to 93 dB and still more than 80 dB at 20 kHz. The frequency response is pretty much dead flat, less than .3 dB down at 20 Hz and less than .1 dB at 20 kHz. As seen in the iXBT review, there's a bunch of power supply harmonics and such crap in the spectrum, but all of the spikes are below -100 dB. (Still, power filtering could be better.) The ADC adds a whole zoo of high-order harmonics, but most of those are around or below -110 dB. At 96 kHz, IMD+N rises to 0.015% (same for sweeping IMD, just higher throughout the spectrum), and you can see a noticeable increase in the noise floor above 30 kHz (noise shaping?). Stereo separation still is better than 75 dB at 40 kHz, the -0.5 dB points of the frequency response are at ~23 Hz and ~40 kHz, respectively. I think I'll upload some RMAA reports later.
For listening tests, I used Winamp 5.08something with the ASIO plugin (resampling to 48 kHz, quality "high") feeding ASIO4All 2.6 (buffer 256 samples, hardware acceleration with 10 ms latency); the Shibatch Super EQ plugin with my usual HD590 compensation setting was used (posted this in the Headphones forum a while back).
The ASIO channels -> output correlation is a bit strange, a channel offset of 2 didn't output the signal on line out 2 but rather a mono downmix on the first output.
Plugging my HD590 directly into the card was no problem. I had to turn down the volume pretty far, the sound itself was quite listenable (not as peaky in the highs are directly out of the Aureon). Audible background noise? Nah. Where is it supposed to come from anyway, at a dynamic range in excess of 100 dB(A)? (I would, however, suggest to turn off all of the unused inputs for playback, just to be sure.) Testing the other line outs, these also could drive my cans without breaking into sweat. I then proceeded with line out tests (using line out 1) involving my amp. I haven't listened for too long yet, but overall the sound is fairly clean, albeit less "groovy" / punchy than on the Aureon (a matter of dynamics, I guess - not too surprising with cheapo capacitors). The soundstage also appears to be a tad smaller. (As always, don't forget the grain of salt. I don't have any audiophile grade equipment, and my hearing is rather average as well.) But overall, pretty good for a budget card. This is one time where the advertising is right, this indeed is a good upgrade from crappy onboard audio, in both D/A and A/D. (EAX and all that, however, is done completely in software.) I bet there's quite a bit of improvement possible if you know where the hot end of the soldering iron is.
I found the CPU load to be a bit on the high side, with basic 48 kHz playback taking about 5-6% on an 800EB. The old XFire takes 2% or so in the same comp, and I don't see why such a simple task should use any more computing power than that. (I mean, hey, a mere 386SX-20 could output 44.1 kHz 16 bit stereo.) The card is apparently sharing an interrupt with a Promise Ultra66, I'll run some cooperation tests later (i.e. "does playback still work flawlessly with a busy harddrive").
This card operates in one of four fixed output modes selectable by choosing the sample rate of either 48 or 96 kHz and the bit depth of either 16 or 24 bits in the "device control" applet. Make sure the settings match the output of the resampler used. (And of course, record in 48 and not 44.1 kHz.)
Using RMAA, loopback measurements were taken with 48 and 96 kHz sample rate and 24 bits. Best results were obtained in 48 kHz, with a THD of 0.0015% and IMD+N of 0.0063%; dynamic range came out at 93.7 dB. Stereo crosstalk levels proved to be excellent, with up to 93 dB and still more than 80 dB at 20 kHz. The frequency response is pretty much dead flat, less than .3 dB down at 20 Hz and less than .1 dB at 20 kHz. As seen in the iXBT review, there's a bunch of power supply harmonics and such crap in the spectrum, but all of the spikes are below -100 dB. (Still, power filtering could be better.) The ADC adds a whole zoo of high-order harmonics, but most of those are around or below -110 dB. At 96 kHz, IMD+N rises to 0.015% (same for sweeping IMD, just higher throughout the spectrum), and you can see a noticeable increase in the noise floor above 30 kHz (noise shaping?). Stereo separation still is better than 75 dB at 40 kHz, the -0.5 dB points of the frequency response are at ~23 Hz and ~40 kHz, respectively. I think I'll upload some RMAA reports later.
For listening tests, I used Winamp 5.08something with the ASIO plugin (resampling to 48 kHz, quality "high") feeding ASIO4All 2.6 (buffer 256 samples, hardware acceleration with 10 ms latency); the Shibatch Super EQ plugin with my usual HD590 compensation setting was used (posted this in the Headphones forum a while back).
The ASIO channels -> output correlation is a bit strange, a channel offset of 2 didn't output the signal on line out 2 but rather a mono downmix on the first output.
Plugging my HD590 directly into the card was no problem. I had to turn down the volume pretty far, the sound itself was quite listenable (not as peaky in the highs are directly out of the Aureon). Audible background noise? Nah. Where is it supposed to come from anyway, at a dynamic range in excess of 100 dB(A)? (I would, however, suggest to turn off all of the unused inputs for playback, just to be sure.) Testing the other line outs, these also could drive my cans without breaking into sweat. I then proceeded with line out tests (using line out 1) involving my amp. I haven't listened for too long yet, but overall the sound is fairly clean, albeit less "groovy" / punchy than on the Aureon (a matter of dynamics, I guess - not too surprising with cheapo capacitors). The soundstage also appears to be a tad smaller. (As always, don't forget the grain of salt. I don't have any audiophile grade equipment, and my hearing is rather average as well.) But overall, pretty good for a budget card. This is one time where the advertising is right, this indeed is a good upgrade from crappy onboard audio, in both D/A and A/D. (EAX and all that, however, is done completely in software.) I bet there's quite a bit of improvement possible if you know where the hot end of the soldering iron is.
I found the CPU load to be a bit on the high side, with basic 48 kHz playback taking about 5-6% on an 800EB. The old XFire takes 2% or so in the same comp, and I don't see why such a simple task should use any more computing power than that. (I mean, hey, a mere 386SX-20 could output 44.1 kHz 16 bit stereo.) The card is apparently sharing an interrupt with a Promise Ultra66, I'll run some cooperation tests later (i.e. "does playback still work flawlessly with a busy harddrive").
