• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Review and Measurements of Sound BlasterX G6

Joined
Jan 11, 2019
Messages
79
Likes
78
#41
Can someone correct me if I'm wrong here? Since the volume control is digital, unless you run everything on the computer wide open, isn't the THD+N/SINAD going to be 112dB (and then reducing accordingly for lower volume levels) under any circumstance that is not running it wide open? The low frequency THD rise is thus a nonissue, since it clearly has enough extra bits in the DAC to reduce the digital volume without raising the noise floor. And, even if you do run it wide open, it would only be an issue on recordings that are constantly beating on 0dBFS.

In other words, unless you beat on it, for $150 this appears to be the best PC audio DAC ever tested under $1000--120dB SNR, 112dB SINAD, and vanishing levels of IMD. It's fairly amazing that in actual use, you get performance of a Benchmark DAC3 for peanuts. Aune did about the same thing, but this half the money of even that! If Creative can do this with a portable DAC chip... that does not bode well for expensive standalone DACs. Combine that with their dongle and some of the other dongles punching in around 108dB, and we're pretty much at the point where designing, building and selling (or buying) a DAC that does nothing else is just a waste of time and and money.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,772
Likes
35,907
Location
Seattle Area
#42
In other words, unless you beat on it, for $150 this appears to be the best PC audio DAC ever tested under $1000--120dB SNR, 112dB SINAD, and vanishing levels of IMD.
Before I blew up my Topping D50 in my recent testing, I dialed its output down by a couple of dB and it too shot up to 112+ SINAD. So you can't compare one DAC with output less than 2 volt to another with 2 volt output. The latter may also benefit from being turned down.
 

Jimster480

Major Contributor
Patreon Donor
Joined
Jan 26, 2018
Messages
1,221
Likes
704
Location
Miami
#43
Before I blew up my Topping D50 in my recent testing, I dialed its output down by a couple of dB and it too shot up to 112+ SINAD. So you can't compare one DAC with output less than 2 volt to another with 2 volt output. The latter may also benefit from being turned down.
How did you blow it up?

Edit:
Nevermind I missed that line where it died from the 15V Linear PSU....
RIP
 
Joined
Jan 11, 2019
Messages
79
Likes
78
#44
Before I blew up my Topping D50 in my recent testing, I dialed its output down by a couple of dB and it too shot up to 112+ SINAD. So you can't compare one DAC with output less than 2 volt to another with 2 volt output. The latter may also benefit from being turned down.
That's interesting. At first I wondered if that wouldn't validate another test point at perhaps, -3dBFS, or simple doing a digital level sweep test, and then cataloging the absolute best performance that can be obtained, and where. Then I did a bit of research and decided (once again) that all of this SINAD drag racing is fruitless, and particularly so if you try to develop a standardized test voltage at 2V. Here's why:

Looking through the measurements at RAA, whose measurement database is just insane, shows that the vast majority of DACs all perform within, say, 10dB of each other at -60dBFS. at -60dBFS, almost every DAC has a noise level of around -50dB. Then there is a very, very tiny handful of outliers which are at -55dBFS. Then there's the Oppo HA-1 at -60dBFS, and an Audiolab at maybe -58dB. Same with Chord Hugo. Whatever Audiolab and Oppo did, well, that's some pretty good engineering, but neither are the winner at 0dBFS. But that are better everywhere else. Then you've got the Aune X1s, which will good on paper at absolute levels, have some pretty anamolous results as you back off the digital signal, developing a weird hump (maybe the ESS hump rearing its head in the THD measurements?).

Then let's consider the 2 volt output level... Well, why? Because that was some sort of redbook standard for CD players? Who cares? All Creative or Topping needs to do is find the lowest point of distortion and adjust the product to spit out 2V at precisely that point, thereby "gaming" your SINAD chart. Plus, 1.6V is enough to drive almost any power amplifier into clipping. On top of that, who really cares what happens at 0dBFS when the DAC is being fed by a digital source which uses a digital volume control? Used as intended, such a product may never see a 0dBFS signal! Unfortunately, the less technically inclined will likely never realize any of this.

I know, I know, you can't test for everything... But still, why keep that SINAD chart going when the test voltages and the dBFS levels have often been varied, and you're not even finding the best a product can do? It's all rather silly. A weighted performance metric might be far more useful.
 
Joined
Dec 27, 2018
Messages
10
Likes
6
#45
There is a reduction in the tallest jitter spikes but some remains since the device is ultimately USB powered.
Heh, time to blow up another DAC with external PSU? I do wonder what that'd do to these jitter spikes... but I guess it wouldn't matter.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,772
Likes
35,907
Location
Seattle Area
#46
Then let's consider the 2 volt output level... Well, why? Because that was some sort of redbook standard for CD players? Who cares?
What do you do with vast majority of DACs with no volume control that output 2 volts or thereabouts? That is what is driving my standardization. It is where the industry is as a whole.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,772
Likes
35,907
Location
Seattle Area
#47
Looking through the measurements at RAA, whose measurement database is just insane, shows that the vast majority of DACs all perform within, say, 10dB of each other at -60dBFS. at -60dBFS, almost every DAC has a noise level of around -50dB. Then there is a very, very tiny handful of outliers which are at -55dBFS.
I don't trust such data as far as I can throw it. Who knows what is used to measure these products and the average noise floor of the ADCs used? That may be the limiting factor than what the DACs are doing.

Regardless, I am not sure what your battle is about. I publish tons of data in reviews. Full graphs are provided with respect to levels (IMD) and frequency (THD vs Frequency). My concluding remarks are comprehensive and don't at all rely on a single SINAD number. Excellent SINAD numbers do in general lead to excellence elsewhere. It usually indicates a design that was verified using careful measurements. It is not an accident to be random and separate from other measurements.
 

bennetng

Addicted to Fun and Learning
Joined
Nov 15, 2017
Messages
626
Likes
491
#48
With all the meaningless talk about SINAD not being in the top ten, fixing that ASIO glitch is much more pratical. While ASIO is not a priority for gamers, if it is an advertised feature then it is supposed to work. Otherwise don't advertise that feature.

My Titanium HD has a working ASIO without showing a square wave, but then the SoundFont synthesizer is as glitchy as hell. Luckily with the advancement of virtual instruments that part is unimportant anyway.

Creative has a habit of making powerful DSP chips with poor implementation. Their 20 years old EMU10kx chips transformed into a monster with the 3rd party kX Driver installed. I was so excited about the features and even wrote a whole tutorial about it (in Chinese):

http://www.geocities.jp/anothergs/kXTut/index.html

Just how sad geocities.jp will stop their service after March 31 and my tutorial about this obsoleted chip will become a part of internet history...
 
Joined
Jan 11, 2019
Messages
79
Likes
78
#49
I don't trust such data as far as I can throw it. Who knows what is used to measure these products and the average noise floor of the ADCs used? That may be the limiting factor than what the DACs are doing.

Regardless, I am not sure what your battle is about. I publish tons of data in reviews. Full graphs are provided with respect to levels (IMD) and frequency (THD vs Frequency). My concluding remarks are comprehensive and don't at all rely on a single SINAD number. Excellent SINAD numbers do in general lead to excellence elsewhere. It usually indicates a design that was verified using careful measurements. It is not an accident to be random and separate from other measurements.
They use a proprietary setup which they designed and engineered. It's documented on the site. I have found that their measurements and third party measurements tend to align well, and have spent a number of hours comparing them. I find them particularly useful because each product has hundreds of measurements all saved and easily accessible--THD, multitones, single tones, CCIF, digital input level, voltage output level, SMPTE. It is simply a tremendous resource for understanding product behavior in detail, and teasing out what is meaningful and what is not.

I don't have any particular battle, except my ongoing complaint about the emphasis on THD+N at either 2V or whatever maximum level can be squeezed out, if it won't put out 2V. THD+N is a single very lossy measure of a device's performance. Unfortunately, it's also very easily digestible and misleading (especially when transposed to the mysterious SINAD). See, e.g., https://superbestaudiofriends.org/i...-of-6-amps-thd-n-and-thd-specifications.6194/. That only resurfaced here because of one of the peculiarities of the product, which was a significant increase when the level was reduced, for a component that will likely never be operated at maximum.

How many $99 "finish it yourself" Khadas toneboards have been sold for no good reason other than it's the cheapest thing with the highest number? On the other hand, this review will probably sell zero SoundBlasterX G6s because it isn't as "good" despite the fact that when used as intended and the way most people will use it--plugged into a computer or phone, not using ASIO or WASAPI, and not cranked to maximum volume--it is actually better. And it comes with a case! And drivers! And a fully discreet headphone amp! And you're not tethered to a wall socket as with a discreet Khadas/amp setup! And you don't get top octave hash of the Khadas! Most people won't notice any of that, though. They'll just notice that the SINAD is lower than a Toneboard, and that when run wide open (which will never occur in actual use) "we are talking nearly 1% THD+N!!!" Even though almost no one actually using this thing will ever have that happen, and measurements elsewhere where it does not happen. Instead, they will get the benefit of IMD that blows away the Toneboard and equals a Benchmark DAC3 in actual use, lower distortion than the Toneboard, a product from a reputable manufacturer, actual drivers designed by the manufacturer, and on, and on...

It's a matter of presentation, I suppose. What you seem to despise is even the slightest technical fault, and when there is one, it tends to be followed with lots of exclamation points and whatnot, instead of an explanation of whether it actually matters. I agree that technical faults are unfortunate, but in many instances they really don't matter.

Plus, I think its more fair to point out flaws in presentation and measurements here than simply griping about it in another forum as many are prone to do. Tone and presentation are critical. Years ago, when I was doing what you are doing with test equipment no one else had, I was a bit dismayed looking back years later. What I had intended the take-away to be was not what it actually was. Ah, well, enough said. Carry on with the good measurements. :)
 
Last edited:

BYRTT

Active Member
Patreon Donor
Joined
Nov 2, 2018
Messages
194
Likes
205
Location
Denmark (Jutland)
#50
...How many $99 "finish it yourself" Khadas toneboards have been sold for no good reason other than it's the cheapest thing with the highest number?...
While i get your points and can see G6 is quite good can't you see the big problem that G6 so far isn't very versatile and one can only get guranteed good numbers over that TOSLINK input, myself curios asked Amir for 32 bit tone test and he couldn't run it because that TOSLINK didn't go that high in rate. Now look how Khadas TB whether using SPDIF or USB input seems run flawless bitperfect on Win7/Win10 and various Linux distributions and probably the point to be popular, that versatility and that it works for bitperfect playback is the thing that Amir's analyzer could extract all the good graphs and also reason myself shopped one Khadas TB. About G6 so far my personal take is why in heaven get a unit that advertise versatility but when tested one is locked to that TOSLINK input else forget about bitperfect good numbers and also reminds me how that brand probably for financial reasons or what do i know think its normal not support that various operating systems is a dynamic developing thing that needs some care down the road so that customers is not bricked on a particular operating system with a particular non perfect but less buggy revision of a Sound Blaster software package.

...And you don't get all of the high frequency hash on the Khadas! Most people won't notice any of that, though...
Is argument the HF hash below then think take into acount that when boxed as for dac1 at right it gets less and also input for Khadas TB/MARCH audio dac1 is USB that for most normal cases use to be few dB worse than SPDIF input and G6 for below case actual use a galvanic isolated TOSLINK input to transport SPDIF.

1000.png


...Instead, they will get the benefit of IMD that blows away the Toneboard, lower distortion than the Toneboard, lower IMD that the DX3Pro, a product from a reputable manufacturer, actual drivers designed by the manufacturer, and on, and on...
Would love it but then please get them fix stuff so that versatility is on par expected or advertised, and don't forget who knows if also Khadas TB will improve numbers dialing down its input a few dBs.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,772
Likes
35,907
Location
Seattle Area
#54
hi,amirm.
thd+n in here is 0.002% , is that blasterx G6 work worse in 48khz?
This test uses 90 kHz of bandwidth for noise and distortion. The dashboard uses a bit more than audible band (22.4 kHz). By definition including more bandwidth hurts DAC (or any audio device) performance. Sometimes a little, sometimes a lot depending on how much ultrasonic junk there is.
 

Conroe

New Member
Joined
Apr 13, 2019
Messages
1
Likes
0
#56
Hi, I'm new here. Thank you for the comprehensive review for the G6. Currently, I am using the FiiO Olympus 2 (E10K). Would the Sound Blaster be a worthwhile upgrade for me?

I appreciate your thought on this.
 

Veri

Major Contributor
Patreon Donor
Joined
Feb 6, 2018
Messages
3,014
Likes
2,986
#57
Hi, I'm new here. Thank you for the comprehensive review for the G6. Currently, I am using the FiiO Olympus 2 (E10K). Would the Sound Blaster be a worthwhile upgrade for me?

I appreciate your thought on this.
If your older E10K can do everything you want, there is little reason to upgrade. Especially if you want more headroom/volume, the G6 is unlikely to offer it. Better to use line-out and get something like a JDS Atom for more volume capabilities.

If you want more of the same then yes, the G6 is a relatively low-power but marginally better device than the E10K. You can always purchase it and send it back if you don't think it is enough of an upgrade.
 
Joined
Apr 13, 2019
Messages
10
Likes
0
#58
If your older E10K can do everything you want, there is little reason to upgrade. Especially if you want more headroom/volume, the G6 is unlikely to offer it. Better to use line-out and get something like a JDS Atom for more volume capabilities.

If you want more of the same then yes, the G6 is a relatively low-power but marginally better device than the E10K. You can always purchase it and send it back if you don't think it is enough of an upgrade.
greetings, I am new to the forum, I have a question about the line output of the sound blaster x g6, is a line output to connect amplifiers auriculres, or is only a line output to connect active speakers, thank you for responding.
 

Veri

Major Contributor
Patreon Donor
Joined
Feb 6, 2018
Messages
3,014
Likes
2,986
#59
greetings, I am new to the forum, I have a question about the line output of the sound blaster x g6, is a line output to connect amplifiers auriculres, or is only a line output to connect active speakers, thank you for responding.
Hi. line output is meant to be amplified, by amplifier or active speakers. The "line level" just means it is supposed to be passed through, as opposed to a pre-amp/variable output which is "active" in the sense that it is a line level that already has been amplified.

Some devices don't have line out, then you can maximise output and use the volume control on the next device in the chain, but if there is line level output it is better for fidelity reasons to use the line out :)
 
Joined
Apr 13, 2019
Messages
10
Likes
0
#60
Hi. line output is meant to be amplified, by amplifier or active speakers. The "line level" just means it is supposed to be passed through, as opposed to a pre-amp/variable output which is "active" in the sense that it is a line level that already has been amplified.

Some devices don't have line out, then you can maximise output and use the volume control on the next device in the chain, but if there is line level output it is better for fidelity reasons to use the line out :)
Hi, thanks for answering, what happens I read in some forums that the line out of the G6 suffers cuts, I also understand that the line output of the G6 can be controlled with the volume knob of the G6, is that true. Thank you.
 
Top Bottom