Creative Sound Blaster G3

[di2ger]

Hi guys,
I wanted a small device for digital/analog audio measurements so I bought Creative Sound Blaster G3.
It has Line-in and mini-Toslink Optical inputs.
I used Topping D10s and REW to do some testing:

Line-in performance is quite underwhelming, but what surprised me is that the digital input has a similar issue with low-frequency noise.
I've reached out to Creative support to see if this can be fixed.

 

[AnalogSteph]

Line-in performance is quite underwhelming,

Well, for starters, not hopelessly overdriving the input would help. Check input gain. Perhaps it can't actually take any more either.

The hardware is also clearly not running at 24/96, looks more like 16/48 analog and 24/48 digital to me (but I don't know how good the ADC is - dynamic range spec is only 100 dB and the input side tends to be worse than that, so about 16-bit performance may be all you can hope for).

but what surprised me is that the digital input has a similar issue with low-frequency noise.

That is odd indeed. If that's a Windows machine, I would be looking for rogue APOs in recording device settings (use the old dialog you can reach if you go looking for system sound schemes).

I've reached out to Creative support to see if this can be fixed.

Good luck with that. Last time I tried to report a bug, they only seemed to have dumb user support, with no escalation even close to development. Mind you, that was in 2014.

 

[di2ger]

I appreciate your response, AnalogSteph, it's very helpful.

Check input gain.

Oh, my! It wasn't nice of me at all . Sorry, G3, I didn't mean to overfeed you.
Here is the updated graph:

Line-in looks much better now, with THD: -92.6 dB and THD+N: -56.4 dB. The second value is still quite underwhelming due to the low-frequency noise issue.

rogue APOs

Sorry, but what are the rouge APOs? I used Windows and Mac with the same results.

Good luck with that

Thanks. So far, so good:

We will review your case further and get back to you.
We appreciate your kind understanding and patience during this time.

 

[di2ger]

Also, I can't stop watching the sweep:

 

[AnalogSteph]

Line-in looks much better now, with THD: -92.6 dB and THD+N: -56.4 dB. The second value is still quite underwhelming due to the low-frequency noise issue.

The apparent ground loop in the measurement setup isn't helping either. You should do this between 2 computers, at least one of which being battery-operated (and hence floating). Otherwise you'd need a USB isolator. More people have a spare laptop floating around than one of those.

Which generator level are you running? I presume you were using RMAA's generator, which is set to -12 dBFS by default. If that gives you near 0 dBFS on the input, input gain may still be cranked a fair bit. I would generally expect a Creative device to be accepting about 2 Vrms.

Seeing that noise floor, I can't help but think "noise cancelling" somehow. Hmm...

You'll still have to figure out how to set up the sample rate properly, too. It needs to be done in RMAA and on OS-level. (If you are lucky, that's it. Creative and Asus Xonar and other cards of yore also needed hardware sample rate to be set manually in the manufacturer's control panel app.) Also note that REW on Windows can't do more than 16-bit samples with the default Java I/O, you need to be using ASIO (potentially via ASIO4All) for that.

 

[di2ger]

The apparent ground loop

Wow, this is something I wanted to avoid. AFAIK the ground loop can happen with the line-in measurements, and not with optical Toslink, as it electrically isolates the source and the destination. I am using just one laptop at a time, normally Mac. The video I posted is about digital input, so I didn't expect any ground loop issues.
BTW, how one can see a ground loop on a frequency response graph?
To be sure I used my phone as a 1kHz signal source for G3 line-in with very similar results.

Which generator level are you running?

Normally at 0 dBFS, but then I adjust the output and input gain to get the best THD+N value. Basically, for each output level I am searching for the maximum input gain, which still doesn't clip. Then I pick the best THD+N value. The reason I'm not using the generator level all that much is that I prefer analog volume control to digital one, hoping that the analog will give me better performance.

You'll still have to figure out how to set up the sample rate properly

Sure, this is the first thing I check, what options OS gives for my devices and then I set them to the same as in REW Soundcard's Preferences.

REW on Windows can't do more than 16-bit samples with the default Java I/O

Oh, wow, that's too bad. Thanks, I didn't know any of that. For measurements I definitely need at least 24 bits. BTW, on Mac I don't see any improvement between 24 and 32 bit test files:

Maybe REW on Mac can deliver only 24 bits?
About G3 sample rate:

The hardware is also clearly not running at 24/96, looks more like 16/48 analog and 24/48 digital to me

As I see in the sweep, these 96kHz are plain deception. Nothing good is happening above 24kHz, only noise, aliasing and reflections. And that'd be ok, if it just stayed above 24kHz, but it goes back to the audible range, and quite loudly. I need to test if, at the 48kHz sample rate, they have a proper low-pass filter at the inputs.

 

[AnalogSteph]

BTW, how one can see a ground loop on a frequency response graph?

Well, not on that, but in the spectrum. The slew of power supply harmonics coming up to about -80 dBFS starting at 120ish Hz is definitely suspect.

Oh, wow, that's too bad. Thanks, I didn't know any of that. For measurements I definitely need at least 24 bits.

It's not a real biggie, since you can just switch to ASIO after installing and setting up ASIO4All, but it's definitely something to be aware of.

As I see in the sweep, these 96kHz are plain deception. Nothing good is happening above 24kHz, only noise, aliasing and reflections. And that'd be ok, if it just stayed above 24kHz, but it goes back to the audible range, and quite loudly.

Something is definitely set up wrong. I can see that the ADC is only running at 48 kHz for sure. You'll need to change that. Both Windows and MacOS have corresponding control panel settings (on Windows, find the old dialog with system sounds and stuff, on MacOS go to Audio / MIDI setup).

 

[di2ger]

Good news! Creative answered the following:

Please use Digital Audio Interfaces/What U Hear to record SPDIF In to see if it helps.

In order to record Line In or Optical In through Digital Audio Interfaces/What U Hear, kindly enable Monitoring for SDPIF Interface or Line in Monitoring section.

So there is an issue and a workaround, which is good.
This method helped remove the low-frequency noise, but I see that it is still not bit-perfect. I'll have to ask the support about that as well.

Something is definitely set up wrong.

Yes, I'm afraid it's on the hardware side. I need to ask the support about it as well. On my side it is correct:

I will try to measure the DUT again soon and update the thread.

 

[AnalogSteph]

And both the input and output sample rate are set up correctly?

I am not sure why you are recording external inputs through What U Hear in the first place. The What U Hear recording input is a special tap-off on the output signal pipeline so you can record things played back there, e.g. a mix of both yourself and a person on the other end of a remote session, or the internet radio station that you are currently playing. If you are recording the line-in like that, the signal is basically recorded, then immediately played back and the playback signal is then diverted into What U Hear to be recorded - a very circuitous route.

Use of this feature should be reserved for those circumstances that actually require it. I would not be surprised if sample rate had to align in no less than 3 places: The actual input itself, What U Hear and the output.

 

[di2ger]

So here are the measurements, taken from "What U Hear input", as advised by Creative Support:

Here are the THD+N values [10..22400 Hz]:
Source 96kHz/24bit test 1kHz file: -138.4 dB.
Digital input: -129.6 dB (8.6 dB worse).
Analog input: -70.2 dB (68.2 dB worse).

And both the input and output sample rate are set up correctly?

Yes.

I am not sure why you are recording external inputs through What U Hear in the first place.

Sorry for not being clear about it. My first and second measurements were taken directly from appropriate inputs, muting all but one input and disabling all features. I didn't like the results and asked Creative support how to fix it. They said to use the "What U Hear" thing with monitoring enabled in the mixer. So I did and the results are above. Much better, but still not very happy.

Now it's time to ask the support where is everybody after 24 kHz, where is bit-perfect input and when it's all going to work properly, and not through... the backdoor.

 

[AnalogSteph]

This thing is pretty quirky indeed then. The digital input has jitter components in it, so there must be an ASRC in there somewhere, too. Clearly there's more going on than meets the eye, and the firmware may not entirely be up to the task.

What exactly is it that you were you hoping to measure? The G3 sure looks like a bust for now.

 

[di2ger]

What exactly is it that you were you hoping to measure?

I wanted to use the digital input for digital audio measurements, as my external USB sound card with Toslink SPDIF input doesn't work in 80% of the cases for whatever reason.

The G3 sure looks like a bust for now

Yeah, I'm a bit disappointed. It looks great, the size is perfect, the connectivity is good, price is low, but I don't feel like using it for measurements with this set of issues.

 

[AnalogSteph]

I wanted to use the digital input for digital audio measurements, as my external USB sound card with Toslink SPDIF input doesn't work in 80% of the cases for whatever reason.

What model was that? Could it be that it doesn't do more than 16/48 and that's your problem? Or that it doesn't work unless input sample rate is manually set up correctly?

Back in the olden days, if you mainly wanted SPDIF I/O on the cheap, the go-to card was some kind of CMedia-equipped job. You can still find internal cards with CM8828 + CM9882A today that would fit the bill, but something external that'll do 24/96 and not break the bank is a much tougher call. Maybe you can find a used X-Fi HD or something.

 

[kemmler3D]

G3 owner here using the toslink input so I can EQ my TV's output...

Did you install the app and update the firmware? Out of the box mine had some really screwy sample rate issues and underruns and all sorts of crap. Has been working like a champ since I updated the FW.

Otherwise, I guess 128dB isn't anything to complain about for listening, but I can see why it would not be suitable for measuring anything.

 

[di2ger]

What model was that?

I've tried these two:

1. Shiphose 7.1 Sound Card USB 2.0 Sound Card Cmi6206 Chipset USB USB Audio Device Class Spec1.0 and USB Hid Class Spec 8 Channel Dac Output, black (222132)
2. 6 Channel Sound Card,USB External Digital Optical SPDIF Audio Output Adapter,Plug and Play,for PC Laptop Recording Compatible with win98 / XP / 2000 / Vista / Win7/ MAC OS9.1

it doesn't do more than 16/48

True, so I was trying 16-bit at 44.1 and 48 kHz.

it doesn't work unless input sample rate is manually set up correctly

Checked all options, some sources work (like Topping d10s), and some don't (like Wiim Mini).

G3 owner here

Welcome!

Did you install the app and update the firmware?

Sure, my firmware version (the app says the latest) is 1.4.201027.1200. What's yours?

128dB isn't anything to complain about for listening

Very true, however, this is only in "What U Hear" mode. In the normal input way, I see THD+N: -71.9 dB [20..20000 Hz], which still might be not noticeable.
It'd be nice if you can measure your digital input with REW to see if your G3 works fine. If you don't have digital Toslink output, you can check line in, it has a similar problem.

 

[di2ger]

Maybe you can find a used X-Fi HD or something.

Thank you for the tips! While searching, I found this one.
It seems to be exactly what I need.

 

[di2ger]

So, I've gotten an answer to my questions from Creative support.
I asked:

I have a number of questions regarding the measurements using the suggested "What U Hear" input:
1. I notice that the digital Toslink input is not bit-perfect. It is close, but not exactly the same as the original. Can this be fixed?
2. With a 96 kHz sampling rate I expect to see some signal between 24 kHz and 48 kHz, but there is only noise. How can I get the full frequency range?
3. Can this low-frequency noise issue be resolved for normal inputs instead of "What U Hear" in future firmware updates?

Creative support answered:

As for the Optical In recording, kindly understand that our sound card is designed for entertainment usage and may not be suitable for professional usage.
We seek your kind understanding on this matter.

Translating from support English to plain English: yes, there are issues, but that's OK as most of the users don't notice them.
That's a pity as I really like the device.

 

[di2ger]

Support is not sure if they have anything bit-perfect:

We have the following model of the sound card that can record SPDIF In with the higher recording quality, but it might not be a bit perfect.

Sound BlasterX G6
Optical-in:
16 / 24 / 32-bit / 44.1, 48.0, 88.2, 96.0, 176.4, 192.0 kHz
https://ca.creative.com/p/sound-blaster/sound-blasterx-g6

Sound Blaster X4
SPDIF Input:
16 / 24-bit, 48.0, 96.0, 192 kHz
https://ca.creative.com/p/sound-blaster/sound-blaster-x4

Sound Blaster GC7
SPDIF Input:
16 / 24-bit, 48.0, 96.0, 192 kHz
https://ca.creative.com/p/sound-blaster/sound-blaster-gc7

 

[sool]

Thank you for the measurements.
I have another question regarding G3: if I would like to use it as a cheap DAC to connect my PC to amplifier - how well it works as a USB DAC? Does it measure any good? Have you any chance to test it as a inexpensive USB DAC?

 

[di2ger]

Sorry, @sool, I just double-checked my measurements and couldn’t find G3’s output measurements. It’s because at the time I was searching for a good input device.
I did a listening test when I still had the device - it was perfect to my ears. The thing is, I also find Mac Pro’s headphones out as perfect, apart from some CPU/GPU noise from time to time. The same with my Samsung smartphone’s out.
My guess would be that G3 will improve the sound mainly due to less interference, but what needs to be done is to measure both PC and G3’s outs and then ideally do an A/B test.