Audigy2 NX - 44.1 -> 48khz upsampling
Jun 19, 2006 at 1:44 AM Thread Starter Post #1 of 10

RichA

100+ Head-Fier
Joined
Jan 19, 2006
Posts
108
Likes
14
Was sitting here watching a film, and thought to myself just how spectacular it sounded. Then it dawned on me - the audio was 48khz, so no upsampling was taking place.

Figured I'd do some testing at the end of the film. Ripped my current benchmark CD(Acoustic Alchemy - Arcan'um) to FLAC. Then to ogg with -q10 - one at 44.1khz, the other at 48khz. Not a massively fair test, as the original was 44.1khz, and both samples will be coloured by ogg(if you believe such things...), but I figured it would be good enough to let me see(hear?) the difference between upsampling with the Audigy, and the upsampling in ogg.

Difference was night and day, when listening critically. Listening to the middle part of a delicate track, I was able to pick the file upsampled by Audigy 80% of the time. But from the start of the track, it was 100%. Silence, followed by the sharp striking of the guitar - I got it every single time.

Don't get me wrong - I still don't think the Audigy2 NX is a bad sounding card, despite what some people here will tell you. I've defended the card quite religiously, and still would. But the upsampling is far from perfect. Not tried it with the Audigy2 SE I have, but I'm guessing that it would share a similar tale...

On a good note, I couldn't tell any difference whatsoever between -q6 and -q10, so when I build my Alien DAC, hopefully my decision to rip to ogg won't have been a bad one.

ABXing was done on my headphones - I'd say the difference would be even clearer if I was using my loudspeakers.

Will gladly post samples if anyone cares
smily_headphones1.gif
.

ETA:- Have just realised you can set the sample rate within FLAC. Had a quick go with two FLAC samples - 44.1 and 48khz. Got it right 7/7 times.
smily_headphones1.gif


--Rich
 
Jun 20, 2006 at 1:18 PM Post #2 of 10
How about 44,1KHz ogg resampled to 48KHz via playback software (winamp, foobar etc.) vs 48KHz ogg? If I read your post right, you didn't try that yet.

I'd be interested which one, encoder or player, does better job at resampling. And if encoder wins, is it worth it to re-encode thousands of songs because one happened to buy a crappy soundcard (I have AUdigy 2 too)
600smile.gif
 
Jun 20, 2006 at 1:41 PM Post #3 of 10
MacGyverille terveisiä täältä napapiirin pohjoispuolelta.


RichA:

Now, when you're in deep testing process
mad.gif
, could you compare those players/plugins you have there w/ NI BeatPort Player using FLAC? BeatPort Player doesn't need external resampler nor dithering plugins (both are build-in and automatic -> you just set the SR for output) so it's all the same which SR/bit-depth you're using on sources.

As BPP supports native ASIO, MME and DS, it's possible to use it w/ NX, but ASIO4ALL could be used too to add 'ASIO' (KS) support for 44.1/48/96 kHz samplerates. Also, if you have the SE (iIrc, it's not ASIO card either) running you can compare w/ it too.

You can also compare NX sound quality w/ USB-ASIO drivers (www.usb-audio.com).

jiitee
 
Jun 20, 2006 at 3:56 PM Post #4 of 10
Quote:

Originally Posted by Ihmemies
How about 44,1KHz ogg resampled via playback software (winamp, foobar etc.) vs 48KHz ogg?


Pass. Will have a fiddle later. Don't use Windows, so no Winamp or Foobar here
smily_headphones1.gif
.

Quote:

Originally Posted by jiiteepee
MacGyverille terveisiä täältä napapiirin pohjoispuolelta.


RichA:

Now, when you're in deep testing process
mad.gif
, could you compare those players/plugins you have there w/ NI BeatPort Player using FLAC? BeatPort Player doesn't need external resampler nor dithering plugins (both are build-in and automatic -> you just set the SR for output) so it's all the same which SR/bit-depth you're using on sources.

As BPP supports native ASIO, MME and DS, it's possible to use it w/ NX, but ASIO4ALL could be used too to add 'ASIO' (KS) support for 44.1/48/96 kHz samplerates. Also, if you have the SE (iIrc, it's not ASIO card either) running you can compare w/ it too.

You can also compare NX sound quality w/ USB-ASIO drivers (www.usb-audio.com).

jiitee



Again, I don't use Windows, so ASIO doesn't apply to me, unless there is something I don't already know?

--Rich
 
Jun 20, 2006 at 4:48 PM Post #5 of 10
Quote:

Originally Posted by RichA
Pass. Will have a fiddle later. Don't use Windows, so no Winamp or Foobar here
smily_headphones1.gif
.

Again, I don't use Windows, so ASIO doesn't apply to me, unless there is something I don't already know?

--Rich



Gosh, yeah, OK, well least now I know what the Kubuntu means ...
k1000smile.gif



jiitee
 
Jun 20, 2006 at 6:44 PM Post #6 of 10
^Heh, Kubuntu rocks
smily_headphones1.gif
. Standard audio driver these days is ALSA - there is a low-latency driver called Jack, but I don't think it offers any improvement in terms of sound quality. But I'm clueless, so expect to be proven wrong
wink.gif
.

Quote:

Originally Posted by Ihmemies
How about 44,1KHz ogg resampled to 48KHz via playback software (winamp, foobar etc.) vs 48KHz ogg? If I read your post right, you didn't try that yet.

I'd be interested which one, encoder or player, does better job at resampling. And if encoder wins, is it worth it to re-encode thousands of songs because one happened to buy a crappy soundcard (I have AUdigy 2 too)
600smile.gif



Tried it. With my ears, amp, cans, loudspeakers, and choice of music, I could find no perceivable difference between the following:-

44.1khz WAV -> 48khz EAC, courtest of FLAC
44.1khz WAV played back directly, but resampled to 48khz via ALSA

No comment on how Windows would sound - I'd expect it to be different, given the moanings on here
smily_headphones1.gif
.

--Rich
 
Jun 20, 2006 at 7:43 PM Post #8 of 10
Why would I need one? I can set ALSA(my sound system) to resample automatically, if needed...

--Rich
 
Jun 21, 2006 at 6:25 PM Post #9 of 10
I abx'd one 320kbps lossy mp3 (just for kicks), only 7 trials (I don't know how to adjust the amount...

Anyways, I converted the mp3 first to 16bit/44,1KHz wav without dsp's, and another with SRC resampler to 16bit/48KHz wav. I switched playback to directsound, 16bit, without any dsp's/resampling or anything like that.

Result was 7/7, so I suppose that Audigy 2's internal resampling is poorer than resampling done by software. To prove it properly to others too, I should do at least the double amount of trials. Today 7 shall be enough proof for me.
rolleyes.gif
 
Jun 22, 2006 at 3:48 AM Post #10 of 10
What I do when running in Windows is switch my Audigy 2 NX to "digital only" mode (inside the "advanced" volume settings), making sure the device is in 44.1 KHz 16-bit mode and set to use external processing (these options are found within the Creative Device Control program).

Then I feed the non-resampled signal to my external DAC and enjoy the pure goodness.
icon10.gif


If you happen to run linux you might've come across a "small" problem, mainly that the ALSA devs have disabled 44.1 KHz playback (because it sounds like crap due to the card not being able to handle it). They also haven't implemented a way to get "digital only" mode working.

However, being the impatient person that I am, I took it upon myself to hack the usb-audio ALSA-driver so that it supports "digital only" mode and reenables 44.1 KHz playback (not usable for the analog outputs since it still sounds like crap, however it's needed to passthrough a non-resampled 44.1 KHz signal for external processing).

Hopefully I'll get the patch submitted to the ALSA-devs this week or the next. Feel free to PM me for more details.
 

Users who are viewing this thread

Back
Top