• 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 Audio-gd DAC19 DAC

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
19,648
Likes
20,902
Location
Seattle Area
#1
This is a review and detailed measurements of the Audio-gd DAC19 DAC. It is on kind loan from a member who sent me two of their DACs for review. The DAC19 costs US $640 plus shipping from the company's website.

The design of DAC19 is similar to the rest of the Audio-gd line:

Audio-gd DAC19 DAC Audio Review.jpg

Controls feel solid and good. The unit also has good bit of heft to it so overall impressions are positive from a male perspective.

The back panel is what you expect plus a couple of their current-driven proprietary "ACSS" connectors:
Audio-gd DAC19 DAC Back Panel Audio Review.jpg
At this price range, XLR balanced output should be mandatory to help avoid ground loops which can readily occur with computer audio. It is missing here which is disappointing.

There have been many revisions of this model over the years. I think the original unit came out in 2010 and seems like that is the marking that is on the back of the unit. I am not an expert in all of that so don't know where this unit ranks so I took a picture of the inside:

Audio-gd DAC19 DAC Teardown.jpg


It doesn't look like their current units that have an FPGA based daughter card. The analog design and choice of DAC chip (PCM1704 ?) seems the same though.

There are some jumpers on this unit but I could not correlate them to anything on the Audio-gd website so went with them as is.

Prior Audio-gd products have not done well in my testing. They tend to get poor measurements and miss the published specifications. Will the DAC19 be an exception? Let's find out.

DAC Audio Measurements
Firing up the DAC19 with full amplitude digital signal of 0 dBFS as I start testing showed an output that was above nominal 2.0 volt we like to see (not a bad thing). So I dialed it down digitally to get us to 2 volts which improved the results a bit:

Audio-gd DAC19 DAC Audio Measurements.png


That second harmonic at almost -80 dB sets the SINAD for the bad channel at 80 dB. The other channel is 8 dB better showing lack of precision in design. Averaging the two channels gets us a SINAD of 84 dB which is best I have measured for Audio-gd (I think) still puts the DAC19 in forth bucket of all DACs tested:
Audio-gd DAC19 DAC SINAD Audio Review.png


There are no specifications for THD+N on Audgio-gd website for DAC19 but there is one for SNR so let's measure that:

Audio-gd DAC19 DAC Dynamic Range Audio Measurements.png


We are missing the specifications by almost 17 dB! Still, we clear the bar for CD/16-bit music so it is not the end of the world.

Jitter performance shows poor attention to clean engineering:

Audio-gd DAC19 DAC Jitter Audio Measurements.png


I lowered the output to -60 dB and most of the spurious tones disappeared (in red). Since real music doesn't have such high amplitude at 12 kHz, actual performance there may be OK.

Linearity measurement is a disaster:
Audio-gd DAC19 DAC Linearity Audio Measurements.png


The oscillating up and down usually indicates predictable error in signal processing/dithering. I changed the levels manually from -110 up and nothing would change in the output until I got into the 90s dB. So something is broken here. The error now goes well into CD territory as we can't get correct output level to -96 dB.

Intermodulation distortion test shows high noise and distortion levels as predicted by the dashboard:
Audio-gd DAC19 DAC IMD Audio Measurements.png


I could not run my multitone test because it is 192 kHz sampling rate and even over USB, I could not get the DAC19 to recognize that. It is beyond the specs for the device (96 kHz).

The reconstruction filter shows a very slow roll off:
Audio-gd DAC19 DAC Filter Response Audio Measurements.png


Even by 24 kHz, we only get 40 dB of reduction of out of band components. Ideally this number would be far, far bigger. At 22.05 kHz, we are only down a negligible 6 dB (theory mandates infinite truncation at this point).

The slow roll off also bleeds into audible band:
Audio-gd DAC19 DAC Frequency Response Audio Measurements.png


We have lost 1.5 dB by 20 khz which is not an issue for us older folks but for younger listeners, there will be somewhat less highs. Maybe this is the appeal of this DAC for some?

Conclusions
The data clearly speaks for itself: the performance of the Audio-gd DAC19 under test is not competitive at all. At the price charged, it both lacks features such as balanced output and pure performance. The newer revisions may have made the jitter and roll off issues less of a problem but I suspect the rest remains due to avoidance of feedback and good design practices.

It is a shame really. Looking inside, the DAC19 warms the cockles of this engineer's heart with old school use of tons and tons of discrete parts. A lot of work has gone into the design of this unit, yet with no aim to verify good performance.

From audibility point of view, the distortion levels are low enough that the average audiophile won't hear. No, there is no magic that they will hear either to make this design superior.

Needless to say, I can't recommend the Audio-gd DAC19. There are so many better choices today at far lower prices.

------------
As always, questions, comments, recommendations, etc. are welcome.

I am going to get a haircut today. One of my pink panthers heard me saying that and says he too wants a haircut. Problem is, I can't afford a panther haircut. They have to be tranquilized and such so it costs a lot more than human haircuts. I feel bad for him so please donate some money so I can give him the grooming he deserves using::

Patreon: https://www.patreon.com/audiosciencereview), or
upgrading your membership here though Paypal (https://audiosciencereview.com/foru...eview-and-measurements.2164/page-3#post-59054).
 
Joined
Jun 20, 2018
Messages
32
Likes
42
#4
Not surprising considering how their other offerings performed. Thanks Amir.

I'm curious how other, proper implementations of the PCM1704 would measure.
 

Newk Yuler

Member
Patreon Donor
Joined
Mar 5, 2019
Messages
57
Likes
32
#7
This is an early version of the DAC sans the DSP that could be added with a small solder mod. Not surprised it measures poorly and glad to know for certain. I have one of the earlier DAC 19 DSP models that gets some occasional use. The DSP board has a number of jumpers that allows operational options going all the way down to a complete bypass. I intend to try running HQPlayer with its excellent resampling filters into it manually bypassing the DSP features, although the resolution is limited by the DIR9001 S/PDIF receiver to 24 bit 96 kHz. The DAC 19 is no longer available as are none of the DACs using the supposedly awesome PCM1704UK R2R DAC chips. The Audio GD stuff has beautiful guts. Too bad it consistently measures like crap. :(

http://www.audio-gd.com/Pro/dac/DAC19/DAC19EN.htm
 
Joined
Apr 25, 2019
Messages
30
Likes
46
#8
Outperformed by a 9$ Apple dongle...:facepalm:
Can anyone imagine an Intel/AMD competitor that comes up with a CPU that costs 50 times more and offers a much worse performance? It just doesn't make any sense, yet in audio field those firms somehow manage to exist. Reminds me of a Louis CK joke about a certain pet food store...
 
Joined
Mar 27, 2019
Messages
8
Likes
33
#9
This is my DAC-19. Bought it used maybe 4 years ago for $350. After reading Amir's other measurements of Audio-Gd stuff, and reading his review of the Topping D50, I bought a D50 off eBay for $150...

FYI Before I sent this off to Amir, I did an A/B comparison of this DAC-19 vs. the Topping D50 and could not tell the difference with statistical reliability. System used: FLAC files made from red book CDs via SPDIF, 50K Alps pots for volume control and A/B level matching, Harman-Kardon Citation II amplifier and Quad ESLs.

The Topping D50 measures a lot better. I don't know how to correlate these DAC measurements with audibility. What are the thresholds? At any rate, I'll stick with DACs that measure better over ones with the kind of performance this unit showed. Especially for $150! And then, too, the Topping D50 is the size of a deck of playing cards vs. this Audio-Gd boat anchor.

I really want to like the Audio-Gd stuff, because of their fanatical overbuilt build quality. I'm an old man- I'm used to having my illusions shattered.
 
Joined
Mar 27, 2019
Messages
80
Likes
87
#10
A lot of people rate the audio-gd devices highly, waxing superlatives of how they “sound”

To me, amirm’s measurement strongly suggest that, for better or worse, they impart their own acoustic signature or “special effects” downstream...
 

invaderzim

Active Member
Joined
Aug 30, 2018
Messages
252
Likes
391
Location
NorCal
#11
A lot of people rate the audio-gd devices highly, waxing superlatives of how they “sound”

To me, amirm’s measurement strongly suggest that, for better or worse, they impart their own acoustic signature or “special effects” downstream...
Sometimes the picture just looks better with the model in soft focus.
 
Joined
Dec 27, 2018
Messages
16
Likes
12
#12
I remember when this company exploded onto the scene many years ago over at headfi. I also remember getting in heated debates about not hearing a lot of the magic that others claim to have heard with his dacs. It's nice to see my posts of many years ago being validated lol. This also reminds me of an a comment a reviewer once wrote "many audiophiles still hear with their eyes".
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
19,648
Likes
20,902
Location
Seattle Area
#13
This also reminds me of an a comment a reviewer once wrote "many audiophiles still hear with their eyes".
That, they do....
 
Joined
Nov 2, 2018
Messages
90
Likes
48
#14
I think boutique manufacturing requires breaking away from best practices. There is no other way to stand apart from the rest. Hopefully, the results are euphonic.
 

jackenhack

Active Member
Patreon Donor
Joined
Oct 25, 2018
Messages
190
Likes
411
Location
Stockholm, Sweden
#16
I don't understand how this company can still be in business. I bought two different op-amps from them, and they were horrible. Fully discrete components and the old NE5532 was many times better, but You can get it for pocket change. I’ve never seen a good measurement of any equipment from that company. Goes to show you that the audio field is the only field I know where reviewers get less educated the longer they work in the business. Steve Guttenberg comes to mind...
 

solderdude

Major Contributor
Joined
Jul 21, 2018
Messages
1,762
Likes
2,726
Location
The Neverlands
#17
It's a good thing that @amirm did not show multitone performance.
Unless one would like so see what a not mowed lawn looks like there probably will not be many flowers visible between the grass.
 

GrimSurfer

Active Member
Joined
May 25, 2019
Messages
292
Likes
230
#20
Outperformed by a 9$ Apple dongle...:facepalm:
Can anyone imagine an Intel/AMD competitor that comes up with a CPU that costs 50 times more and offers a much worse performance? It just doesn't make any sense, yet in audio field those firms somehow manage to exist. Reminds me of a Louis CK joke about a certain pet food store...
Reminds me of the audio cable industry!
 
Top Bottom