IPB

Welcome Guest ( Log In | Register )

3 Pages V   1 2 3 >  
Reply to this topicStart new topic
Onboard vs Cheap USB SoundCard?
Tall-Guy
post May 24 2012, 13:08
Post #1





Group: Members
Posts: 143
Joined: 22-November 05
Member No.: 25923



Hey,

Up until now I have been using my Onboard Sound Card - ALC1200(Realtek). It suppose to be quite decent according to reviews (I'm mainly using it with Headphones).
More technical information about it can be found here.

Recently I purchased a set of new headphones that comes with a USB Soundcard: Steelseries USB Dongle. Technical information can be found here.

Few questions based on the above:
1. Is any one of them significantly better than the other (spec-wise)?
2. Does Onboard card getting Amp on the way out by the motherboard? What about the USB soundcard?

I tried to compare them using my ears, but the time it takes me to unplug and plug everything again (and change the output device in Windows) just takes too much for me to 'Remember' how to previous sounded. Or perhaps it just the difference was not that major for me to notice.

Thanks.
Go to the top of the page
+Quote Post
pdq
post May 24 2012, 13:38
Post #2





Group: Members
Posts: 3372
Joined: 1-September 05
From: SE Pennsylvania
Member No.: 24233



Without looking at the specs I would guess that both are more than adequate for your purposes, with one possible difference. With no music playing do you hear any extraneous noise at all? This sometimes happens with internal sound cards, but if you don't hear any then I would stick with the internal card.
Go to the top of the page
+Quote Post
Brand
post May 24 2012, 15:45
Post #3





Group: Members
Posts: 314
Joined: 27-November 09
Member No.: 75355



Do a RMAA test with both of them. Also using different inputs (when you test the onboard output test it also with the Steelseries input in addition to the onboard input and vice versa), so that you can see which one is the bottleneck.
I think you can even test the RMAA performance with the headhpones load if you use a splitter on the output and plug the headphones in. Headphones tend to influence the frequency response, crosstalk etc.

Whether these tests translate into audible differences is, of course, a different question.
Go to the top of the page
+Quote Post
DVDdoug
post May 24 2012, 17:58
Post #4





Group: Members
Posts: 2535
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE
I tried to compare them using my ears, but the time it takes me to unplug and plug everything again (and change the output device in Windows) just takes too much for me to 'Remember' how to previous sounded. Or perhaps it just the difference was not that major for me to notice.
You are right... That's not the best way to do an AB (or ABX) test. It's also very important to match levels.

But if you can't hear an OBVIOUS difference, does it matter? If you use one sound card one day and a different sound card the next day and you don't hear a difference, does it matter?

Careful, scientific, double-blind ABX tests are important to prove that there is an audible difference when people think they hear a difference. It's not so important to prove that you don't hear a difference. Think about that... If I claim that I don't hear a difference, can you prove that I actually do hear a difference? biggrin.gif

I agree with pdq. If you hear a defect from a soundcard, it's usually noise. There's noise in every analog signal, but as long as it's below audibility in both cards you obviously won't hear a difference.

If the impedance is too high, you can get frequency response variations with the (relatively) low-impedance headphone-load attached.

If you like to listen loud, some soundcards put-out a higher level than others.

Go to the top of the page
+Quote Post
AndyH-ha
post May 25 2012, 07:34
Post #5





Group: Members
Posts: 2205
Joined: 31-August 05
Member No.: 24222



Not to create any big argument, but I see the most important use of ABX tests as being to decide that I really can't hear a difference when I know differences exist. Audio processing software often provides more than one choice for how to handle a problem. Quite often one, or more, of those ways is much less labor intensive than another.

Is the easy way good enough? Especially if I'm feeling lazy, can I depend on a simple AB comparison? If I don't want to have to take the hard way, it can sometimes be too easy to convince myself there isn't any real difference in results. Sometimes that is true, sometimes it is not.

As far as "does it matter?" goes, it may be not unlike those people who say "I can't really hear any difference at X quality level ... but I feel better if I use the higher quality Y settings. Even though a simple AB test doesn't seem to show me any difference, I'm frequently not satisfied with the easier approach if an ABX test says I can tell the sample apart, even if I'm not quite sure why I can.
Go to the top of the page
+Quote Post
probedb
post May 25 2012, 08:51
Post #6





Group: Members
Posts: 1194
Joined: 6-September 04
Member No.: 16817



QUOTE (DVDdoug @ May 24 2012, 17:58) *
But if you can't hear an OBVIOUS difference, does it matter? If you use one sound card one day and a different sound card the next day and you don't hear a difference, does it matter?


I think that's actually all that needs to be said smile.gif
Go to the top of the page
+Quote Post
Garf
post May 25 2012, 09:50
Post #7


Server Admin


Group: Admin
Posts: 4883
Joined: 24-September 01
Member No.: 13



I'd second the advice to do an RMAA test. The quality of both is probably good enough that an ABX won't tell you anything. And yes, this does mean that it doesn't actually particularly matter and you can pick for convenience/features.

That said, the Steelseries USB thing looks like it's more meant to mangle the sound than anything (and that might be detectable in ABX).

If you do the RMAA, get RMAA 6.2.3, not 6.2.4 as the latter gives wrong results. Make sure the output levels for whatever device you are testing are set as high as possible, and input level low enough that it just doesn't clip in RMAA. Also make sure both playback and recording use the same format in the Windows Sound control panel (not only in RMAA), preferably 48kHz 24-bit.

The RMAA you posted for the ALC1200 isn't particularly good for a modern, built-in soundchip, so I suspect it might not have been done in optimal conditions. (Unfortunately I've seen shitloads of mainboard reviews with exactly the same problem sad.gif)
Go to the top of the page
+Quote Post
Tall-Guy
post May 25 2012, 10:31
Post #8





Group: Members
Posts: 143
Joined: 22-November 05
Member No.: 25923



I'm quite new to this, so let me know if I got that right:

I downloaded the 6.2.3 version as suggested and I'll start testing the Onboard Card First. I removed the Mic as I mainly care about the Out quality then In, and I prefer not to make everything more complex.

Few questions:

1. Should I plug my headphones to the 'Speakers' jack or to the dedicated Headphones Jack? or it shouldn't matter? Will it be OK to have both of them connected (Speakers and Headphones)
2. I set my Output to max (100) and disabled everything else (Mic, CD Audio, Line in). You mentioned the input level should be 'Low as possible' to avoid clipping. Does it mean I should enable those and set them to the lowest possible? (0 in the scale?)
3. The format set on my Onboard card is 24/192Khz, should I lower this to 24/48Khz? and if so, why?
4. Im using the Default Test-Scheme, is that ok?
5. I'm not sure if I'm using is Direct Sound or MME. I remember I was able to choose that prior to Windows 7 though.

When I select DirectSound with my Device (and MME Wave mapper is my only option on the recording), set 24/48Khz. Am I good to go now by clikcing 'Playback' only?
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post May 25 2012, 11:18
Post #9





Group: Members
Posts: 3647
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (Tall-Guy @ May 25 2012, 05:31) *
I'm quite new to this, so let me know if I got that right:

I downloaded the 6.2.3 version as suggested and I'll start testing the Onboard Card First. I removed the Mic as I mainly care about the Out quality then In, and I prefer not to make everything more complex.

Few questions:

1. Should I plug my headphones to the 'Speakers' jack or to the dedicated Headphones Jack? or it shouldn't matter? Will it be OK to have both of them connected (Speakers and Headphones)


Normally tests are done without anything connected but the necessary signal cables for running the test. No headphones, no speakers.

QUOTE
2. I set my Output to max (100) and disabled everything else (Mic, CD Audio, Line in). You mentioned the input level should be 'Low as possible' to avoid clipping. Does it mean I should enable those and set them to the lowest possible? (0 in the scale?)



Try some different output settings and see what works best. RMAA forces you to set the input control right. No clipping!

QUOTE
3. The format set on my Onboard card is 24/192Khz, should I lower this to 24/48Khz? and if so, why?


While some onboard cards will do *something* with a 24/192 file, I know of very few that will actually record a signal that high. Sample rates > 16/44 generally have no audible benefits. If an audio interface works better at some sample rate, its probably the lower one.


QUOTE
4. Im using the Default Test-Scheme, is that ok?


Yes.

QUOTE
5. I'm not sure if I'm using is Direct Sound or MME. I remember I was able to choose that prior to Windows 7 though.


Shouldn't matter if RMAA works at all.

QUOTE
When I select DirectSound with my Device (and MME Wave mapper is my only option on the recording), set 24/48Khz. Am I good to go now by clikcing 'Playback' only?


AFAIK there is no such thing as Directsound for recording. Its a playback-only feature. Make sure that all of your format settings agree.
Go to the top of the page
+Quote Post
soulsearchingsun
post May 25 2012, 11:18
Post #10





Group: Members
Posts: 145
Joined: 27-January 05
Member No.: 19370



QUOTE (Garf @ May 25 2012, 09:50) *
If you do the RMAA, get RMAA 6.2.3, not 6.2.4 as the latter gives wrong results.

Could you please point me to a resource further explaining this problem? I'm having problems telling the forum search engine to look for the relevant thread, if it exists.

Edit: There is a bug-fix release available as of May 02 2012. Could it already resolve this issue?

This post has been edited by soulsearchingsun: May 25 2012, 11:23
Go to the top of the page
+Quote Post
Brand
post May 25 2012, 11:20
Post #11





Group: Members
Posts: 314
Joined: 27-November 09
Member No.: 75355



QUOTE (Garf @ May 25 2012, 10:50) *
get RMAA 6.2.3, not 6.2.4 as the latter gives wrong results.

What makes you say so? I did a quick test now and I got pretty much the same results.


Tall-Guy: generally, you don't need to do much from the default.
Connect the output to an input with a stereo cable. Select the in and out in RMAA and the sample rate and bit depth (also in Windows/driver settings) and then run Playback/recording. Adjust the levels if needed (it will tell you when it's ok), keeping in mind that highest software volume out should give the best result.
Connecting the headphones with a splitter to the tested output is optional, but if you're using headphones it makes sense to do it.

This post has been edited by Brand: May 25 2012, 11:35
Go to the top of the page
+Quote Post
Tall-Guy
post May 25 2012, 14:35
Post #12





Group: Members
Posts: 143
Joined: 22-November 05
Member No.: 25923



Well, everything just become much easier. I was boosting the volume all the way and I can hear annoying 'crackling' from the headphones Jack. That must be what you guys warned me about. Moreover, I found this (the numbers are being showed at 1:50) - and it seems the USB is pretty decent. I guess i'll save me some RMAA education and stick with the USB.

One last question though, I have a three places to adjust the volume:
1. in Windows
2. On the USB Dongle itself
3. On the line between the 3.5 jack and the headphones themselves.

I should put the Windows and USB Dongle volume at max and just adjust the volume using the knob on the headphone?
Go to the top of the page
+Quote Post
Garf
post May 25 2012, 14:36
Post #13


Server Admin


Group: Admin
Posts: 4883
Joined: 24-September 01
Member No.: 13



QUOTE (Tall-Guy @ May 25 2012, 11:31) *
1. Should I plug my headphones to the 'Speakers' jack or to the dedicated Headphones Jack? or it shouldn't matter? Will it be OK to have both of them connected (Speakers and Headphones)
2. I set my Output to max (100) and disabled everything else (Mic, CD Audio, Line in). You mentioned the input level should be 'Low as possible' to avoid clipping. Does it mean I should enable those and set them to the lowest possible? (0 in the scale?)


You need a cable that connects speaker out (or line out, should be the same) to line in. You should enable line out for playback at maximum volume (without amplification) and line in (and *only* line in) for recording. RMAA will allow you to tune the volume when you start the test, do this by gradually increasing the line in volume until it just doesn't clip.

QUOTE
3. The format set on my Onboard card is 24/192Khz, should I lower this to 24/48Khz? and if so, why?


Going higher than 48kHz is pointless but it shouldn't hurt either. Note that you not only need to set this in RMAA but also in the Windows Control Panel Sound configuration, for both recording and playback.

In Windows 7 this means Control Panel -> Sound -> Playback -> [device] -> Properties -> Advanced. Also make sure in Enhancements everything is disabled.
Then do the same for Sound -> Recording.



Go to the top of the page
+Quote Post
Garf
post May 25 2012, 14:46
Post #14


Server Admin


Group: Admin
Posts: 4883
Joined: 24-September 01
Member No.: 13



QUOTE (soulsearchingsun @ May 25 2012, 12:18) *
QUOTE (Garf @ May 25 2012, 09:50) *
If you do the RMAA, get RMAA 6.2.3, not 6.2.4 as the latter gives wrong results.

Could you please point me to a resource further explaining this problem? I'm having problems telling the forum search engine to look for the relevant thread, if it exists.

Edit: There is a bug-fix release available as of May 02 2012. Could it already resolve this issue?


RMAA 6.2.4 has (or had) a bug that causes it to mark down every frequency response to "Very Poor", pretty much regardless of the actual result, i.e. even if its almost perfectly flat. It's possible the hotfix solves that as it appears to be published after I last used it.
Go to the top of the page
+Quote Post
Tall-Guy
post May 25 2012, 15:10
Post #15





Group: Members
Posts: 143
Joined: 22-November 05
Member No.: 25923



QUOTE
You need a cable that connects speaker out (or line out, should be the same) to line in. You should enable line out for playback at maximum volume (without amplification) and line in (and *only* line in) for recording. RMAA will allow you to tune the volume when you start the test, do this by gradually increasing the line in volume until it just doesn't clip.


It basically means however I can not perform this test on the External USB Card? as It doesn't have Line in, but only out (and a USB).
Go to the top of the page
+Quote Post
Garf
post May 25 2012, 15:17
Post #16


Server Admin


Group: Admin
Posts: 4883
Joined: 24-September 01
Member No.: 13



QUOTE (Tall-Guy @ May 25 2012, 15:35) *
Well, everything just become much easier. I was boosting the volume all the way and I can hear annoying 'crackling' from the headphones Jack.


You mean that your inboard sound starts to distort at maximum volume? Try the test with the line out, I'm guessing that when it's set in headphone mode it probably tries to amplify if you increase the volume and that is causing it to fail.

QUOTE
Moreover, I found this (the numbers are being showed at 1:50) - and it seems the USB is pretty decent.


I'd be somewhat suspicious of that "review" [1] for several reasons:

a) All cards got exactly the same result. (within measurement errors)
b) The numbers are insanely good. I didn't even know it was electronically possible to get -133dB SNR out of a DAC unless the card is cooled to 0 Kelvin or something. Let alone a noise level of -200dB. Did they put it in a lead box to shield it from background earth radiation?

He's effectively saying that the SteelSeries USB is better than the RealTek HD (whatever model isn't mentioned) because it has -205.6dB stereo seperation instead of -204.1dB.

To explain how out of whack (or ridiculous) this is, allow me to compare. Stereo crosstalk for high end audio cards is usually around the 100-110dB mark. That is, when tested properly by someone who understands what they're doing, of course. Now, the 205dB result posted doesn't mean the card is twice as good. That's not how decibels work. This means the card would be about 57 THOUSAND times better.

My best guess at what happened here is that he used DIGITAL loopback, i.e. the DAC/ADC (essentially the soundcard) were not involved at all, and he basically just copied the audio data to memory and back. Which obviously gives exactly the same result regardless of the soundcard. This is a pointless test of course as you always need a DAC to convert the signal to something audible.

So, do yourself a favor and run a proper test. You're not going to reach the numbers posted. Probably not by a factor of 65000 smile.gif
Go to the top of the page
+Quote Post
Garf
post May 25 2012, 15:25
Post #17


Server Admin


Group: Admin
Posts: 4883
Joined: 24-September 01
Member No.: 13



QUOTE (Tall-Guy @ May 25 2012, 16:10) *
It basically means however I can not perform this test on the External USB Card? as It doesn't have Line in, but only out (and a USB).


You can loop back the line output from the USB card into the line in of another soundcard. This will limit the maximum possible result to the performance of the second soundcard, though, so you'd typically want to use the best you can get your hands on and RMAA that first.

If you only have the built-in audio on your mainboard, you can RMAA that, and then use this set up to see if you get approximately the same result when you use the USB card line output into the mainboard line input. If the results are close, this means the USB is "no worse" than the mainboard.

So you can only see if the USB is worse or not worse. You can't measure if it is better.

Go to the top of the page
+Quote Post
Tall-Guy
post May 25 2012, 15:42
Post #18





Group: Members
Posts: 143
Joined: 22-November 05
Member No.: 25923



Ermm, Im getting:
"PROBLEM: The input level is low. Try to increase recording or playback levels in your mixer."

My Line in volume is at max at the moment.
The setup:
Nothing connected to the onboard card besides a 3.5 Jack cable with one leg in the Line in, and the other side in line out (Speakers). Am I missing something?
Go to the top of the page
+Quote Post
saratoga
post May 25 2012, 15:58
Post #19





Group: Members
Posts: 4854
Joined: 2-September 02
Member No.: 3264



QUOTE (Tall-Guy @ May 25 2012, 10:42) *
Ermm, Im getting:
"PROBLEM: The input level is low. Try to increase recording or playback levels in your mixer."

My Line in volume is at max at the moment.
The setup:
Nothing connected to the onboard card besides a 3.5 Jack cable with one leg in the Line in, and the other side in line out (Speakers). Am I missing something?


Either the volume isn't actually at max, or you're attenuating the line in volume in Windows.
Go to the top of the page
+Quote Post
Tall-Guy
post May 25 2012, 16:17
Post #20





Group: Members
Posts: 143
Joined: 22-November 05
Member No.: 25923



Found the issues. Next one :-)
PROBLEM: Inter-channel leakage in the recorded signal. Possible cause: connection cables, swapped channels.

By the way, It seems that on the LIne in I can only set 16 Bit and not 24 like the LIne out. Is that OK?
Go to the top of the page
+Quote Post
Tall-Guy
post May 25 2012, 17:32
Post #21





Group: Members
Posts: 143
Joined: 22-November 05
Member No.: 25923





Thats the Onboard.
Go to the top of the page
+Quote Post
Garf
post May 25 2012, 17:41
Post #22


Server Admin


Group: Admin
Posts: 4883
Joined: 24-September 01
Member No.: 13



It's maybe farfetched, but I hope it's a stereo cable, and it's plugged in correctly?

QUOTE
By the way, It seems that on the LIne in I can only set 16 Bit and not 24 like the LIne out. Is that OK?


It will limit the best possible SNR results to 96dB, which reasonable onboard chips are able to reach.
Go to the top of the page
+Quote Post
Garf
post May 25 2012, 17:44
Post #23


Server Admin


Group: Admin
Posts: 4883
Joined: 24-September 01
Member No.: 13



QUOTE (Tall-Guy @ May 25 2012, 18:32) *
Thats the Onboard.


Not bad, looks reasonable for a modern but relatively low end soundchip. Note that it's already 6dB better than the result you posted earlier. (i.e. twice as good)
Go to the top of the page
+Quote Post
Tall-Guy
post May 25 2012, 17:50
Post #24





Group: Members
Posts: 143
Joined: 22-November 05
Member No.: 25923



Well, it's a 3.5 jack in both sides. Is that a Stereo cable? :-)
I just tried the USB Sound Card. I connected the Headphones slot (Out I guess) into the same Line in in the previous test (my Onboard Card). For some reason the Recording Level is very low, around 27dB, and that's with the highest recording level possible. Is USB sound card being handed differently?
Go to the top of the page
+Quote Post
Tall-Guy
post May 25 2012, 18:53
Post #25





Group: Members
Posts: 143
Joined: 22-November 05
Member No.: 25923



OK. I'm recapping everything and adding information about the USB Sound Card.

Onboard Card:
http://i49.tinypic.com/16abfhl.jpg

I wasn't able to level both right and left channels perfectly, but the difference was not that far away

Steelseries USB:
It gap between the Right and left channels were pretty big this time (around 2-3dB). In the Device settings (In Win7) there is an option called "Listen to Device" (under "Listen"). It playback the Line in back to the USB Soundcard. When this option was being checked, I wasn't able to reach -1dB. My only options were -3dB, or over clipping above 0dB. I decided not to over-clip and run the test with -3-7dB:

http://i48.tinypic.com/24pb3n6.jpg

With "Listen to Device" checked, with everything maxed out I was only able to reach to -5dB (same gap between channels again), but It the Recording Volume kept jumping Between -5dB and -26dB for some reason. Results:

http://i46.tinypic.com/34zzwv8.jpg
Go to the top of the page
+Quote Post

3 Pages V   1 2 3 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 26th July 2014 - 11:48