How good is the DAC on my motherboard?

[Cirrus Incus]

Hi, I'm currently using my onboard PC audio for my headphones (Fidelio X2) and desktop speakers (Edifier R1850DB). The motherboard I've got is an Asus Maximus Hero VIII.
https://www.asus.com/Motherboards/MAXIMUS-VIII-HERO/specifications/

I don't notice any background noise except an occasional very faint high pitched sound on the right side of my headphones when starting/stopping audio tracks. Music sounds good to my untrained ears.

The motherboard has an ESS ES9023P DAC:
http://www.esstech.com/index.php/en...acs/sabre-hifi-stereo-integrated-dacs/es9023/

The amp is a TI RC4580 2VRMS audio OP Amp:
https://www.ti.com/product/RC4580

How would these be regarded?
Would I notice a substantial improvement if I upgraded to an external DAC and amp?
I'm considering to get a Tempotec Sonato HD Pro for use with my PC and phone.

Thanks

 

[dasdoing]

use REW to measure your DAC: https://www.roomeqwizard.com/help/help_en-GB/html/soundcard.html
I am no DAC fetichist like many here. If the thing is flat and has no bad distorsion it wont make a diference (that is if it is well shieled like it should on a MB like this).
about the Amp, it obviously will be very limited, but your speakers are active anyways.

 

[dfuller]

In my experience, built in DAC/Amps on motherboards are... okay. Definitely nothing groundbreaking or SOTA, but for most people they're fine.

Will you notice a difference with an external DAC/Amp? Depends on how sensitive your headphones are. My motherboard's headphone output hisses like crazy with my (fairly sensitive) ATH-M50s and it's enough that it bugs me.

 

[m_g_s_g]

use REW to measure your DAC: https://www.roomeqwizard.com/help/help_en-GB/html/soundcard.html

I did read this and decided to make a loopback measurement of my motherboard's DAC (ASUS Z390-E Gaming SupremeFX). This is a summary of what I did (so I will remember that in the future. The information was scattered from several places):

1. Calibrate voltage level. Connect a multimeter to your motherboard line out RCA connector. Then: REW -> Generator -> sine tone (500Hz -inside my multimeter AC V specs-, -6dBFS) -> Play -> Measure voltage at line out RCA connector and finally -> Calibrate level. I double checked from the Generator that a 4dBu test tone level showed 1.228V at the multimeter. So far, so good.
2. Calibrate soundcard in REW preferences. Directly connect your soundcard/mobo line in to line out. Then: REW -> Preferences -> Select highest supported sample rate and -12 dBFS as sweep level -> Calibrate soundcard -> next/next -> Change your input volume level; make sure that input and output dBFS levels are similar (slightly lower in than out) ->next -> A measurement starts.
3. Measure. REW -> Distorsion: This is my result. Not very nice. My feeling is that the motherboard's ADC is way worse than the DAC (but I don't know how to prove it, since I don't have a measurement-level ADC interface).

My listening tests (;-)): This motherboard drives my 300 ohms HD6XX more than happily (it's supposed to have opamps to drive high impedance headphones). I bought a Shiit Heresy amp some time ago, but I have sold it since (I couldn't hear the difference). I'm not a gamer, so I didn't test under high GPU load. When I heavily load my nVidia GPU for work (I do), I can move somewhere else to hear music.

NOTE: I didn't have any luck doing a RTA measurement. I guess the configuration or sine tone (200Hz, -12dBFS) I was using was not appropriate, or maybe some other calibration is needed for that (I calibrated the -12dBFS tone as SPL 0 using the SPL Meter as suggested somewhere). The noise floor was absurd (-46dB, see second screen capture).

 

[dasdoing]

My feeling is that the motherboard's ADC is way worse than the DAC (but I don't know how to prove it, since I don't have a measurement-level ADC interface).

It's been a very long time since I used onboard sound, but back in the days the inputs were much worse indeed. This was before they started to shield the onboard audio chips though.
If you think about it. A gamer needs no quality inputs

 

[AnalogSteph]

NOTE: I didn't have any luck doing a RTA measurement. I guess the configuration or sine tone (200Hz, -12dBFS) I was using was not appropriate, or maybe some other calibration is needed for that (I calibrated the -12dBFS tone as SPL 0 using the SPL Meter as suggested somewhere). The noise floor was absurd (-46dB, see second screen capture).

That's not noise, that's non-ideal windowing. May indicate either nonideal REW settings (waveform captured with a time offset) or resampling going on.

First of all, I would make sure that all of the following are set up identically:
* Playback device sample rate and bit depth
* Recording device sample rate and bit depth
* Measurement software sample rate and bit depth
For starters, try 48 kHz, 24 bit everywhere (if the recording device does not offer 24 bit, use 16). I've had Realtek HDA ADCs woefully underperform at 44.1 kHz and multiples.

In recording device properties, also make sure input level is set properly. Be sure to right-click the level field and select dB for a better idea of what you're doing. Start out with an input level of 0 dB, with input gain (if any, for microphone / combo inputs) turned down to 0 dB as well.

REW is a bit of an odd choice for soundcard measurements. I'd try RMAA first. That definitely requires a stereo input though.

And yes, the recording side is usually worse than playback. You may be able to cheat a bit by using passive attenuation if you can either solder something up (using some metal film resistors and heatshrink) or have a suitable headphone amplifier handy, the idea being that distortion related only to the input would diminish with input level. Levels can then be brought up digitally (or perhaps using a built-in switchable preamplifier) at the expense of noise. So you still need to take measurements both with and without attenuation, but would have a better idea of performance levels in the end.
In a similar manner, a (known low-distortion, low noise) external preamplifier stage of known gain plus a corresponding output level reduction can be used to magnify output noise, allowing it to be quantified more accurately. So if for example you have an output stage with noise @ -103 dBFS and an input stage at -97 dBFS, you'll normally get -96 dBFS total (assuming same full-scale levels) with little idea of where output noise level actually is. Add 20 dB of low-noise external amplification, and you'll get -82.8 dBFS, so you can infer ca. -102.8 dBFS of output noise which would be super close to the actual value.

The crux of routing audio through cabling outside a system and back is the inevitable ground loops you end up encountering. It's usually manageable on the same sound device but from one to another in the same machine it's generally a lost cause with unbalanced cabling.

 

[m_g_s_g]

@AnalogSteph, thanks for your detailed suggestions.

I expect my motherboard audio to measure similarly to the Gigabyte Z390 that Amir reviewed recently.

I'm well aware of the fact that I'm not using (or have access to) the right recording gear to properly measure my motherboard audio. I was just trying to find out the limits of what I have available today. And I'll definitely look at RMAA, since I see that many people use it for computer motherboard audio measurement.

Anyway, I still feel that I'm making a newbie mistake using the RTA feature. While the FR and Distorsion graphs I got look Ok to me, the FFT looks weird. You are right, it doesn't look like a noise floor. Today I was only able to repeat the measurement (disabling Equalizer APO, it was engaged the other day!!), and it looks sligthly better. I'll recheck sample rates / bit depths / levels and REW options, but maybe this is my current limit before trying your other suggestions (passive attenuation, using a preamp with known gain).

 

[joe hues]

I have asrock a320m-hdv r4.0 as my MOBO, i just bought an amp and a nice pair of headphones? should i get a dac too? id really appreciate any and all help!

 

[joe hues]

mmx 300 and topping l30!

 

[zermak]

@AnalogSteph, thanks for your detailed suggestions.

Use Hann or Blackman-Harris window while testing single tones (it should be good for dual tones too) and switch to Rectangular window while testing multitones (but use higher samples).
16K samples is fine for a fast check but like you did you can go for a full sweep with higher values for more precise results.
By the way my cheap on-board loopback is around 77dB SINAD and it's definitle the ADC the weak link; the DAC is probably around 80dB SINAD and it is enough for normal listenings levels

 

[Mike B]

I have an old (2014) desktop. I had an old (probably 2000) set of powered speakers - two small satellites and a subwoofer. I bought a $40 sound card and used it in place of my onboard sound card, using the 3.5mm analog connection both times. The new sound card improved the sound noticeably. Then I bought a set of powered Klipsch speakers (>$300) and connected them to USB - they sounded a lot better. They made a much bigger difference than the swapped soundcard. As an experiment I also connected the Klipsch speakers to the optical digital out jack of the $40 soundcard. They also sounded better than the old speakers. I couldn't tell a lot of difference between Klipsch via USB and Klipsch via optical digital from the sound card. My original (2014) onboard sound card had no digital out so I had no way to try that. So my answer, with my cheap and old PC, is that a $40 sound card made a difference but $300 speakers made a much bigger difference. The fact that the new speakers could connect direct to the USB port made the new sound card unnecessary. But I was experimenting so it was not really wasted money.