• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Exasound E32 DAC Review and Measurements

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
46,033
Likes
256,913
Location
Seattle Area
I have been reviewing a lot of low cost DACs recently. I wanted to have a high-performance DAC as a reference. My best DAC is in my audio system and pain in the neck to relocate back and forth. So I set out to find a DAC in $2000 range (my cost). Idea was to have a very low noise/low distortion DAC that would be revealing of any tweaks put in front of them.

After getting suggestions from the forum, I short-listed it to Benchmark and Exasound. Both of them provide measurements on their site which gave me confidence of good design. As it turns out, I reached out to Exasound and they immediately answered so I went with them. I asked for accommodation pricing on E32 DAC and I received a good discount. As a professional courtesy I don't want to say how much that is. Suffice it to say it is similar to dealer margin for high-end products. Should this have been a transaction with someone I know, the discount is usually 10 to 20% higher. Still, I thought it was a reasonable offer and I received the unit from Canada a couple of weeks ago.

upload_2017-10-11_20-45-39.png

The retail price of E32 is $3,500 so not cheap at all.

Overview
Having been used to mostly DIY and cheesey packing of $300 and under DACs, the Exasound E32 is a breath of fresh air. The unit while small, is quite sturdy and very attractive. The case work is sold and has a lot of custom touches such as curved front panel. It also has a nice display that shows what it is receiving which is very useful for testing to confirm what I am sending, is what is being received by the DAC.

Mind you, it is not to the same level of multi-thousand dollar DACs which come in larger enclosures, weigh a ton, etc. But for this tier devices, it seems quite suitable and no complaints.

As expected the E32 comes with both balanced and unbalanced outputs. Inputs are USB, coax and optical SPDIF. Would have preferred one of those inputs to have been AES/EBU. It doesn't look like they had enough room to put one in there.

Two things were disappointing here:
1. It comes with a switchmode power supply as an external wall-wart. As with my other tests, this also leaks AC mains into the DAC and comes right out of its analog outputs. In my testing below, I opted to test with my lab supply to eliminate that. I can't find the darn switching power supply right now anyway :). When I do, I will post its result too.

2. It comes with proprietary drivers! I wish I had noted this before or I would not have purchased it. The issue is that it won't work with any Linux product which includes all streamers in the market. Furthermore, if the company goes out of business one day and Microsoft changes the driver interface, this will become a door stop even on Windows. This has happened plenty of time in the past so it is a major negative against the unit this pricey which I expect to use well into the future.

Measurements
I started with my usual J-test measurement at 24-bits/48 Khz. For reference, I used my $79 Behringer UMC204HD -- the winner of low cost DACs. I started with unbalanced/RCA output first:

upload_2017-10-11_20-30-11.png


I did not bother to label the graph because as you see, it is uneventful. There is a tiny spike where the marker is which might be USB packet frame timing. The differential is 130 db though so not an issue.

Compared to Behringer, the output is higher by a few db which is good. The differential between that peak and the signal is about 1 db in favor of Exasound E32.

I then ran a harmonic distortion test of 1 Khz tone (24-bit/44.1 Khz):

1Khz unbalanced.png


The picture here is quite disappointing. While the noise floor is slightly lower in Exasound, it has tons more harmonic distortion than Behringer. The Behringer nails this department with just a single third harmonic which is 10 db lower than Exasound. There is some genius analog designer working at Behringer!

In other words, I took many steps back when it comes with the Exasound. Very, very disappointing.

Fortunately things looked up when I tested the balanced output. First the J-test again:

J-test.png


Here the much higher output of Exasound comes through to deliver 7 db better signal to (measurement) noise floor.

Repeating the same test with 1 Khz we get:

1Khz balanced.png


Here the reverse picture appears vs unbalanced. The Exasound is 6 db better when we are comparing the peak signal to 2nd harmonic distortion. The Exasound still has higher order harmonics that the Behringer does not have. So it is not a complete home run.

As always, note that these distortions and noise products are at very low levels. I doubt that any differences are audible at all. But since I wanted to use the high-end DAC for measurement reference, they matter directly to the task at hand.

Summary
It pains me to say this but I really can't recommend the Exasound E32. I think this is the first time any reviewer had gotten an audio device at accommodation pricing and turned around and poke the manufacturer in the eye :). But the reality is the reality. In unbalanced mode, the performance simply is not acceptable when a $79 DAC beats it left and right. In balanced, there is an improvement but the requirement for proprietary driver is a showstopper and I can't good in good conscious give a thumbs up for it.

I am not sure what I am going to do at this point. I can't afford to keep buying multi-thousand dollar DACs to evaluate and find the ideal one :). At RMAF 2017 I spoke to a couple of manufacturers about evaluating their units. If so, then I my sell this unit and upgrade one of those.

For now, I am in somber mood. Everyone please shed a tear for me. :)

As always, I am open to suggestions, corrections, additional data/comments from everyone.
 
Last edited:
Here is a quick comparison using J-test comparing the Exasound E32 against Mytek Brooklyn:

Exasound vs Mytek.png


The Exasound clearly beats the Mytek here by 10 db. Their analog output levels are essentially identical and the Exasound has 10 db lower noise floor.

The Brooklyn retails for $2,000 though so quite a bit cheaper.
 
As always pertinent and interesting, but if you carry on being honest soon manufacturers will refuse to supply you, oh wait that’s already occurred hasn’t it!
Keith
 
I can see some subjectivists arguing that you measurements versus sound. To that end, would it be possible to do A/B listening tests of this dac and others via a switch or something similar?
 
If I was to have the spare cash to buy an Exasound, I would likely feed it the highest bit rate I could upsample to - I think DSD256 is its limit.
Would testing using a DSD256 input make any difference?
 
After getting suggestions from the forum, I short-listed it to Benchmark and Exasound.

Odd harmonics...

Both channels?

Clipping?

Inter-sample "overs"?

What does it do when fed less than full scale?

Do you have a Benchmark coming? DAC2 or DAC3?
 
Do you have a Benchmark coming? DAC2 or DAC3?
I have asked for a DAC3. Reception to my request was "OK" at the show. I have not heard from them but will ping them to see what happens.
 
As always pertinent and interesting, but if you carry on being honest soon manufacturers will refuse to supply you, oh wait that’s already occurred hasn’t it!
Keith
While some companies are willing to have their products tested and even offer them without me asking, those seem to be quite an exception. So you are right: expecting manufacturers to sending products for review is not a strategy that will work.

I think the better plan and one that I will pursue is to test products that users have already purchased.
 
Odd harmonics...

Both channels?.

Clipping?

Inter-sample "overs"?

What does it do when fed less than full scale?
Let me give a general answer to this: I am trying best to not post a ton of charts and graphs in reviewing products. Others do that and that is fine. But I think I lose the audience when that is done. To wit, I don't see a reason to post frequency response and such when it is so flat and similar in this class of product. I am open to feedback on this front.

As to your question, the 1 Khz tone was at -6 dbFS. So clipping is not an issue here. Clipping effects tend to be quite a bit more horrendous. Here is the Mytek again:

index.php
 
I can see some subjectivists arguing that you measurements versus sound. To that end, would it be possible to do A/B listening tests of this dac and others via a switch or something similar?
I hear the criticism. In this case though, I specifically bought this DAC as a measurement tool. That said, I will do an AB test when I get a chance but I would have to adjust its output level relative to Behringer which may impact the results.
 
I have been using a balanced output version of an Exasound E28 happily for the last 3-4 years. That unit has been superceded by the E38, for which there is not ( yet, maybe never?) a balanced output version. But, balanced is important in my 7.1 setup. I just don't like the idea of long unbalanced interconnects or of recabling under my listening room floor for unbalanced. And, I go directly from the DAC to my amps. So, I am sticking with the E28 for now.

I knew about the software driver issue, which has been absolutely no problem for me coming out of a Win 7 HT PC. I have found the instantaneous channel by channel level and sampling rate displays available from that driver on the PC screen to be quite useful diagnostically from time to time. It helped me solve some problems and reach some understandings which I might not have otherwise.

I have ignored most of the computer audiophile fluff about LPS vs. SMPS, though it appears from the E32 measurements that might be an issue, though apparently not via balanced outs. So, I think I will stick with George Klissarov's wall wart.

Kal did a quick survey of Mch DACs in Stereophile a few months ago. He felt that the Exasound E28/38 occupied something of a sweet spot among the slim pickings in Mch DACs available. I agree. I do not see anything else that temps me near its price point.

I do not use the ultra hi sampling choices much, listening primarily to SACD rips converted to 176k PCM for compatibility with Dirac Live on my PC. But, I have been able to listen to some DSD256 and PCM352k Mch recordings I have at their native sampling rates. Interesting, but not by enough to change my basic preference for 176k with Dirac.

Sorry for your disappointment, Amir. In my case, though, the E28 via USB was a definite sonic upgrade to the Integra prepro I had been using via HDMI, which is no surprise. I am quite happy with my sound, and it still stacks up well subjectively to the finest systems I have heard.
 
Excuse my ignorance but can we assume that the Behringer would perform as well on the harmonic distortion test of 1 Khz tone (24-bit/44.1 Khz) if it had unbalanced outputs?
 
As to your question, the 1 Khz tone was at -6 dbFS. So clipping is not an issue here. Clipping effects tend to be quite a bit more horrendous.

Some discussion on clipping.

https://benchmarkmedia.com/blogs/application_notes/intersample-overs-in-cd-recordings

"Every D/A chip and SRC chip that we have tested here at Benchmark has an intersample clipping problem! To the best of our knowledge, no chip manufacturer has adequately addressed this problem. For this reason, virtually every audio device on the market has an intersample overload problem. This problem is most noticeable when playing 44.1 kHz sample rates."
 
Last edited:
AFAIK from the plots Amir did not oversample in his tests, though I do not know if this is a problem in the DAC itself (e.g. if it oversamples always, or due to the delta-sigma modulation inside). The interpolation problem is well-known and old stuff, you'd think DAC FW designers would have that handled by now.
 
Given your goal of getting a DAC for clean measurements, you likely would not have done better than the Benchmark. For one you have JA's measurements showing superb results. I believe he commented it was possible the AP he measured with was not equal to the Benchmark and the actual capabilities may have been better than measured in some parameters.

I am curious though, as the Exasound doesn't meet specs or even come close. Is it possible yours is not operating to normal specifications? I think I would ask them about that possibility.

https://www.exasound.com/StartHere/CompareSpecifications.aspx

https://www.exasound.com/e32/e32DACMeasurements.aspx
 
I think I noticed, on my DAC2, that the S/PDIF and USB produce a different Impulse Response.

I'll have to go back and try again, but what I remember seeing was a symmetrical IR - wiggle, hump, wiggle - on the S/PDIF. Using the USB input - flat, hump, more pronounced wiggles after.

(here lacking the education to properly name, nor remember from past reading, those response types (due to filters) without looking them up someplace).

I'll have to repeat the observation.
 
I think I noticed, on my DAC2, that the S/PDIF and USB produce a different Impulse Response.

I'll have to go back and try again, but what I remember seeing was a symmetrical IR - wiggle, hump, wiggle - on the S/PDIF. Using the USB input - flat, hump, more pronounced wiggles after.

(here lacking the education to properly name, nor remember from past reading, those response types (due to filters) without looking them up someplace).

I'll have to repeat the observation.
Hump with wiggle after might be minimum phase filter. Symmetrical might be what is colloquially called a linear phase IR filter. Or do I have that wrong? Don can straighten us out.
 
Excuse my ignorance but can we assume that the Behringer would perform as well on the harmonic distortion test of 1 Khz tone (24-bit/44.1 Khz) if it had unbalanced outputs?
Behringer has both just the same and I showed those results:

index.php


I called it "RCA" because more people are familiar with that term than unbalanced.
 
Back
Top Bottom