Skip to main content
Topic: Why Is The Sb Live So Bad? (Read 6632 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Why Is The Sb Live So Bad?

Everywhere I read that the SB Live is a crappy sounding piece of crap directly to be tossed into the bin. The question is WHY? Is there any website explaining? And I think on this forum was mentioned that the SB 128 PCI was be much better...  The SB Live was a great improvement on my system (before I used a SB16... ISA, with IDE, CT2290 to be precise. Yep... the modell from 1994. Now THAT is noise. But still a great card).

Why Is The Sb Live So Bad?

Reply #1
pci 128 does output 44 khz

so you 44khhz mp3' are play at 44 khz

sblive/audigy outputs 48khz meaning that you 44khz mp3 are opampelt to 48khz witgh alters the sound a bit

audigy does a much better upsampling job then live
Sven Bent - Denmark


Why Is The Sb Live So Bad?

Reply #3
hmm, I see... is this upsampling so bad and is it impossible for it to output at 44 kHz?


Why Is The Sb Live So Bad?

Reply #5
uhm, ok zZz... soz for typing 2 words big. Or is there anything else wrong? Or do you mean the first sentence which is indeed a bit long? 

Why Is The Sb Live So Bad?

Reply #6
Quote
uhm, ok zZz... soz for typing 2 words big. Or is there anything else wrong? Or do you mean the first sentence which is indeed a bit long? 

kadajawi: I believe he mentioned Sven's "opampelt"

Why Is The Sb Live So Bad?

Reply #7
Oohmpahloompa? Aaaaaaaaaah! Help it is that foriegn language stuff!

Why Is The Sb Live So Bad?

Reply #8
Quote
Oohmpahloompa? Aaaaaaaaaah! Help it is that foriegn language stuff!

Isn't that "Willy Wonka's Amazing Chocolate Factory"?

Why Is The Sb Live So Bad?

Reply #9
Quote
hmm, I see... is this upsampling so bad and is it impossible for it to output at 44 kHz?

No, in a blind test using a capture of the resampled digital output of the SB live vs accurate resampling in SoundForge/CoolEdit/SSRC/Lame, half of the listeners (that is one out of two) couldn't make any audible difference between the SBlive resampling / CoolEdit etc resampling / original not resampled (I still have the samples).

SB live mostly earned its bad reputation because of unsolvable bugs with MOBO using the Via133 chipset (in fact, a different patch is realized for each 4-in-1 drivers version), and for endless skipping problems with Winamp.

Why Is The Sb Live So Bad?

Reply #10
Quote
...half of the listeners (that is one out of two)

LOL

Now getting back to the subject, bad resampling on the Live will sound more like a strange colorization in highs due to added harmonics. I don't think it's that easy to pick up...

BTW you should do a search for topics related to the subject :D

http://www.audio-illumination.org/forums/i...7&highlite=live

Why Is The Sb Live So Bad?

Reply #11
SB Live and Audigy also have a frequency rolloff starting at 16 KHz, and IIRC going above -9 dB somewhere at 19 KHz. Then, they are both hogging the PCI bus (always reserving bandwidth), which triggers erratic behaviour especially with VIA chipsets. Not to mention, their drivers belong to the worst ones ever programmed.

Why Is The Sb Live So Bad?

Reply #12
 I think the search for live wasn't THAT successful 
Quote
maybe not so great for classical & live music (but that's OK) .

although there was indeed some info. I don't use a KT133 chipset, but Winamp skipping problems? Any short summary? (nothing long needed  )

Ps: sven doesn't deserve to be flamed... just look where he comes from. It just might be that he is not a native english speaker...

Why Is The Sb Live So Bad?

Reply #13
In the middle of the thread about the SSRC plugin for Winamp, a serious flaw about SB live reviews was uncovered :
http://www.audio-illumination.org/forums/i...1&t=2101&st=104

In fact, with some (most ?) Windows/live version combination, it seems that the sound quality is set to low by default !

The trick is to go to the control panel / multimedia / audio / playback / advanced settings, and choose hifi speakers and best quality instead of desktop speakers and/or good quality.

For me however, it had no effect (Live CT4620, Windows 98)

Why Is The Sb Live So Bad?

Reply #14
Quote
I don't use a KT133 chipset, but Winamp skipping problems? Any short summary? (nothing long needed   )

Well, in fact it all related to the VIA chipset :

http://forums.winamp.com/showthread.php?threadid=36850

Skipping / freezing / clicking & popping during playback section

4) VIA Chipset woes subsection

Why Is The Sb Live So Bad?

Reply #15
Quote
In the middle of the thread about the SSRC plugin for Winamp, a serious flaw about SB live reviews was uncovered :
http://www.audio-illumination.org/forums/i...1&t=2101&st=104

In fact, with some (most ?) Windows/live version combination, it seems that the sound quality is set to low by default !


Ah yes..  for some moving the slider to "best" (if your operating system didn't do that already) made a big difference..  for other systems and setups it didn't.

My Audigy and XP didn't gain, or even lose anything, (at 44.1Khz) for example..  but you will find that recording/playing back at 48Khz will get rid of the >16Khz rolloff,  as well as a lot of alialsing error on high freq's, and improve the quality of several paramaters..  on both cards..

The other guy did see much better performance just with moving the slider,  with a Live and 98..  more than worth a try!

YMMV and all that..  but read the thread thru,  get RMAA,  and start experimenting!

I might mention,  I never experienced the dreaded breakups, with my Live or Audigy,  over 2+ yrs running my Abit KT7 board..  always using current 4in1 drivers.

Why Is The Sb Live So Bad?

Reply #16
Just to add another data point, my SB Live Value would skip in Winamp under Win2k fairly often, but increasing the size of the output plugin's buffers fixed this.  No problems in XMMS under Linux either.

Why Is The Sb Live So Bad?

Reply #17
Quote
hmm, I see... is this upsampling so bad and is it impossible for it to output at 44 kHz?

I can tell the SB Live samples in http://www.pcabx.com/product/ct4830/index.htm (french horns) at an ABX test with confidence over 99%.

Didn't try Pio's test.

Why Is The Sb Live So Bad?

Reply #18
with my extigy I can set the freq to 44.1

Why Is The Sb Live So Bad?

Reply #19
Quote
with my extigy I can set the freq to 44.1

u can do that in audigy too.. using the app that comes with it.

they're all ghey anyway :S

the problem is, whe ur a gamer u have no choice!!

Why Is The Sb Live So Bad?

Reply #20
I don't think I could ABX the pcabx samples.

However, the experiment is not the same as mine. These samples were playback in the SB Live digital out, then recorded into the SB Live digital in. So they came through 44.1 to 48 kHz resampling (during playback), then, at the recording, 48 to 44.1 resampling, leveling, dithering then another level compensation to restore their original level.
Moreover, AC97 specifies playback resampling for fixed frequencies, while the recording uses Asynchronous Sample Rate Conversion. So there may be more loss at the recording than at the playback (but the SB live may also use ASRC for playback, I don't know)

In my tests, the SB live outpout was recorded by the Marian Marc 2 digital input in slave mode. Therefore asynchronous resampling from 48 to 44.1, leveling, dithering and level compensation again were avoided. My sample is a raw lossless 48 kHz capture of the SB live output.

You can get it here :

Right-click to download original.pac
Right-click to download 1.pac
Right-click to download 5.pac

All three files are about 1 MB. Besides the original, there are two 48 kHz files (1 and 5.pac). Together with the SB live capture, there is another one made by SSRC, with ATH dither. I provide it for reference, as the 48 kHz one can be told apart from the 44.1 kHz one by AC97 soundcards owners, because 48 kHz playback is supposed to sound better than 44.1 kHz playback.
Even for true soundcards owners, cheap converters as the ones in PC soundcards may behave better at 48 kHz than at 44.1 kHz. So it's better trying to ABX 1 and 5 (both 48 kHz) than Original and SB Live (44.1 versus 48 kHz).

By the way, even setup at 44.1 kHz, there can be losses if the soundcard uses ASRC (from 44.1 to 44.1 with losses !). Someone should record the digital output with another soundcard in slave mode in order to check for differences between the files.

Why Is The Sb Live So Bad?

Reply #21
With the Audigy, and I believe the Extigy too, you can play 44.1 KHz files, but the card internally resamples them to 48 KHz before playing them, because it only can play by hardware 48 KHz data. Same as with the Live, the difference is that the live does this quite badly.

Why Is The Sb Live So Bad?

Reply #22
Quote
I don't think I could ABX the pcabx samples.

The Live samples sound a little more distorted in the high frequencies, I hadd less problems ABXing them with speakers than with headphones.

Quote
However, the experiment is not the same as mine.


Yes, that's true.

I will try your samples tonight or tomorrow, I can't right now. I haven't downloaded them yet, but I think it would be easier to ABX the card's effect with samples that exploit its faults.

I have access too to a Live card (not in my computer, though), so maybe I could try to replicate your experiment with the PCABX french horn sample, when I have some more time.

Why Is The Sb Live So Bad?

Reply #23
I noticed this effect on my SB Live Value, I was resampling from 22khz to 44khz, and noticed it sort of reflected the frequencies at 8-11khz across the 11khz boundry, into the 14-11khz region.  I asked around in the IRC channel, and found that it was because the card didn't resample perfectly, but to be honest, I preferred the sound it caused than resampling "perfectly" in Cool Ed.

One question, where do you find this slider that lets you set sound quality to low or maximum?  I can't find one in Win2K?
< w o g o n e . c o m / l o l >

Why Is The Sb Live So Bad?

Reply #24
Quote
I will try your samples tonight or tomorrow.

I've tried to ABX the original versus the 1_5.pac file (SB resampling). First I tried concentrating on highs on different parts of the clip, but didn't succeeded. At last, concentrating just on same slight reverb or volume difference at the vocal part, I got 10/11, p=0.6%. It was quite tough, however, the differences came and went continuosly.

But, to discard different performance of my non resampling card (Audiophile 2496) at 44.1 and 48 KHz (shouldn't happen, however), I should try to ABX the 1_1.pac file (SSRC) versus the 1_5.pac file.

 
SimplePortal 1.0.0 RC1 © 2008-2020