• 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!

Review, Measurement and Comparison of Schiit Modi 2 Uber and Topping D30

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,384
Location
Seattle Area
This is a detailed review and measurements of the popular Schiit Modi 2 Uber DAC and comparison to Topping D30. I own both units. The Modi 2 Uber edition is one of the three versions of this DAC and retails for $149 plus shipping. The Topping D30 is available on Amazon Prime for $119 and includes shipping so good bit cheaper.

Topping D30 versus Schiit Modi 2 DAC review.jpg


The two are similar in features in that both have three inputs: USB, S/PDIF coax and Toslink optical inputs. Both have external power bricks. The Topping has its power switch on the front which I prefer to the back switch of Modi 2 Uber. I also like the English labels on Topping as opposed to alien symbols used in Schiit Modi 2 Uber.

A while back I reviewed the plain Modi 2 and unfortunately it did not perform well. This prompted people to ask me to measure the Uber edition since it has an external power supply and as such, should not be affected by the USB power as Modi 2 was.

The Topping D30 was also reviewed earlier but this is an extended set of measurements than I used to perform then.

So let's get into the data. As usual, if you are not familiar with what these graphs mean, refer to my tutorial on understanding audio measurements: https://www.audiosciencereview.com/forum/index.php?threads/understanding-audio-measurements.2351/

Measurements
Let's start with the test of Jitter and Noise using the USB input. The test signal which is essentially a 12 kHz tone is played by Roon in exclusive mode and the output of each DAC is measured by my analyzer. Anything other than the 12 kHz tone is unwanted:

Topping D30 versus Schiit Modi 2 Uber Jitter and Noise Measurement over USB.png


The Topping D30 has a nice flat noise floor with a couple of tiny spikes here and there. The Schiit Modi 2 Uber on the other hand suffers from three maladies:

  1. Noise spikes at lower frequencies, possibly caused by power supply mains
  2. Rising noise level "skirt" around our main 12 kHz tone, indicating random clock jitter
  3. Lower output level by nearly 3 dB. This means its dynamic range is actually worse than what is shown in the graph
Fortunately the levels in #1 and #2 from audibility point of view are pretty low so not a major concern. Just lack of good engineering at play.

Let's look at the distortion and noise from a 1 kHz tone that itself is filtered (i.e. everything shown is unwanted):

Topping D30 versus Schiit Modi 2 DAC 1 kHz Residual Noise and Distortion Measurement.png


We see that the Modi 2 Uber has much higher noise at lower frequencies (< 6 kHz) where our hearing is most sensitive. It also has higher harmonic distortion spikes than Topping D30 with the exception of 5th harmonic.

We can put this in the form of classical THD+N measurement by having my analyzer add up the noise and distortion contributions:

Topping D30 versus Schiit Modi 2 Uber THD+N Measurement.png


We see that the Modi 2 Uber underperforms Topping D30 by 7 dB. But wait, there is more!

As I was running this test the bar graph kept jumping from time to time when analyzing Modi 2 Uber. I was puzzled by this so tested the Topping D30 again and it was rock solid. As you see, Modi 2 Uber min THD+N is -96 dB but then increases way up to -59 dB (smaller numbers are worse). The D30 on the other hand only varies between -103 and -102 dB.

To see what was going on I ran the previous spectrum measurement of 1 kHz in repeat mode on Modi 2 Uber and got this:

Topping D30 versus Schiit Modi 2 DAC Repeat Noise and Distortion Measurement.png


Bottom group of graphs shows the lower distortion level. But once in a while, the distortion and noise jump up as shown in green and yellow. So clearly something is wrong. Way wrong.

I then hooked up the Modi 2 to my Stax headphone amp and listened. The problem became super obvious: the DAC would play the 1 kHz tone cleanly for some 8 to 10 seconds and then there would be a soft glitch/static. Sometimes it would do this once. Other times it would do it twice and once it did it three times in a row!

I then tested the S/PDIF input using my generator as the source and there was no glitch at all. This leads to the conclusion that the Schiit Modi 2 Uber has a broken USB implementation. This is completely repeatable on my system.

Before any owners say they have this unit and don't have this problem, I played a bunch of music with the DAC and I could NOT detect the glitch there. Whether it was masked by music or is a problem that occurs at 0 dBFS, I don't know. What I do know is that in identical situation with dozens of DACs including the Topping D30, this problem does not occur.

Let's move on and test DAC linearity over S/PDIF. Here we send a digital sample to the DAC and keep changing its value linearity and see if the analog voltage coming out of the DAC does the same. An ideal DAC would produce a flat line at zero with no deviation:

Topping D30 versus Schiit Modi 2 Uber Linearity Measurement.png


The Schiit Modi 2 Uber turns in pretty poor performance. Using my self-imposed 0.1 dB error limit, it is only able to output a signal that is less than this error at 72 dB and higher. This translates into 12 bits of clean output. The Topping D30 on the other hand marches on to -90 dB which is equivalent to 15 bits. This is significantly better performance than Schiit Modi 2 Uber.

The picture is even worse than that. At -120 dB, the error from Schiit Modi 2 Uber is 15 dB. At the same level, the Topping D30 error is just 2.5 dB.

Feeding both DACs a very low level -90 dB sine wave generates this:

Topping D30 versus Schiit Modi 2 DAC -90 dB Sine Wave Measurement.png


While the Modi 2 Uber sine wave is pretty clean, its level has dropped by a fair bit which reflects poorly in linearity tests (noise dominates over the signal).

I could run more tests but I think the picture here is quite consistent in each measurement. Schiit Modi 2 trails the Topping D30 by a significant margin.

Subjective Listening Tests
I connected both DACs to the inputs of my Stax SRM-007t driving the Omega pro headset. Using its AB switch I could instantly go between the two DACs while Roon played the same music to both.

There was no comparison here: the Topping D30 had far more authority, detail, bass impact, air, etc. In every way it was superior to the Schiit Modi 2 Uber. A child could tell the difference.

What is that you say? You thought DACs don't sound very different? Well, that is the case if you match their levels. The Topping has nearly 3 dB higher output than the Schiit Modi 2. So if this is not compensated for, the extra level absolutely translates into better fidelity.

Why did I run the test knowing what I know about the levels? This was to dispute the subjective tests online where people say Topping D30 doesn't sound as good as Schiit Modi 2 Uber. That is an impossibility since their testing is like above without level matching. If they can't hear the louder DAC, i.e. Topping D30 sounding better, they need to turn in their audiophile card. It is simple as that.

To match levels, I used the Roon DSP volume control and notched down the Topping D30. Unfortunately the resolution was too coarse. I could either have it be a bit louder than Modi 2 Uber or a bit less loud. Regardless, almost all of the above differences I heard when I didn't level match went out the window. Both DACs sounded pretty good. Trying hard I thought the Schiit Modi 2 Uber sounded a tiny bit brighter. This is not a very reliable observation so please don't hold me to it.

Conclusions
The measurements show that the Schiit Modi 2 Uber DAC is simply not competitive in performance. It has lower output, higher noise and jitter and much worse linearity than Topping D30.

Worst yet, its USB implementation appears to be faulty causing glitches in the output.

Time and time again when I test Schiit products they demonstrate lack of design verification and targets. The basics are not done. Just aren't. The company needs a reboot in its development process and must establish competitive targets and strive to achieve them in new products. Just because a product produces music, it doesn't mean it is working correctly.

Fortunately our hearing is nowhere as sensitive as our instrumentation. So if you own a Modi 2 Uber, I am not telling you to run out and buy something else. But if you are in the market for a new DAC, I am compelled to recommend buying a different product whether that is Topping D30 or other DACs in this price range. They have all gone though proper engineering and design verification to produce performance that at least meet the CD specification in noise and resolution.


As always, questions, comments, corrections, etc. are all welcome.

----
If you like this review, please consider donating funds for these types of hardware purchase 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).
 
Last edited:

Wombat

Master Contributor
Joined
Nov 5, 2017
Messages
6,722
Likes
6,459
Location
Australia
Mcdac's.
 

nuko

New Member
Joined
Apr 30, 2018
Messages
4
Likes
2
Great review, Schiit is pretty much out of the question for me. It definitely doesn't help that their customer service is apparently Schiit as well...

But Magni 3 and O2 review/comparison soon, @amirm? :D It'd definitely go well with the recent reviews/comparisons like this one.
 

Duckeenie

Active Member
Joined
Apr 7, 2018
Messages
103
Likes
112
This is new...
 

sonci

Active Member
Joined
Feb 18, 2018
Messages
233
Likes
112
Amirm, do you think Topping would benefit from a linear power supply?
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,384
Location
Seattle Area
Great review Amir! Since it appears that in this dac, it is probably a better option to use spdif than usb, I am wondering if there are spdif jitter issues like the SMSL sanskrit had here https://audiosciencereview.com/forum/index.php?threads/battle-of-s-pdif-vs-usb-which-is-better.1943/
Thanks. :) Note the the linearity measurement was made using S/PDIF so even there, performance of Modi 2 Uber is not competitive.

I did test the Modi 2 Uber against three different USB to S/PDIF converters here: https://audiosciencereview.com/foru...s-of-smsl-x-usb-digital-audio-converter.2432/

As you see, the low frequency noise components can be removed but the random jitter cannot be eliminated.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,384
Location
Seattle Area
Amirm, do you think Topping would benefit from a linear power supply?
From memory, I tested it and it did make a small improvement. But it is not an audible difference. So I would not bother.
 

mindbomb

Active Member
Joined
Nov 11, 2017
Messages
284
Likes
175
Thanks. :) Note the the linearity measurement was made using S/PDIF so even there, performance of Modi 2 Uber is not competitive.

I did test the Modi 2 Uber against three different USB to S/PDIF converters here: https://audiosciencereview.com/foru...s-of-smsl-x-usb-digital-audio-converter.2432/

As you see, the low frequency noise components can be removed but the random jitter cannot be eliminated.

oh wow, I had assumed the random jitter had something to do with their usb receiver, but it happens regardless of input.
 

Johnny2Bad

Member
Joined
Dec 22, 2017
Messages
42
Likes
28
You've got a typo (why do these things stick out like neon to me and others read along and don't notice? One of the mysteries of life).

" ... The Topping D30 on the other hand marches on to -90 dB which is equivalent to 15 db. ..."

Should be:

... equivalent to 15 bits.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,384
Location
Seattle Area
You've got a typo (why do these things stick out like neon to me and others read along and don't notice? One of the mysteries of life).

" ... The Topping D30 on the other hand marches on to -90 dB which is equivalent to 15 db. ..."

Should be:

... equivalent to 15 bits.
Thanks. I made the correction. It is one of those things when you write something, you read it the way you want, rather than what it is :).
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,384
Location
Seattle Area
Hi there. Welcome to the forum.

Thanks for the link. Let me start by saying that I am happy to see others measure and try to replicate my measurements. Agree or disagree, it is a good thing to see more people paying attention to such data.

As to their measurements, there is actually fair amount of agreement. Alas, it is hard to see due to way they have presented the data. Let's take DAC linearity:

index.php


This seems to show that the Schiit Modi 2 Uber is doing well. But it may not be.

Let's step back and understand this measurement. The system generates digital audio samples and then keeps reducing their values and measures the analog output of the DAC. If I halve the digital sample value for example, the analog output of the DAC better follow. Any variation is an error from ideal performance of a DAC. If we plot the input digital sample values against the output analog voltage, we would get a graph like they are showing above. An ideal DAC would have a perfectly flat line.

So what is the problem in the above graph? Lack of resolution. We are traversing some 120 db in so few pixels. So it is hard to see any deviations. For this reason, the standard linearity graph does NOT show the above line. Instead, we show the error from ideal value. Since the error range is small, we can zoom way in and see any variations.

Let me demonstrate that with my measurements where now I show both the graph above and the variation I showed in my review:

Topping D30 versus Schiit Modi 2 DAC Linearity Expanded Measurement.png


The green lines are the same as what they are showing. As we see, they look of similar. But look at the red line which shows the error. In the case of Schiit Modi 2 Uber on the left, the error starts at just -72 db. Look at the part of the green graph and you cannot with naked eye see that kind of error. One would think that the green line is perfect up to some -110 dB but that is not remotely the case.

Alas, not all measurement systems can show the deviation graph. Without it, the linearity information is next to impossible to interpret except for extreme cases of errors.

So in summary, their graph may very well be in full agreement with mine. It is just that it is impossible to see it there. Wrong presentation for this type of data.

Note again that what I am presenting is industry standard. See for example this from many such measurements in Stereophile magazine:
1295EADFIG05.jpg


Notice that it is only showing the error graph just as I do.
 

Johnny2Bad

Member
Joined
Dec 22, 2017
Messages
42
Likes
28
Thanks. I made the correction. It is one of those things when you write something, you read it the way you want, rather than what it is :).

Oh, I understand completely when you're the author. There were a number of comments and I'm sure many readers who didn't comment, before I posted though. Not much gets by many of the active posters here but typos seem to be one. In my normal web-surfing habits, I send a lot of eMails to webmasters who operate professionally designed and edited sites, often large corporations who presumably get good help at presumably good salaries. Yet I see every one.

For some reason, news sites are the worst, but I don't bother notifying them. It's their job, and if they do it poorly, they deserve to have it hang out there for everyone to see. But a site like this, unpaid and community based, I will, because as part of the community I feel it's upon me to help.

I had a boss, husband & wife team, and there was a large post-and-beam hall where you needed four tiers of scaffolding to change the light bulbs. She could see spider webs in the farthest corners of the building from 80 feet away, and would direct people to brush them off. I once mentioned to her about her "excellent vision" and she replied "It's not a skill, it's a disease."

-Regards
 
Last edited:

Shazb0t

Addicted to Fun and Learning
Joined
May 1, 2018
Messages
643
Likes
1,230
Location
NJ
Hi there. Welcome to the forum.

Thanks for the link. Let me start by saying that I am happy to see others measure and try to replicate my measurements. Agree or disagree, it is a good thing to see more people paying attention to such data.

As to their measurements, there is actually fair amount of agreement. Alas, it is hard to see due to way they have presented the data. Let's take DAC linearity:

index.php


This seems to show that the Schiit Modi 2 Uber is doing well. But it may not be.

Let's step back and understand this measurement. The system generates digital audio samples and then keeps reducing their values and measures the analog output of the DAC. If I halve the digital sample value for example, the analog output of the DAC better follow. Any variation is an error from ideal performance of a DAC. If we plot the input digital sample values against the output analog voltage, we would get a graph like they are showing above. An ideal DAC would have a perfectly flat line.

So what is the problem in the above graph? Lack of resolution. We are traversing some 120 db in so few pixels. So it is hard to see any deviations. For this reason, the standard linearity graph does NOT show the above line. Instead, we show the error from ideal value. Since the error range is small, we can zoom way in and see any variations.

Let me demonstrate that with my measurements where now I show both the graph above and the variation I showed in my review:

View attachment 12713

The green lines are the same as what they are showing. As we see, they look of similar. But look at the red line which shows the error. In the case of Schiit Modi 2 Uber on the left, the error starts at just -72 db. Look at the part of the green graph and you cannot with naked eye see that kind of error. One would think that the green line is perfect up to some -110 dB but that is not remotely the case.

Alas, not all measurement systems can show the deviation graph. Without it, the linearity information is next to impossible to interpret except for extreme cases of errors.

So in summary, their graph may very well be in full agreement with mine. It is just that it is impossible to see it there. Wrong presentation for this type of data.

Note again that what I am presenting is industry standard. See for example this from many such measurements in Stereophile magazine:
1295EADFIG05.jpg


Notice that it is only showing the error graph just as I do.
What do you make of the FFT measurements showing the humps at 9kHz and 15kHz on the 12kHz tone chart for the D30 and the higher noise floor of the D30 in comparison to your measurements?
 

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
Thanks. I made the correction. It is one of those things when you write something, you read it the way you want, rather than what it is :).

So, D30 has 18-bits resolution on its USB input and only 15-bits resolution on SPDIF input?
 
Top Bottom