This will be a run down on the Schitt Modi Multi-bit, test review sample provided by a forum member.
This unit's claim to fame is being the lowest cost DAC on the market with a closed form HF upsamping filter (original samples are kept, oversampling interpolates between them), and a true multi-bit DAC chip, the AD5547, a 16 bit resolution multiplying DAC from Analog devices.
The Modi offers Optical, S/PDIF, and USB inputs, with stereo unbalanced RCA outputs.
Basic specs are on the Schiit site; THD 10-20kHz is rated at 0.006% or less at full rated output.
All testing was done on the S/PDIF RCA input.
First, a frequency response sweep was performed, 20-20kHz. Seems basic, but you might be surprised at what I've seen just on that. Drive level was -6dBFS (that means, 6dB below full scale output in digital terms)
Response is down only a about 4 tenths of a dB at 20kHz, a fairly good result in this price class.
To get an initial idea of low level nonlinearity and noise performance, I usually go next to 1kHz at -60 dB, to check for idle tones (common in Delta-Sigma converters) and look at the noise floor- to see idle tones or other artifacts clearly, I use a large FFT sampling size and average 20 samples, which filters out random noise.
FFT
Scope Mode
This is not as clean as I would like to see- one of the challenges with lower cost true multi-bit DACs is switching spikes when going through the zero crossing region.
Here's that 1kHz at -60dBFS in a narrower measurement span, ending at 2 kHz.
Now, let's put things in perspective, these artifacts are at about -110 dB or lower; we could simply argue that having a very good analyzer makes things look a bit worse than they are...
If we drop the digital input level further, to -70 dBFS, the situation, or problem, becomes clearer, though.
Here, the scale transition spikes are quite clear.
To confirm if the low level spikes are all harmonically related, I repeated this test with two new lower test frequencies, 50Hz and 100Hz. The spacing interval should change depending on the fundamental, and it does.
The other low level test I usually do is 1kHz sine at -90dB, both for 16bit data and 24 bit- as this is a 16bit DAC, the latter test isn't possible.
Normally with a good DAC the result is a stepped staggered up and down pulse wave, as -90 dBFS is just toggling the LSB of the DAC. Here we have something I've never seen before... (I probably just don't get out enough...)
This is a good example of what it should look like, on a Cambridge Audio 851D, which has a fairly sophisticated implementation of the AK4490.
Last up, let's take a look at the high frequency CCIM test (19+20kHz Intermodulation Distortion). This is pretty demanding, with inputs at -6dB at 19 and 20kHz.
This is what the output looks like in scope mode, showing how the signals beat together and reach a much higher peak level when they sum in phase (which is why each individual signal is no more than -6dB).
And the FFT spectrum-
This result is fairly good- the difference frequency component at 1kHz is down about -82 dB, and the image components near 25 kHz and 38kHz are mostly under -80dB. Possibly more important, there are virtually no side bands around the 19 and 20kHz product.
Now, this is only the 2nd low cost DAC I've tested on the AP, and I don't have the digital interface necessary for testing with phones or iPods, so It's hard for me to put in fairly in perspective. So I'll let the results speak for themselves, and post results on another low cost DAC sent to me to test...
I will say that the very low level performance is somewhat of concern, and I've also heard that even the Schiit Yiggy has some issues with zero crossing behavior, but not at all to this degree. For 4x the money, (I paid $999 for the one I bought new) one does get a very large step up in performance and capability with the Cambridge 851D, including a sophisticated output volume control. .
This unit's claim to fame is being the lowest cost DAC on the market with a closed form HF upsamping filter (original samples are kept, oversampling interpolates between them), and a true multi-bit DAC chip, the AD5547, a 16 bit resolution multiplying DAC from Analog devices.
The Modi offers Optical, S/PDIF, and USB inputs, with stereo unbalanced RCA outputs.
Basic specs are on the Schiit site; THD 10-20kHz is rated at 0.006% or less at full rated output.
All testing was done on the S/PDIF RCA input.
First, a frequency response sweep was performed, 20-20kHz. Seems basic, but you might be surprised at what I've seen just on that. Drive level was -6dBFS (that means, 6dB below full scale output in digital terms)
Response is down only a about 4 tenths of a dB at 20kHz, a fairly good result in this price class.
To get an initial idea of low level nonlinearity and noise performance, I usually go next to 1kHz at -60 dB, to check for idle tones (common in Delta-Sigma converters) and look at the noise floor- to see idle tones or other artifacts clearly, I use a large FFT sampling size and average 20 samples, which filters out random noise.
FFT
Scope Mode
This is not as clean as I would like to see- one of the challenges with lower cost true multi-bit DACs is switching spikes when going through the zero crossing region.
Here's that 1kHz at -60dBFS in a narrower measurement span, ending at 2 kHz.
Now, let's put things in perspective, these artifacts are at about -110 dB or lower; we could simply argue that having a very good analyzer makes things look a bit worse than they are...
If we drop the digital input level further, to -70 dBFS, the situation, or problem, becomes clearer, though.
Here, the scale transition spikes are quite clear.
To confirm if the low level spikes are all harmonically related, I repeated this test with two new lower test frequencies, 50Hz and 100Hz. The spacing interval should change depending on the fundamental, and it does.
The other low level test I usually do is 1kHz sine at -90dB, both for 16bit data and 24 bit- as this is a 16bit DAC, the latter test isn't possible.
Normally with a good DAC the result is a stepped staggered up and down pulse wave, as -90 dBFS is just toggling the LSB of the DAC. Here we have something I've never seen before... (I probably just don't get out enough...)
This is a good example of what it should look like, on a Cambridge Audio 851D, which has a fairly sophisticated implementation of the AK4490.
Last up, let's take a look at the high frequency CCIM test (19+20kHz Intermodulation Distortion). This is pretty demanding, with inputs at -6dB at 19 and 20kHz.
This is what the output looks like in scope mode, showing how the signals beat together and reach a much higher peak level when they sum in phase (which is why each individual signal is no more than -6dB).
And the FFT spectrum-
This result is fairly good- the difference frequency component at 1kHz is down about -82 dB, and the image components near 25 kHz and 38kHz are mostly under -80dB. Possibly more important, there are virtually no side bands around the 19 and 20kHz product.
Now, this is only the 2nd low cost DAC I've tested on the AP, and I don't have the digital interface necessary for testing with phones or iPods, so It's hard for me to put in fairly in perspective. So I'll let the results speak for themselves, and post results on another low cost DAC sent to me to test...
I will say that the very low level performance is somewhat of concern, and I've also heard that even the Schiit Yiggy has some issues with zero crossing behavior, but not at all to this degree. For 4x the money, (I paid $999 for the one I bought new) one does get a very large step up in performance and capability with the Cambridge 851D, including a sophisticated output volume control. .
Comment