Line In noise: is it the same for Live, Audigy and Augidy 2? Do them all use AC97?

Daniel Drummond · Oct 5, 2003, 10:35 PM

I was talking to LeMury the other day (via e-mail) and he made me the favor of checking the noise level of the Line In in his Audigy 2. They were around -60db. That's the same I get in my sblive. With a LiveDrive I get -78db when faders are set to 0db and -65db when faders are set to +12db. Why does that happen?
How good is being able to record at 24bit 96kyz in an Audigy 2 if its input is so noisy?
Is an Audigy 2 only a benefit from a Live card when it's associated with an AudigyDrive?
Also, can an AudigyDrive record sound at 24/96 or that can be achieved only through the cards own Line In connector?

I really hope some one has those answer because I need to decide if it is worth trading a Live+LiveDrive IR into an Audigy 2 or not... Please help me decide!!!

**Eugene Gavrilov** · Oct 6, 2003, 12:48 AM

>> Why does that happen?

line-in uses ac97 codec -- and it is a very noisy device

(with rather bad technical specifications)

livedrive uses (better) i2s codecs

>> can an AudigyDrive record sound at 24/96 or that can be achieved only through the cards own Line In connector

audigy2's line-in is still using ac97 codec
(its results are slightly better than live's since it is wired better)

24/96 has sense only for audigy/live-drive recordings

>> Live+LiveDrive IR into an Audigy 2 or not

audigy2's line in is worse than live+livedriveIR combo (if using line2/mic2)

/E

Daniel Drummond · Oct 6, 2003, 12:56 AM

"audigy2's line in is worse than live+livedriveIR combo (if using line2/mic2)" Well, I guess that solves my problem. I'm not "upgrading" them!!! I'll wait till I have enough money to buy Audigy2 + audigydrive so it makes sense... But one point is still not clear: does Audigy 2 record 24/96 through audigydrive??? Thanks.

Lex Nahumury · Oct 6, 2003, 01:24 AM

To be more specific, and confirming what Eugene mentioned;

Test conditions:
1. Total reset of driver settings
2. Clear whole dsp
3. connect only Prolog AC97 -> Peak Meter (in Horizontal mode)
4. at Mixer's 'In's& Out's' page Unmute AC97
5. No external cabels/signals connected/applied

Results:

-Audigy2 OEM (SigmaTel STAC9721/23) = ~60dB

-SBLive5.1 OEM (SigmaTel STAC9708) = ~ 60dB

*Remark;
Test was done with 3534 drivers.
Perhaps 3535 gives other results due to different AC97 gain setting (reg 1C),
but I don't think so because I did a quick test with lowest AC97 Gain setting.
Results were the same.

On the other hand; what really matters is 'Signal to Noise' ratio.
(See Max.M's post in dsp section on this item)

Anyway, it doesn't bother me since I have never heard noise in my recordings
either with SBlive or Audigy 1 and 2.
Analog playback is a different story. Audigy has far less noise.

/LeMury

Daniel Drummond · Oct 6, 2003, 01:56 AM

LeMury, I understand the what matters is signal to noise ratio, but when I record from the card's line in, in quiet passages (piano playing) I'm able to hear noise in the final mix and that doesn't happen with the livedrive. That's why I'm thinking that I'd be better of keeping my current setup and waiting till I have more money to buy the who audigy 2 pack. I understand that for playback audigy is better. Well, it just ocurred to me that I can get an Audigy 1 used for a good price and keep the livedrive IR, so I can improve playback and keep low noise recording!!! That's what I'll do then! Thanks Eugene and LeMury for making my choice clearer...

Lex Nahumury · Oct 6, 2003, 02:32 AM

Daniel,

>> but when I record from the card's line in, in quiet passages (piano playing) I'm able to hear noise
>> in the final mix and that doesn't happen with the livedrive.

Ah I see. Yep of course, when that is the case you better not use AC97 based input.
(hmmm...probably I dont make music with that much dynamic range, or my ears suck..hehehe)

>>That's why I'm thinking that I'd be better of keeping my current setup and waiting till I have more
>>money to buy the who audigy 2 pack.

Sounds like the best thing to do in your case.

>> Well, it just ocurred to me that I can get an Audigy 1 used for a good price and keep the livedrive IR,
>> so I can improve playback and keep low noise recording!!!
>>That's what I'll do then!

Nice alternative solution..!

/LeMury

snop · Oct 6, 2003, 07:58 AM

Hi all,

I'm not an expert but a newbie... Reading this post makes me think about audigy/live recording quality. Apart from the signal/noise ratio, is there any test to see (hear) how good are audigys and lives when recording ? I mean, for instance, to plug a signal generator to line in and try to record different signals (with different frequencies and/or amplitudes). I don't know if this test would be useful or not (or if it has been already done).

On the other hand: what device could be used to transform an analog signal to a digital signal (through spdif) ? If they exist, are they expensive ? it's worth to buy one or it's better to get a professional sound card ?

Thanks in advance for your replay (and please forgive me if my English is not fine...)

SnOp

RIV@NVX · Oct 6, 2003, 09:42 AM

Audigy2 Platinum eX doesn't have AC97 Codec; it should be better on this one.

But it's also a lot more expensive...

TravelRec. · Oct 6, 2003, 11:34 AM

Hi around here...

If you want to improve your inputs and outputs, it´s not really necessary to buy another card. A guy in germany is working on HIGH END Converters from analogue to SPDIF and opposite. Please take a look to:

http://www.beis.de/Elektronik/ADDA2496/ADDA2496.html

The more the interest, the faster some results, I think...

Cheers!

TravelRec.

Daniel Drummond · Oct 6, 2003, 01:53 PM

For audigy and live sound quality reviews, check 1)Live 5.1 http://www.digit-life.com/articles/sbliveplatinum51/ 2)Audigy Platinum EX: http://www.digit-life.com/articles/a...mex/index.html 3)Audigy 2 family: http://www.digit-life.com/articles2/...gy2/index.html 4)Audigy 2 Platinum Ex: http://www.digit-life.com/articles2/...-ex/index.html

Lex Nahumury · Oct 6, 2003, 05:27 PM

Daniel,

This was the post I was looking/refering for/to:
http://www.hardwareheaven.com/showthre...threadid=26832

When I got the time I will;
- attach an external Signal Generator to Line-In
- decrease ac97 faders and
- increase external input signal untill analog clipping occurse

(Or we could just ask Mattcro's bevindings)

It might reveal not so bad specs afterall.

/LeMury

Daniel Drummond · Oct 6, 2003, 07:13 PM

Yep, maybe the input is amplified and that is causing the bad singal to noise ratio (is that what you mean?).

In this review http://www.digit-life.com/articles/sbliveplatinum51/ using creative drivers, they have obtained this results:

Test type: External loopback, 48000 Hz sampling, -3dB@1kHz
SNR, dB 73.2 input in the card and db 82

They have also done other tests. Please take a look at them. If my previous posts there are links to tests with audigy and audigy2.

I don't quite understand those kind of testes and results, but they may shed us some more light on the subject...

Lex Nahumury · Oct 6, 2003, 09:37 PM

Daniel,

>>Yep, maybe the input is amplified and that is causing the bad singal to noise ratio
>>(is that what you mean?).

Well,.more precise but in 'mortal' terms;

"a to low analog input signal which (once converted) gets digitaly amplified to the desired recording level"

So if one could raise the analog input level up to using ADC's full range,
no dsp amplification is needed thus improving SNR (Signal to noise Ratio).

In practise;
Move the AC97 slider to ~25% and you will see noise level drop to ~75dB.
Now apply a test (sine) signal to Line-In (e.g. from your mixer console) and increase it untill
clipping occurs at the ADC. (you can hear but not see this) Then back of a little.
Now watch your record level. If it is ~ 0dB you have improved your SNR.
If it's much lower, you have to turn up the AC97 slider untill record level reaches ~0dB.
Now remove test signal and measure noise level.
That result is as good as it gets with that particular Codec.

That is basicly the test/issue as described in that thread but now in "lamers" terms.
(Anybody; Fixme..)

Read that thread carefully since it also depends on driver version.

Hope this helps a bit,

/LeMury

Daniel Drummond · Oct 7, 2003, 04:38 AM

LeMury, I have just followed your steps and tested it. The sweet spot for my soundblaster live (4.1 CT4760) is setting the AC97 slider to 28.5. That way I achieve -69 to -73, with the mixer faders all the way down, -67 to -69 with the mixer faders outputting the maximum volume without distiortion (I just kept making as much noise as I could with my keyboard and it achieves 0db but nevers surpasses it). That's great! But I have good news! I'll be getting an used Audigy 1 OEM tomorrow, so I'll experience better output quality sound while mantaining my livedrive IR! They guy will send me the audigy via express mail and it'll probably be with me on wednesday or thursday! I'll then do all this tests again to achieve the best SNR. Man, am I learning good stuff from this forum!? Thanks!

Daniel Drummond · Oct 7, 2003, 04:48 AM

LeMury, I was also reading the post you recommended and there's something I can't get... Eugene stated the in 3535b he had removed the 4x amplification. If he really did it (and I think he did, comparing both codes of older and newer prologs) I had to able to attain a lower SNR leaving the AC97 slider at 100 and not at 28.5 (where it is set now...). Am I missing something else here??? By the way I have a different AC97 codec than that guy has. Mine is SigmaTel STAC9721/23 and his is STAC9708.

Lex Nahumury · Oct 7, 2003, 05:43 AM

Daniel,

Cool,..Congrats with your purchase of the adg1.
So you will install both cards right?

As for the 4x amplification;
I currently haven't any 3535 version installed so I can't check right know.
But to be honest, It's not (yet) clear to me if/were and how Eugene removed that 4x amp.
I had took a quick look at 3535 SDK source code some time ago, but can't remember finding
any indication to this. If it's done in kmixer.exe, there is no way I can tell since it's not open source.
I do remember the change in Epilog's output code though.
O well, I still have to check 3535c, so I will dig deeper then.

And/or maybe Eugene could shed some light on this...

However;
The same rules still apply. Maybe with other ac97slider settings but who cares
as long as the SNR result are improved by properly adjusting the input level vs the ac97 level slider.

/LeMury

P.S.
As a test signal you could simply use a Midi synth/module putting out a non enveloped sustaining
sine sound. Makes testing a lot easier than banging on your keyboard while adjusting settings

P.P.S
Come to think of it; you could also use a sustaining Sine soundfont, play it back and amplify it
through your external Mixer and feed it to Line-IN. Right?
Duhh..i'm tired. cant think straight anymore..i'll go to bed now...cya..

Daniel Drummond · Oct 7, 2003, 06:03 AM

If you right click epilogs from version 3534 and 3535 and click on edit you'll see the differences. 3535 has both versions available.

Yep, I could use a sine wave signal but the problem is that when I play more than one note simultaneously, the sound goes louder. This way, if I set the AC97 slider volume using a sine wave, when I do real playing, clipping will occur. So I just made my keyboard make sounds as loud as it can, to never have clipping.

I'll probably not use both cards. I need to sell the live to get back part of the money used with audigy. Unless I can make some extra money this week... Will try hard!

Lex Nahumury · Oct 7, 2003, 07:01 AM

>>If you right click epilogs from version 3534 and 3535 and click on edit you'll see the differences.
>>3535 has both versions available.

yes I know, but if I remember correctly, there was no change in the amplifier code.

>>Yep, I could use a sine wave signal but the problem is that when I play more than one note
>>simultaneously, the sound goes louder. This way, if I set the AC97 slider volume using a sine wave,
>>when I do real playing, clipping will occur. So I just made my keyboard make sounds as loud
>>as it can, to never have clipping.

No. The right procedure is to determine the 'right' ac97slider setting and never touch it again!
(Using a single sustained test signal is really more accurate for this.)
Then when you start recording Live stuff (which of course may be louder or softer),
you ONLY adjust the output levels from your external Mixer console while
watching the Record Vu meters in Cubase for clipping/level check.
Done this way you will always achieve a constant and best possible SNR without the risk of clipping.

I realy hope you understand this,

/LeMury

Daniel Drummond · Oct 7, 2003, 07:22 AM

Yes, that really makes sense... I got it. I'll do that them! Thansk once more...

RIV@NVX · Oct 7, 2003, 08:28 AM

Quote:

Originally posted by Daniel Drummond
I'll probably not use both cards. I need to sell the live to get back part of the money used with audigy. Unless I can make some extra money this week... Will try hard!

Are you planning to sell the LiveDrive? I think it can work with Audigy1 too, so I suggest you to try it before selling.

tubus · Oct 7, 2003, 12:32 PM

As for external A/D Converters, I use (otherwise broken) DAT-recorders with good results.
It doesn't simplify the setup and handling, but I got them for free and it's very easy to add a digital input even to the oldest lives ...

tubus

Daniel Drummond · Oct 7, 2003, 01:52 PM

Riv@nx, I'm keeping the livedrive... I decided to keep it and get an Audigy instead of selling it to get an Audigy 2. It made more sense to me.

**Eugene Gavrilov** · Oct 7, 2003, 06:19 PM

yep, I've removed an (invalid) amplification
but it doesn't mean the amplification is impossible

in the pre-3535 releases the ac97 signal was amplified, while the others weren't
in 3535c and later releases there are no amplified signals, however, the level slider
is used for amplification instead

that is:
level = 0.0 ... 1.0

pre-3535:
output(ac97) = input*level
output(all the rest) = (input * level) / 4
(/4 is necessary for any hardware input)

post-3535:
output = (input * level)

thus, you get a possibility to use +12dB gain (if necessary)

new epilog / prolog / xrouting are used for that

legacy settings should function correctly (but the amplification feature won't be available)
new kX Mixer displays the signal levels in dB (if possible)

the 'ac97' slider in the kX Mixer is displayed in %% since it can be 'assigned' for WinMM/ Windows Mixer compatibility
if you un-assign it, the slider will be displayed in dB

/E

Lex Nahumury · Oct 7, 2003, 06:32 PM

Thanks Eugene for clearing that up.

/LeMury

RIV@NVX · Oct 7, 2003, 09:34 PM

Quote:

Originally posted by Daniel Drummond
Riv@nx, I'm keeping the livedrive... I decided to keep it and get an Audigy instead of selling it to get an Audigy 2. It made more sense to me.

That's a very good decision!

Lex Nahumury · Oct 7, 2003, 11:47 PM

Daniel,

I just finished some testing with an external signal generator and osciliscope.

With 3535c drivers;

-Only Prolog AC97 -> Epilog Rec L/R
-AC97 slider on 'Ins & Outs' page at 0dB

SBLive5.1 OEM Line-In ~-72dB

(no time to test adg2 OEM, but probably the same)

Quite an improvement I'd say.
Well, thanks Mattcro, Max and Eugene...!

/LeMury

Daniel Drummond · Oct 10, 2003, 09:13 PM

Ok LeMury, but why do you get +-72db testing it that way (with faders at 0db) and Peak plugin "only" gives +-62db when connected straight to prolog's line in outputs? Thanks.

Lex Nahumury · Oct 10, 2003, 09:41 PM

Quote:

Originally posted by Daniel Drummond
Ok LeMury, but why do you get +-72db testing it that way (with faders at 0db) and Peak plugin "only" gives +-62db when connected straight to prolog's line in outputs?

????

See;

http://members.home.nl/nahutec/kxplugins/snr.jpg

/LeMury

Daniel Drummond · Oct 11, 2003, 05:07 AM

Well, I have just figured out what is going on... I had made a setup with drivers version 3534 and loaded them with drivers 3535rc2. So in the INS AND OUTS pages the AC97 was not show in db and didn't have a little button to set up 0 db. And I could only get that +-72 when I lowered it to number 25. I then cleared the DSP window and started building the same setup from scratch. For my surprise, now the AC97 fader is shown in db, has the button to set it to 0db and I got the exact same values as you did... hehehe Sorry for the mess... Thanks for making me re-check it.

Lex Nahumury · Oct 11, 2003, 03:36 PM

Daniel,
>>..For my surprise, now the AC97 fader is shown in db, has the button to set it to 0db and I got
>>the exact same values as you did...

Yes, In my case, for some reason the AC97 slider keeps switching randomly between dB and numeric readout.
No big deal, if set to "No amplification 0db" (cant remember in numeric) then it's fine.
(besides; I dont use Prolog, but my own 'Scr' in which amplification is 'hardcoded' at 0dB with no sliders to mess with)

Since Eugene fixed the amplification there is no need to tweak anymore.
Simlpy setting Line-In (ac97) to 0dB and only adjusting external Analog signal level should give best SNR.
For all AC97 based input this would seem to around ~-72dB.
(I checked Sblive, adg1 and adg2 AC97 based Line-In. They were all he same)
I dont have LiveDrive so I can't tell for I2S<-->Codec based inputs.

btw;
You also may checkout 'rc3 update files';
http://www.hardwareheaven.com/showthre...5&pagenumber=1
They fix new epilog's -6db output issue.
In simple words; Ouput sounds as loud as it used to be (when sliders are at 0dB).

/LeMury

Oct 5, 2003, 10:35 PM	#1
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	Line In noise: is it the same for Live, Audigy and Augidy 2? Do them all use AC97? I was talking to LeMury the other day (via e-mail) and he made me the favor of checking the noise level of the Line In in his Audigy 2. They were around -60db. That's the same I get in my sblive. With a LiveDrive I get -78db when faders are set to 0db and -65db when faders are set to +12db. Why does that happen? How good is being able to record at 24bit 96kyz in an Audigy 2 if its input is so noisy? Is an Audigy 2 only a benefit from a Live card when it's associated with an AudigyDrive? Also, can an AudigyDrive record sound at 24/96 or that can be achieved only through the cards own Line In connector? I really hope some one has those answer because I need to decide if it is worth trading a Live+LiveDrive IR into an Audigy 2 or not... Please help me decide!!! __________________ www.palcomp3.com.br/ed

Oct 6, 2003, 12:56 AM	Thread Starter #3
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	"audigy2's line in is worse than live+livedriveIR combo (if using line2/mic2)" Well, I guess that solves my problem. I'm not "upgrading" them!!! I'll wait till I have enough money to buy Audigy2 + audigydrive so it makes sense... But one point is still not clear: does Audigy 2 record 24/96 through audigydrive??? Thanks. __________________ www.palcomp3.com.br/ed

Oct 6, 2003, 01:56 AM	Thread Starter #5
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	LeMury, I understand the what matters is signal to noise ratio, but when I record from the card's line in, in quiet passages (piano playing) I'm able to hear noise in the final mix and that doesn't happen with the livedrive. That's why I'm thinking that I'd be better of keeping my current setup and waiting till I have more money to buy the who audigy 2 pack. I understand that for playback audigy is better. Well, it just ocurred to me that I can get an Audigy 1 used for a good price and keep the livedrive IR, so I can improve playback and keep low noise recording!!! That's what I'll do then! Thanks Eugene and LeMury for making my choice clearer... __________________ www.palcomp3.com.br/ed

Oct 6, 2003, 07:58 AM	#7
snop DriverHeaven Newbie Join Date: Oct 2003 Posts: 1 Rep Power: 0	Not only noise Hi all, I'm not an expert but a newbie... Reading this post makes me think about audigy/live recording quality. Apart from the signal/noise ratio, is there any test to see (hear) how good are audigys and lives when recording ? I mean, for instance, to plug a signal generator to line in and try to record different signals (with different frequencies and/or amplitudes). I don't know if this test would be useful or not (or if it has been already done). On the other hand: what device could be used to transform an analog signal to a digital signal (through spdif) ? If they exist, are they expensive ? it's worth to buy one or it's better to get a professional sound card ? Thanks in advance for your replay (and please forgive me if my English is not fine...) SnOp

Oct 6, 2003, 09:42 AM	#8
RIV@NVX Freedom is a feature. Join Date: Jul 2002 Location: Croatia, Rijeka Posts: 4,476 Rep Power: 0 System Specs	Audigy2 Platinum eX doesn't have AC97 Codec; it should be better on this one. But it's also a lot more expensive... __________________ -- Vedran

Oct 6, 2003, 12:48 AM	#2
Eugene Gavrilov kX Project Lead Programmer and Coordinator Join Date: Dec 2002 Posts: 3,119 Rep Power: 75	>> Why does that happen? line-in uses ac97 codec -- and it is a very noisy device (with rather bad technical specifications) livedrive uses (better) i2s codecs >> can an AudigyDrive record sound at 24/96 or that can be achieved only through the cards own Line In connector audigy2's line-in is still using ac97 codec (its results are slightly better than live's since it is wired better) 24/96 has sense only for audigy/live-drive recordings >> Live+LiveDrive IR into an Audigy 2 or not audigy2's line in is worse than live+livedriveIR combo (if using line2/mic2) /E

Oct 6, 2003, 01:24 AM	#4
Lex Nahumury DH Senior Member Join Date: Jan 2003 Location: The Netherlands Posts: 1,932 Rep Power: 64	To be more specific, and confirming what Eugene mentioned; Test conditions: 1. Total reset of driver settings 2. Clear whole dsp 3. connect only Prolog AC97 -> Peak Meter (in Horizontal mode) 4. at Mixer's 'In's& Out's' page Unmute AC97 5. No external cabels/signals connected/applied Results: -Audigy2 OEM (SigmaTel STAC9721/23) = ~60dB -SBLive5.1 OEM (SigmaTel STAC9708) = ~ 60dB *Remark; Test was done with 3534 drivers. Perhaps 3535 gives other results due to different AC97 gain setting (reg 1C), but I don't think so because I did a quick test with lowest AC97 Gain setting. Results were the same. On the other hand; what really matters is 'Signal to Noise' ratio. (See Max.M's post in dsp section on this item) Anyway, it doesn't bother me since I have never heard noise in my recordings either with SBlive or Audigy 1 and 2. Analog playback is a different story. Audigy has far less noise. /LeMury

Oct 6, 2003, 02:32 AM	#6
Lex Nahumury DH Senior Member Join Date: Jan 2003 Location: The Netherlands Posts: 1,932 Rep Power: 64	Daniel, >> but when I record from the card's line in, in quiet passages (piano playing) I'm able to hear noise >> in the final mix and that doesn't happen with the livedrive. Ah I see. Yep of course, when that is the case you better not use AC97 based input. (hmmm...probably I dont make music with that much dynamic range, or my ears suck..hehehe) >>That's why I'm thinking that I'd be better of keeping my current setup and waiting till I have more >>money to buy the who audigy 2 pack. Sounds like the best thing to do in your case. >> Well, it just ocurred to me that I can get an Audigy 1 used for a good price and keep the livedrive IR, >> so I can improve playback and keep low noise recording!!! >>That's what I'll do then! Nice alternative solution..! /LeMury

Oct 6, 2003, 11:34 AM	#9
TravelRec. Alternative Audioproductions Join Date: Sep 2003 Location: Germany / Sachsen-Anhalt Posts: 1,710 Rep Power: 0	Hi around here... If you want to improve your inputs and outputs, it´s not really necessary to buy another card. A guy in germany is working on HIGH END Converters from analogue to SPDIF and opposite. Please take a look to: http://www.beis.de/Elektronik/ADDA2496/ADDA2496.html The more the interest, the faster some results, I think... Cheers! TravelRec. __________________ Always take a look behind.

Oct 6, 2003, 01:53 PM	Thread Starter #10
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	For audigy and live sound quality reviews, check 1)Live 5.1 http://www.digit-life.com/articles/sbliveplatinum51/ 2)Audigy Platinum EX: http://www.digit-life.com/articles/a...mex/index.html 3)Audigy 2 family: http://www.digit-life.com/articles2/...gy2/index.html 4)Audigy 2 Platinum Ex: http://www.digit-life.com/articles2/...-ex/index.html __________________ www.palcomp3.com.br/ed

Oct 6, 2003, 05:27 PM	#11
Lex Nahumury DH Senior Member Join Date: Jan 2003 Location: The Netherlands Posts: 1,932 Rep Power: 64	Daniel, This was the post I was looking/refering for/to: http://www.hardwareheaven.com/showthre...threadid=26832 When I got the time I will; - attach an external Signal Generator to Line-In - decrease ac97 faders and - increase external input signal untill analog clipping occurse (Or we could just ask Mattcro's bevindings) It might reveal not so bad specs afterall. /LeMury

Oct 6, 2003, 07:13 PM	Thread Starter #12
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	Yep, maybe the input is amplified and that is causing the bad singal to noise ratio (is that what you mean?). In this review http://www.digit-life.com/articles/sbliveplatinum51/ using creative drivers, they have obtained this results: Test type: External loopback, 48000 Hz sampling, -3dB@1kHz SNR, dB 73.2 input in the card and db 82 They have also done other tests. Please take a look at them. If my previous posts there are links to tests with audigy and audigy2. I don't quite understand those kind of testes and results, but they may shed us some more light on the subject... __________________ www.palcomp3.com.br/ed

Oct 6, 2003, 09:37 PM	#13
Lex Nahumury DH Senior Member Join Date: Jan 2003 Location: The Netherlands Posts: 1,932 Rep Power: 64	Daniel, >>Yep, maybe the input is amplified and that is causing the bad singal to noise ratio >>(is that what you mean?). Well,.more precise but in 'mortal' terms; "a to low analog input signal which (once converted) gets digitaly amplified to the desired recording level" So if one could raise the analog input level up to using ADC's full range, no dsp amplification is needed thus improving SNR (Signal to noise Ratio). In practise; Move the AC97 slider to ~25% and you will see noise level drop to ~75dB. Now apply a test (sine) signal to Line-In (e.g. from your mixer console) and increase it untill clipping occurs at the ADC. (you can hear but not see this) Then back of a little. Now watch your record level. If it is ~ 0dB you have improved your SNR. If it's much lower, you have to turn up the AC97 slider untill record level reaches ~0dB. Now remove test signal and measure noise level. That result is as good as it gets with that particular Codec. That is basicly the test/issue as described in that thread but now in "lamers" terms. (Anybody; Fixme..) Read that thread carefully since it also depends on driver version. Hope this helps a bit, /LeMury Last edited by LeMury; Oct 7, 2003 at 03:59 AM.

Oct 7, 2003, 04:38 AM	Thread Starter #14
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	LeMury, I have just followed your steps and tested it. The sweet spot for my soundblaster live (4.1 CT4760) is setting the AC97 slider to 28.5. That way I achieve -69 to -73, with the mixer faders all the way down, -67 to -69 with the mixer faders outputting the maximum volume without distiortion (I just kept making as much noise as I could with my keyboard and it achieves 0db but nevers surpasses it). That's great! But I have good news! I'll be getting an used Audigy 1 OEM tomorrow, so I'll experience better output quality sound while mantaining my livedrive IR! They guy will send me the audigy via express mail and it'll probably be with me on wednesday or thursday! I'll then do all this tests again to achieve the best SNR. Man, am I learning good stuff from this forum!? Thanks! __________________ www.palcomp3.com.br/ed

Oct 7, 2003, 04:48 AM	Thread Starter #15
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	LeMury, I was also reading the post you recommended and there's something I can't get... Eugene stated the in 3535b he had removed the 4x amplification. If he really did it (and I think he did, comparing both codes of older and newer prologs) I had to able to attain a lower SNR leaving the AC97 slider at 100 and not at 28.5 (where it is set now...). Am I missing something else here??? By the way I have a different AC97 codec than that guy has. Mine is SigmaTel STAC9721/23 and his is STAC9708. __________________ www.palcomp3.com.br/ed

Oct 7, 2003, 05:43 AM	#16
Lex Nahumury DH Senior Member Join Date: Jan 2003 Location: The Netherlands Posts: 1,932 Rep Power: 64	Daniel, Cool,..Congrats with your purchase of the adg1. So you will install both cards right? As for the 4x amplification; I currently haven't any 3535 version installed so I can't check right know. But to be honest, It's not (yet) clear to me if/were and how Eugene removed that 4x amp. I had took a quick look at 3535 SDK source code some time ago, but can't remember finding any indication to this. If it's done in kmixer.exe, there is no way I can tell since it's not open source. I do remember the change in Epilog's output code though. O well, I still have to check 3535c, so I will dig deeper then. And/or maybe Eugene could shed some light on this... However; The same rules still apply. Maybe with other ac97slider settings but who cares as long as the SNR result are improved by properly adjusting the input level vs the ac97 level slider. /LeMury P.S. As a test signal you could simply use a Midi synth/module putting out a non enveloped sustaining sine sound. Makes testing a lot easier than banging on your keyboard while adjusting settings P.P.S Come to think of it; you could also use a sustaining Sine soundfont, play it back and amplify it through your external Mixer and feed it to Line-IN. Right? Duhh..i'm tired. cant think straight anymore..i'll go to bed now...cya..

Oct 7, 2003, 06:03 AM	Thread Starter #17
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	If you right click epilogs from version 3534 and 3535 and click on edit you'll see the differences. 3535 has both versions available. Yep, I could use a sine wave signal but the problem is that when I play more than one note simultaneously, the sound goes louder. This way, if I set the AC97 slider volume using a sine wave, when I do real playing, clipping will occur. So I just made my keyboard make sounds as loud as it can, to never have clipping. I'll probably not use both cards. I need to sell the live to get back part of the money used with audigy. Unless I can make some extra money this week... Will try hard! __________________ www.palcomp3.com.br/ed

Oct 7, 2003, 07:01 AM	#18
Lex Nahumury DH Senior Member Join Date: Jan 2003 Location: The Netherlands Posts: 1,932 Rep Power: 64	>>If you right click epilogs from version 3534 and 3535 and click on edit you'll see the differences. >>3535 has both versions available. yes I know, but if I remember correctly, there was no change in the amplifier code. >>Yep, I could use a sine wave signal but the problem is that when I play more than one note >>simultaneously, the sound goes louder. This way, if I set the AC97 slider volume using a sine wave, >>when I do real playing, clipping will occur. So I just made my keyboard make sounds as loud >>as it can, to never have clipping. No. The right procedure is to determine the 'right' ac97slider setting and never touch it again! (Using a single sustained test signal is really more accurate for this.) Then when you start recording Live stuff (which of course may be louder or softer), you ONLY adjust the output levels from your external Mixer console while watching the Record Vu meters in Cubase for clipping/level check. Done this way you will always achieve a constant and best possible SNR without the risk of clipping. I realy hope you understand this, /LeMury

Oct 7, 2003, 07:22 AM	Thread Starter #19
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	Yes, that really makes sense... I got it. I'll do that them! Thansk once more... __________________ www.palcomp3.com.br/ed

Oct 7, 2003, 12:32 PM	#21
tubus DriverHeaven Junior Member Join Date: May 2003 Location: downtown switzerland Posts: 22 Rep Power: 0	As for external A/D Converters, I use (otherwise broken) DAT-recorders with good results. It doesn't simplify the setup and handling, but I got them for free and it's very easy to add a digital input even to the oldest lives ... tubus

Oct 7, 2003, 01:52 PM	Thread Starter #22
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	Riv@nx, I'm keeping the livedrive... I decided to keep it and get an Audigy instead of selling it to get an Audigy 2. It made more sense to me. __________________ www.palcomp3.com.br/ed

Oct 7, 2003, 06:19 PM	#23
Eugene Gavrilov kX Project Lead Programmer and Coordinator Join Date: Dec 2002 Posts: 3,119 Rep Power: 75	yep, I've removed an (invalid) amplification but it doesn't mean the amplification is impossible in the pre-3535 releases the ac97 signal was amplified, while the others weren't in 3535c and later releases there are no amplified signals, however, the level slider is used for amplification instead that is: level = 0.0 ... 1.0 pre-3535: output(ac97) = inputlevel output(all the rest) = (input level) / 4 (/4 is necessary for any hardware input) post-3535: output = (input * level) thus, you get a possibility to use +12dB gain (if necessary) new epilog / prolog / xrouting are used for that legacy settings should function correctly (but the amplification feature won't be available) new kX Mixer displays the signal levels in dB (if possible) the 'ac97' slider in the kX Mixer is displayed in %% since it can be 'assigned' for WinMM/ Windows Mixer compatibility if you un-assign it, the slider will be displayed in dB /E

Oct 7, 2003, 06:32 PM	#24
Lex Nahumury DH Senior Member Join Date: Jan 2003 Location: The Netherlands Posts: 1,932 Rep Power: 64	Thanks Eugene for clearing that up. /LeMury

Oct 7, 2003, 11:47 PM	#26
Lex Nahumury DH Senior Member Join Date: Jan 2003 Location: The Netherlands Posts: 1,932 Rep Power: 64	Daniel, I just finished some testing with an external signal generator and osciliscope. With 3535c drivers; -Only Prolog AC97 -> Epilog Rec L/R -AC97 slider on 'Ins & Outs' page at 0dB SBLive5.1 OEM Line-In ~-72dB (no time to test adg2 OEM, but probably the same) Quite an improvement I'd say. Well, thanks Mattcro, Max and Eugene...! /LeMury

Oct 10, 2003, 09:13 PM	Thread Starter #27
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	Ok LeMury, but why do you get +-72db testing it that way (with faders at 0db) and Peak plugin "only" gives +-62db when connected straight to prolog's line in outputs? Thanks. __________________ www.palcomp3.com.br/ed Last edited by Daniel Drummond; Oct 10, 2003 at 09:32 PM.

Oct 11, 2003, 05:07 AM	Thread Starter #29
Daniel Drummond DriverHeaven Senior Member Join Date: Dec 2002 Location: Brazil Posts: 1,137 Rep Power: 0	Well, I have just figured out what is going on... I had made a setup with drivers version 3534 and loaded them with drivers 3535rc2. So in the INS AND OUTS pages the AC97 was not show in db and didn't have a little button to set up 0 db. And I could only get that +-72 when I lowered it to number 25. I then cleared the DSP window and started building the same setup from scratch. For my surprise, now the AC97 fader is shown in db, has the button to set it to 0db and I got the exact same values as you did... hehehe Sorry for the mess... Thanks for making me re-check it. __________________ www.palcomp3.com.br/ed

Oct 11, 2003, 03:36 PM	#30
Lex Nahumury DH Senior Member Join Date: Jan 2003 Location: The Netherlands Posts: 1,932 Rep Power: 64	Daniel, >>..For my surprise, now the AC97 fader is shown in db, has the button to set it to 0db and I got >>the exact same values as you did... Yes, In my case, for some reason the AC97 slider keeps switching randomly between dB and numeric readout. No big deal, if set to "No amplification 0db" (cant remember in numeric) then it's fine. (besides; I dont use Prolog, but my own 'Scr' in which amplification is 'hardcoded' at 0dB with no sliders to mess with) Since Eugene fixed the amplification there is no need to tweak anymore. Simlpy setting Line-In (ac97) to 0dB and only adjusting external Analog signal level should give best SNR. For all AC97 based input this would seem to around ~-72dB. (I checked Sblive, adg1 and adg2 AC97 based Line-In. They were all he same) I dont have LiveDrive so I can't tell for I2S<-->Codec based inputs. btw; You also may checkout 'rc3 update files'; http://www.hardwareheaven.com/showthre...5&pagenumber=1 They fix new epilog's -6db output issue. In simple words; Ouput sounds as loud as it used to be (when sliders are at 0dB). /LeMury