This is a review and detailed measurements of Schiit modi 3 DAC. It is a replacement for two previous products: the Modi 2 and Modi 2 Uber. I purchased the unit a few days ago for $99 plus $11.26 for shipping for a total of $110.26.
The Modi 3 DAC has the functionality of the latter with three inputs: USB, S/PDIF and Toslink which is nice. A front toggle switch (as opposed to momentary in Modi 2 Uber) selects the input. Schiit has fixed the output voltage so that it is now the full 2 volts as opposed to 1.5 in previous products. They say this is enabled by addition of an external switching power supply which you see here on top of the unit:
For some reason I think having a black supply would go better with the black USB cord and overall gray look of the unit. Since this is a standard USB supply, I suppose you get your own version in however color or format you like.
Schiit says that this is the best measuring DAC they have designed, courtesy of purchasing the same analyzer I have: the Audio Precision APx555. So I expected my review to be boring but alas, that is not completely the case.
Let's get into the measurements and see how she did.
Measurements
As usual, I start with the Dashboard view of a 1 kHz tone at full amplitude using USB input:
While I was pleased by the nice, nearly 2 volt output, I was taken back by the spray of spurious tones in the FFT spectrum in top right. At first I thought the initial spike was at 120 Hz so I replaced the power supply with a linear and another switcher but none made a difference. I played with grounding and while that got me another dB or so in SINAD, it did nothing to resolve the spikes. I then performed a higher resolution FFT and realized that the frequency is actually 100 Hz, not 120 in which case, it is not AC mains related (we live in 60 Hz world so harmonics would be 120 Hz, etc).
EDIT: When feeding the signal from Roon through WASAPI, the problem is resolved. See this post: https://www.audiosciencereview.com/...rements-of-schiit-modi-3-dac.4742/post-104838
To see if this is a USB or core DAC issue, I tested the S/PDIF input:
Wow, the problem went away! So clearly this is a USB input issue. Strangely, when I look at the published Audio Precision report from Schiit, the output looks very close to my S/PDIF even though the report says they were testing USB input:
To troubleshoot further, I plugged in the Schiit Modi 2 Uber into same setup and got this:
At first blush, it may look like the 100 Hz spikes are not here but they are. The Modi 2 noise floor is a lot higher than Modi 3 and is hiding them on the left. If you look carefully though, a tiny spike is poking through. But more easily, we see the large spray in high frequencies on the right where the noise floor is lower, allowing them to show up. More confirmation of this later.
For now, we see 5 dB or so improvement in SINAD using USB over Modi 2. Had the 100 Hz spikes not been there and we were getting the S/PDIF performance, that would have been stellar at 105 dB. Putting that in context of recently reviewed DACs, we get this:
As it is, the cheaper Topping D10 gets 103 dB SINAD over USB versus just 96 for Schiit Modi 3. That aside, Schiit's statement is correct that it is the best measuring DAC they have ever produced, easily beating even their $2,400 Schiit Yggdrasil. It also beats the Modi 2 Uber even though it is cheaper.
Moving on to Jitter we have this:
The 100 Hz spikes in USB input (blue trace) are clearly visible in this high resolution spectrum display. They are gone with S/PDIF in red but then replaced with new ones (in red) around the main tone. Interestingly enough, S/PDIF also has lower noise floor than USB. So there is no clear winner between the two inputs here.
Comparing Modi 2 Uber to Modi 3 over USB, we get this:
Looking to the left, we see that both DACs have the same 100 Hz components. This tells me that they used the same USB subsystem as is from Modi 2. The 100 Hz may be created in my host computer and travelling into both DACs just the same.
Fortunately the rest of the news is positive with the output of Modi 3 being cleaner otherwise without the noise floor modulation (scalloping) that we see in Modi 2 Uber.
The 100 Hz problem dug Modi 3 Uber's grave in USB testing as I use a 200 Hz tone and the spikes completely corrupt those values when the source amplitude is small:
Switching to S/PDIF allows Modi 3 to nail this measurement due to absence of those 100 Hz tones.
Note that this says that the spurious 100 Hz are not level dependent. So they have more impact on the fidelity than if they were a ratio of our full amplitude signal.
If I had picked a different test tones, USB measurement could be shown to be as good as S/PDIF.
The Modi 3 DAC uses an AKM DAC so it does not suffer from intermodulation "hump" at mid-levels as budget ESS DACs tend to do:
These graphs are residual noise dominated so once again we see USB input (red) underperforming S/PDIF (green). Both beat Modi 2 Uber though (pink).
Topping D50 is quieter (blue) but as mentioned, it suffers from higher distortion at mid-levels.
Frequency response is off 0.6 dB at 20 Khz:
I think that is it.
Conclusions
The S/PDIF performance of Schiit Modi 3 Uber put a smile on my face and easily garners my recommendation for that use. The DAC is a bargain at that price with that level of performance. USB input unfortunately falls behind in my testing, likely due to sensitivity to input noise. Since Schiit doesn't show that problem in their testing, your specific host and DAC configuration may perform better.
EDIT: Using WASAPI instead of ASIO4ALL wrapper (which my analyzer requires) resolved the issue with USB. As such, I am recommending the Schiit Modi 3 without reservations.
I will test Schiit Modi 2 Multibit against this later but for now, I am confident to say that this is the best Modi option to purchase.
On a more general note, it is wonderful to see Schiit taking objective measurements and good engineering practices to heart. We have the tools to make sure our designs are right and a company the size of Schiit needs to design products in this manner. I expect better execution from Schiit in future products.
-----
If you like this review, please consider donating funds 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).
The Modi 3 DAC has the functionality of the latter with three inputs: USB, S/PDIF and Toslink which is nice. A front toggle switch (as opposed to momentary in Modi 2 Uber) selects the input. Schiit has fixed the output voltage so that it is now the full 2 volts as opposed to 1.5 in previous products. They say this is enabled by addition of an external switching power supply which you see here on top of the unit:
For some reason I think having a black supply would go better with the black USB cord and overall gray look of the unit. Since this is a standard USB supply, I suppose you get your own version in however color or format you like.
Schiit says that this is the best measuring DAC they have designed, courtesy of purchasing the same analyzer I have: the Audio Precision APx555. So I expected my review to be boring but alas, that is not completely the case.
Let's get into the measurements and see how she did.
Measurements
As usual, I start with the Dashboard view of a 1 kHz tone at full amplitude using USB input:
While I was pleased by the nice, nearly 2 volt output, I was taken back by the spray of spurious tones in the FFT spectrum in top right. At first I thought the initial spike was at 120 Hz so I replaced the power supply with a linear and another switcher but none made a difference. I played with grounding and while that got me another dB or so in SINAD, it did nothing to resolve the spikes. I then performed a higher resolution FFT and realized that the frequency is actually 100 Hz, not 120 in which case, it is not AC mains related (we live in 60 Hz world so harmonics would be 120 Hz, etc).
EDIT: When feeding the signal from Roon through WASAPI, the problem is resolved. See this post: https://www.audiosciencereview.com/...rements-of-schiit-modi-3-dac.4742/post-104838
To see if this is a USB or core DAC issue, I tested the S/PDIF input:
Wow, the problem went away! So clearly this is a USB input issue. Strangely, when I look at the published Audio Precision report from Schiit, the output looks very close to my S/PDIF even though the report says they were testing USB input:
To troubleshoot further, I plugged in the Schiit Modi 2 Uber into same setup and got this:
At first blush, it may look like the 100 Hz spikes are not here but they are. The Modi 2 noise floor is a lot higher than Modi 3 and is hiding them on the left. If you look carefully though, a tiny spike is poking through. But more easily, we see the large spray in high frequencies on the right where the noise floor is lower, allowing them to show up. More confirmation of this later.
For now, we see 5 dB or so improvement in SINAD using USB over Modi 2. Had the 100 Hz spikes not been there and we were getting the S/PDIF performance, that would have been stellar at 105 dB. Putting that in context of recently reviewed DACs, we get this:
As it is, the cheaper Topping D10 gets 103 dB SINAD over USB versus just 96 for Schiit Modi 3. That aside, Schiit's statement is correct that it is the best measuring DAC they have ever produced, easily beating even their $2,400 Schiit Yggdrasil. It also beats the Modi 2 Uber even though it is cheaper.
Moving on to Jitter we have this:
The 100 Hz spikes in USB input (blue trace) are clearly visible in this high resolution spectrum display. They are gone with S/PDIF in red but then replaced with new ones (in red) around the main tone. Interestingly enough, S/PDIF also has lower noise floor than USB. So there is no clear winner between the two inputs here.
Comparing Modi 2 Uber to Modi 3 over USB, we get this:
Looking to the left, we see that both DACs have the same 100 Hz components. This tells me that they used the same USB subsystem as is from Modi 2. The 100 Hz may be created in my host computer and travelling into both DACs just the same.
Fortunately the rest of the news is positive with the output of Modi 3 being cleaner otherwise without the noise floor modulation (scalloping) that we see in Modi 2 Uber.
The 100 Hz problem dug Modi 3 Uber's grave in USB testing as I use a 200 Hz tone and the spikes completely corrupt those values when the source amplitude is small:
Switching to S/PDIF allows Modi 3 to nail this measurement due to absence of those 100 Hz tones.
Note that this says that the spurious 100 Hz are not level dependent. So they have more impact on the fidelity than if they were a ratio of our full amplitude signal.
If I had picked a different test tones, USB measurement could be shown to be as good as S/PDIF.
The Modi 3 DAC uses an AKM DAC so it does not suffer from intermodulation "hump" at mid-levels as budget ESS DACs tend to do:
These graphs are residual noise dominated so once again we see USB input (red) underperforming S/PDIF (green). Both beat Modi 2 Uber though (pink).
Topping D50 is quieter (blue) but as mentioned, it suffers from higher distortion at mid-levels.
Frequency response is off 0.6 dB at 20 Khz:
I think that is it.
Conclusions
The S/PDIF performance of Schiit Modi 3 Uber put a smile on my face and easily garners my recommendation for that use. The DAC is a bargain at that price with that level of performance. USB input unfortunately falls behind in my testing, likely due to sensitivity to input noise. Since Schiit doesn't show that problem in their testing, your specific host and DAC configuration may perform better.
EDIT: Using WASAPI instead of ASIO4ALL wrapper (which my analyzer requires) resolved the issue with USB. As such, I am recommending the Schiit Modi 3 without reservations.
I will test Schiit Modi 2 Multibit against this later but for now, I am confident to say that this is the best Modi option to purchase.
On a more general note, it is wonderful to see Schiit taking objective measurements and good engineering practices to heart. We have the tools to make sure our designs are right and a company the size of Schiit needs to design products in this manner. I expect better execution from Schiit in future products.
-----
If you like this review, please consider donating funds 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:
