Exasound E32 DAC Review and Measurements

[Thomas savage]

I must say, I am very surprised in the tone of voice in their response. It gets personal which is unprofessional. Anyway, answers below.

They should have contacted me if they thought this was unusual. What stopped them???

• He was within the 30-days trial when he published the review, yet he said “Everyone please shed a tear for me”.​

30-day trial? I purchased the unit direct from them at a discounted price (NOT through any ecommerce transaction). It would never occur to me that it would come with any kind of trial. Had I known, I would have likely returned it.

• He never asked us to return or to test his DAC. The source of distortion was never investigated.​

Per above, I didn't think I had return privileges, nor that the device was broken.

• He asked for industry accommodation and he states publicly that he is not happy with the discount after he agreed with the purchase.​

I didn't say I was unhappy with it. This is what I said:

"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."​

I get accommodation pricing from other companies at much deeper discounts. So I stated that as a fact. And that readers should be aware that I am not in Exasound pockets because I received deep/ridiculous discount.

• He talks against the use of a switching power supply, but he lost it and he couldn’t repeat the tests.​

Why are they complaining about this? My lab supply produced better results. They should bank that and not complain.

• It is not clear if he is short-cutting the galvanic isolation during measurements, this can be the explanation of the extra power supply noise that he is talking about.​

Huh? How does an end-user shortcut galvanic isolation inside the unit?

• He didn’t calibrate his measurements at 0dBfs , which is lazy, unprofessional, disrespectful and misleading.​

This is impossible to do. In all of my measurements, I show two different devices with two different outputs. As such, it is impossible to calibrate 0 db to both of them at the same time. I have pleaded with Audio Precision to help solve this problem somehow but for now, the presentation does not allow such calibration. If I were just testing their device, sure, I could set it to 0 dBR.

• He is using equipment with noise floor that is higher than the noise floor of the DAC.​

This shows extreme misunderstanding of how Audio Analysis using FFT works. They should read my tutorial on audio measurements here: https://www.audiosciencereview.com/forum/index.php?threads/understanding-audio-measurements.2351/. Briefly, due to power of wider FFTs, we are able gain some 30 dB of noise reduction. That allows us to now see distortion spikes that would normally be hidden to noise.

And why would they complain about this anyway? Using an even better analyzer would uncover more issues in their design, not less!

He is also confused thinking older Audio Precisions don't work as well as the new ones. See this current comparison on Audio Precision website:

View attachment 11824

Mine is a close cousin of the 2700 on the right. Note that the 2700 beats all but the highest end new generation AP to the left. And it loses by just a few dbs to the APX-555. I have been in communication with Audio Precision to potentially purchase the newer generation units but I have to tell you, the performance gap is exceedingly small, making the huge ($28,000) expense hard to justify.

For their measurements by the way, they used very wide FFTs resulting in artificially lower noise floors.

• I am not sure whether he is using the USB or the SPDIF. They are not equal.​

My bad for not being specific here. This is one of my earlier reviews. Answering anyway, it is USB. Later in other reviews I have shown S/PDIF.

What is not equal about them anyway in a high-end DAC?

• His choice of test signal format is less than optimal (44.1)and with unknown distortions.​

Boy are they confused. The test signals are coming from a file in digital domain. They have no in-built distortions because they are computer generated (sans dither as appropriate).

• He never discussed the test conditions and the test results with us before publishing them.
As it should have been. It is not like they loaned me a free unit for review. I bought the unit for my own use and measured/published the results. And clearly mentioned at the end: "As always, I am open to suggestions, corrections, additional data/comments from everyone." They sat around to complain now and in a different set of communications???

I am not beholding to them on such purchases anyway. If they saw something they did not like, they could contact me just as other manufacturers have. Why would I want to go and chase them when it is not important for them to follow up with me???

• He never contacted us after the purchase and we don’t feel like registering to his forum to play his game. Replying to another forum may just bring more bad publicity, so for now we are ignoring Amir.

Playing games? This is so unprofessional. This is their business. They need act like proper business people. I don't care how they feel. They need to bite their lip and follow up with me and help get better outcome for their devices. And what game am I playing that involves me spending real thousands of dollars to buy this unit???

So that it is clear: my #1 issue with this DAC is the requirement for proprietary drivers. I cannot recommend any DAC this way. Period. Such DACs won't work on Linux devices for example since they have no drivers so forget about using them with any streamer. They know this and it is absent in this reply here. If I were them, I would revise the design and make it UAC2 compliant. I don't care what the performance of the device is. I will always recommend against it if it requires proprietary drivers.

From engineering point of view, I have frequently shown the Exasound to beat all other DACs such as their competitor, Schiit Yggdrasil: https://audiosciencereview.com/foru...ents-and-review-of-schiit-yggdrasil-dac.2358/

And in Bryston review: https://audiosciencereview.com/foru...w-and-measurements-of-bryston-bda-2-dac.2388/

Clearly I give them credit where it is due.

So no, I play no games. I measure things and if there is an issue with them, I welcome manufacturers to contact me. When give a discount, I also don't think about returning gear to manufacturers. I am assuming I am stuck with them and don't have return privileges like retail customers do.

From reading the forum and your reviews of other DAC’s I have always been under the impression the Exasound was often regarded as a benchmark for good or excellent measured performance.

I’m left mystified by the tone of their response tbh it dose not seem to service their best interests .

 

[Blumlein 88]

It seems they read only the Exasound review, but they should check all the reviews, because Exasound is very good imo..
Why everybody is sure that Benchmark and RME will measure better?

As with its predecessor, the DAC2 HGC, which Erick Lichte reviewed for Stereophile in February 2014, Benchmark's DAC3 HGC offers state-of-the-art measured performance. All I can say is "Wow!"—John Atkinson
Read more at https://www.stereophile.com/content...ne-amplifier-measurements#06bH9Y4usvVrsPyB.99

Measurements of Benchmark gear typically matches what they claim. And typically is among the best possible results.

RME is a no BS pro audio company. Their claimed measures would be excellent though maybe not equal to the Benchmark. So it is reasonable to think the RME will be excellent though I don't know of measurements of it yet.

 

[amirm]

Since this discussion is continuing on another forum, I thought I run a fresh set of new measurements. Since a lot has happened since this original review, I thought I make a comparison to the Topping DX7s which retails for $499 versus $3499 for the Exasound E32. The measurements are all through S/PDIF output of my Audio Precision analyzer so should be repeatable by others and eliminate arguments over what file was used for USB testing. All tests are with unbalanced RCA outputs at 0 dB volume.

Also, I have since found the original Exasound power supply and that is what is used now in these tests. The original tests in this review used my lab power supply.

Since there is a lot of back and forth on harmonic distortion, I thought I run my ultra-low-noise residual tests of 1 kHz tone which itself is filtered out. Everything remaining is noise and distortion created by the DAC (the analyzer's own contributions are negligible):

Compared to Topping DX7s, the magnitude of the harmonic distortions are lower. But look at all of those spikes in between them! And at lower frequencies before out main 1 kHz tone. This is just unacceptable for any DAC above $100 let alone at $3,500. Something is generating a lot of noise there. This is why I switched to my lab supply which quieted down some of them.

Next let's look at DAC linearity:

This was a shocker although is something I had noticed before. Previously I had posted much better outputs from Exasound E32 (although not as good as Topping DX7s). At that time I noticed that what I measured was variable. And it was the case here too but I could not get it to substantially look better. We lose linearity at around 96 to 98 dB and things progressively get worse.

I immediately plugged all the cables into Topping DX7s and got the super nice performance of nearly flat line out of Topping DX7s. I moved the cables back to Exasound E32 and the same problem remained.

Let's look at -90 dB sine wave and see how we do as far as linearity there:

The sine waves are similarly clean but as you see in the expanded time scale of the Exasound on the right, they are dancing up and down! A low frequency signal is modulating them as we have seen with some less than performant DACs.

This variation is clearly visible on a -90 db signal which is within the envelope of a 16-bit audio format (96 dB signal to noise ratio). So this is a confirmation of why we saw linearity get lost in the previous graph.

At this point I am just going to stop. The message is very clear. The performance of Exasound E32 is not competitive here. I looked at the measurement page for Exasound E32 and sadly none of these tests are there. So maybe they don't even run such tests. I don't know but excellence in engineering is lacking here.

Performance here is likely impacted by power supply. Why on earth they would complain that I did not use their power supply is beyond me. They should know it makes things worse because they themselves used battery as power source.

Anyway, this is a huge bummer for me because I was intending to sell the Exasound to free funds for other use. Now I am stuck with my own data damning any such sales.

 

[Wombat]

Since this discussion is continuing on another forum, I thought I run a fresh set of new measurements. Since a lot has happened since this original review, I thought I make a comparison to the Topping DX7s which retails for $499 versus $3499 for the Exasound E32. The measurements are all through S/PDIF output of my Audio Precision analyzer so should be repeatable by others and eliminate arguments over what file was used for USB testing. All tests are with unbalanced RCA outputs at 0 dB volume.

Also, I have since found the original Exasound power supply and that is what is used now in these tests. The original tests in this review used my lab power supply.

Since there is a lot of back and forth on harmonic distortion, I thought I run my ultra-low-noise residual tests of 1 kHz tone which itself is filtered out. Everything remaining is noise and distortion created by the DAC (the analyzer's own contributions are negligible):

View attachment 11851

Compared to Topping DX7s, the magnitude of the harmonic distortions are lower. But look at all of those spikes in between them! And at lower frequencies before out main 1 kHz tone. This is just unacceptable for any DAC above $100 let alone at $3,500. Something is generating a lot of noise there. This is why I switched to my lab supply which quieted down some of them.

Next let's look at DAC linearity:

View attachment 11852

This was a shocker although is something I had noticed before. Previously I had posted much better outputs from Exasound E32 (although not as good as Topping DX7s). At that time I noticed that what I measured was variable. And it was the case here too but I could not get it to substantially look better. We lose linearity at around 96 to 98 dB and things progressively get worse.

I immediately plugged all the cables into Topping DX7s and got the super nice performance of nearly flat line out of Topping DX7s. I moved the cables back to Exasound E32 and the same problem remained.

Let's look at -90 dB sine wave and see how we do as far as linearity there:

View attachment 11853

The sine waves are similarly clean but as you see in the expanded time scale of the Exasound on the right, they are dancing up and down! A low frequency signal is modulating them as we have seen with some less than performant DACs.

This variation is clearly visible on a -90 db signal which is within the envelope of a 16-bit audio format (96 dB signal to noise ratio). So this is a confirmation of why we saw linearity get lost in the previous graph.

At this point I am just going to stop. The message is very clear. The performance of Exasound E32 is not competitive here. I looked at the measurement page for Exasound E32 and sadly none of these tests are there. So maybe they don't even run such tests. I don't know but excellence in engineering is lacking here.

Performance here is likely impacted by power supply. Why on earth they would complain that I did not use their power supply is beyond me. They should know it makes things worse because they themselves used battery as power source.

Anyway, this is a huge bummer for me because I was intending to sell the Exasound to free funds for other use. Now I am stuck with my own data damning any such sales.

Offer it for sale on the forums that denigrate your measurements.

 

[Bjorn]

Unless Exasound has a good explanation and can prove otherwise, this is an embarrassment for the company. But how on earth did they achieve class A status in Stereophile?

 

[Blumlein 88]

Unless Exasound has a good explanation and can prove otherwise, this is an embarrassment for the company. But how on earth did they achieve class A status in Stereophile?

Did you notice they did this without a formal review. No measurements. Just some of their writers having some experience of them etc. They made class A without a review. Wow.

This is the closest to a review.
https://www.audiostream.com/content/exasound-audio-design-e32-dac

And it merely parrots the claims of Exasound.

 

[Bjorn]

Did you notice they did this without a formal review. No measurements. Just some of their writers having some experience of them etc. They made class A without a review. Wow.

This is the closest to a review.
https://www.audiostream.com/content/exasound-audio-design-e32-dac

And it merely parrots the claims of Exasound.

I wasn't aware of that. Thanks for pointing it out.

I've tried an earlier Exasound (e18) myself and did a direct comparison to a Lynx Two-B sound card (true high end stuff!). The e18 was clearly lacking in transparency to my ears.

 

[Ron Kuper]

Coming from the other forum

Amir I believe that you have the community best interest that's why I hope you'll see that this is NOT helping.

Two parties with different measurement results, who are unwilling to collaborate or even communicate directly to get to the bottom of this.

Also, this hurts a segment of the community in a way which might not be obvious at first, so I wanted to bring it to your attention.

The e32's brother e38 is about the only consumer 8ch DAC between the $300 (miniDSP) and $11k (Merging) price ranges.

It's not like there's a "Topping" for 8ch DACs.

Lacking alternatives, not getting to the bottom of this or allow more scrutiny leaves a big hole of uncertainty and confusion for those in the market for a quality 8ch DAC. At least that's were I'm at.

Call me naive but I am not ready to accept that easily the insinuations made by both sides towards each other's integrity or professionalism. I am willing to bet something else, technical, is at play here.

Perhaps you can take the high road and contact George?

I know you do this voluntarily as a community service and I admire your contribution, which is exactly why I allow myself to express my feelings this way, since I believe they represents a specific part of the community and I believe that you would've care to listen.

 

[Burning Sounds]

I wasn't aware of that. Thanks for pointing it out.

I've tried an earlier Exasound (e18) myself and did a direct comparison to a Lynx Two-B sound card (true high end stuff!). The e18 was clearly lacking in transparency to my ears.

I tried an Exasound E28 when I was looking for a relacement for my 4 Benchmark DAC1s running in parallel. I also tried a Mytek 8x192ADDA at the same time, which had a good reputation in the pro world. I found the Mytek and the Benchmarks to subjectively sound very similar. The Exasound was very smooth and had a relaxing top end which I liked. However, I found its bass to be soft compared to the Benchmarks and mentioned this to the UK dealer when I returned the unit. He said he would mention it to Exasound and true to his word he got back to me a few days later. Exasound had told him the Benchmarks had a bass "hump" which would account for the difference I was hearing. I found this a strange response from Exasound and it put me off their products. I ended up purchasing the Mytek.

Coming from the other forum

Amir I believe that you have the community best interest that's why I hope you'll see that this is NOT helping.

Two parties with different measurement results, who are unwilling to collaborate or even communicate directly to get to the bottom of this.

Also, this hurts a segment of the community in a way which might not be obvious at first, so I wanted to bring it to your attention.

The e32's brother e38 is about the only consumer 8ch DAC between the $300 (miniDSP) and $11k (Merging) price ranges.

It's not like there's a "Topping" for 8ch DACs.

Lacking alternatives, not getting to the bottom of this or allow more scrutiny leaves a big hole of uncertainty and confusion for those in the market for a quality 8ch DAC. At least that's were I'm at.

Call me naive but I am not ready to accept that easily the insinuations made by both sides towards each other's integrity or professionalism. I am willing to bet something else, technical, is at play here.

Perhaps you can take the high road and contact George?

I know you do this voluntarily as a community service and I admire your contribution, which is exactly why I allow myself to express my feelings this way, since I believe they represents a specific part of the community and I believe that you would've care to listen.

In addition to the Mytek, which is still in production there are numerous 8 channel dacs from inexpensive units like the M-Audio M-trak 8 (about £300 - I have one of these also) to the Motu 8A and RME 8 channel dacs. I realise these are not "consumer" units and may not always be the most aesthetically pleasing, though.

 

[DonH56]

The E32's envelope looks like it is around 60 Hz; mains leakage? One of the problems I used to fight was when the test system ground and equipment ground were separate and low-level leakage/ground loops were introduced. I used an isolation transformer and star ground system but sometimes the leakage paths were tough to ferret out. I saw similar things when the transformer field and/or internal ground scheme was not "good". (Put "good" in quotes because 60 or 120 Hz components 90-120 dB down were not likely to be audible, just seen on the bench.) I am NOT saying that is the problem here, just noting the frequency seems about right. Might help troubleshooting, though at this point I suspect both parties are just tied of the debate.

 

[garbulky]

So now we have previous good measurements and now poor measurements. This doesn't strike me as a good method. If power supplies can make such large differences in the measurements, then it makes sense that both versions be tested the first time. Otherwise you have a first time review that is incorrect in regular usage. What other DAC measurements could be affected?

 

[amirm]

So now we have previous good measurements and now poor measurements. This doesn't strike me as a good method. If power supplies can make such large differences in the measurements, then it makes sense that both versions be tested the first time. Otherwise you have a first time review that is incorrect in regular usage. What other DAC measurements could be affected?

Well, the manufacturer itself used a battery bank and not any power supply so it is not like there is no precedent for that.

I usually perform both measurements but in this case, I literally could not find the power supply when I needed to get that measurement out. So I noted it in my review and posted the results. Prior to that I had tested it with its own supply and my lab one and knew the lab one performed better.

As for recent measurements, they are not the same ones I ran before. They are additional tests I have developed post that review.

 

[garbulky]

Well, the manufacturer itself used a battery bank and not any power supply so it is not like there is no precedent for that.

I usually perform both measurements but in this case, I literally could not find the power supply when I needed to get that measurement out. So I noted it in my review and posted the results. Prior to that I had tested it with its own supply and my lab one and knew the lab one performed better.

As for recent measurements, they are not the same ones I ran before. They are additional tests I have developed post that review.

Yeah I don't get the battery bank either. If they are not confident of the performance of a $3000 product with the included power supply, they should not release it without an adequate power supply. If the battery bank doesn't come with the product they shouldn't be able to say "those are the specs."
Because the specs are not what people are going unless they specify "must use with battery". So essentially they DON'T list the products in a way people are really going to use it.

If I were you, I would always use the included power supply, force them to actually include good performing PS.

 

[Bjorn]

The next business plan for Exasound is to sell a battery power supply for only $1000; taking their products to new heights!

 

[Final]

@Ron Kuper -Lynx Aurora is an 8 channel DAC you could consideras an alternative.

 

[Blumlein 88]

The next business plan for Exasound is to sell a battery power supply for only $1000; taking their products to new heights!

Kal mentioned he uses his with a battery for power.

 

[amirm]

Coming from the other forum

Amir I believe that you have the community best interest that's why I hope you'll see that this is NOT helping.

Two parties with different measurement results, who are unwilling to collaborate or even communicate directly to get to the bottom of this.

Hi Ron. Welcome to the forum. Your point is well taken. I just reached to George to see if he can replicate my latest tests with his analyzer using the same settings. I will report back on he says.

I am not home until later tomorrow so can't do any more measuring myself. Once back, I may try to see if I can replicate their measurements in any way.

 

[Panelhead]

If I was looking for an 8 channel dac the Focusrite Clarett 8Pre would be on the list. Like most it does not offer volume control for all eight channels of D to A conversion. There is a USB version for Windows and a Thunderbolt version for Mac users.

 

[Fitzcaraldo215]

Kal mentioned he uses his with a battery for power.

I thought it was an LPS.

 

[Blumlein 88]

I thought it was an LPS.

Can't remember if it was here or on CA, but he said he uses a battery, and has for I think two years.