Hello Everyone,
This is a review and detailed measurements of the Fiio DM13 portable CD player.
Fiio DM13 - Presentation
I'm not sure I need to provide an extensive presentation as everything is available on the vendor's website. In a nutshell, this is a portable CD player that plays on batteries and be used to rip your CDs too.
The one I got is the non-bluetooth version, and I like the fact it offers unbalanced and balanced outputs, as well as a digital one via a a single 3.5mm jack output. Note that a 3.5mm to RCA coax converter is provided and you'll need to buy another one if you want to use the optical out.
I bought a 4.4mm to 3.5mm balanced cable from Fiio, to perform the below measurements.
I must admit it was real fun to be back to use a pseudo "Discman". That reminded me of when I was young. I could appreciate a good resistance to shocks. I had pleasure listening to it and I could not spot obvious flaws.
Fiio DM13 - Measurement (Analog out)
I performed all measurements as I described them in the post “More than we hear”, and as I reported them for the Onkyo C-733 review. Over time, this will help comparing the devices I reviewed.
The Fiio DM13 outputs a high 5Vrsm from balanced outputs (I did not test unbalanced). There was a slight channel imbalance of 0.17dB which is decent. The balanced outputs are noninverting.
All the below measurements are from balanced output with volume down to 4.2Vrms, to kind of align with Amir's standard. But the measurements are the same at full 5Vrms output, it made no differences.
Let's start with the standard 999.91Hz sine @0dBFS (without dither) from the test CD:
The Fiio DM13 does a little better than its published specs with 0.0015% of THD+N (0.002% published). The resolution is indeed 16bits.
That said, note these spikes of odd harmonics (H3, H5, H7, ...). That reminded me of some issue we already discussed here when trying the burn a test CD with Windows Media Player. The distorsion you see here does no look like coming from the conversion itself... So, let's investigate.
I usually add the same measurement at -6dBFS:
Ah, again the same distorsion profile (odd harmonics slowly decreasing) at a lower level too, that is very uncommon. With this, we can suspect a digital change of the signal prior to conversion, some sort of level equalization as we saw with WMP. Using shaped dither, we can lower the noise floor to analyze deeper.
Below is the same 999.91Hz @0dBFS but with shaped dither (hence the double wave of low level noise you see starting at 6kHz):
Well, here we reveal an effect of processing the digital signal before conversion. Too bad because I think otherwise the DAC in use here is very good (unless this is a function of it, but I doubt).
Let's park this issue for the moment and move on to the good news. First, this is an very silent player:
No visible components from mains, this is perfect and too rare. You might argue with the previous view that was showing a 50Hz spike, but that was my error, the PC in use for measurements was not on batteries and I got lazy to run the measurement again, sorry.
Let's continue with the good news as the bandwidth is very flat:
You can see the channel mismatch (0.17dB) which is not a concern. The roll off at the end is an effect of the oversampling filter in use.
Talking oversampling filter, this is how it works:
The filter attenuates by -100dB maximum, which is decent but we see noise shaping from the conversion. So a lot of noise is created outside audio band, which is common with DSD (SACD) for instance. There might be an upsampling to DSD happening before conversion.
Multitone (1/10 decade) shows a happy CD player, not having issue to clear 16bits of data:
The Jitter test is relatively good:
We can see some unwanted artifacts but they will remain hidden into music.
Started with the Teac VRDS-20 review, and on your request + support to get it done (more here), I'm adding an "intersample-overs" test which intends to identify the behavior of the digital filtering and DAC when it come to process near clipping signals. Because of the oversampling, there might be interpolated data that go above 0dBFS and would saturate (clip) the DAC and therefore the output. And this effect shows through distorsion (THD+N measurement up to 96kHz):
I kept some references and will keep the same for other reviews, so you can quickly compare. The results of the Fiio DM13 mean the oversampling filter has 3dB+ headroom when outputting 4Vrms. This is very good but means there is digital attenuation. That said, even a full output (5Vrms), there is still 2dB headroom. So I would again suspect the presence of an SRC used to perform digital attenuation, or something equivalent. For once I've put in parenthesis the THD (without noise) because the noise shaping is added into the calculation and would look like beginning of clipping but it's not.
Stereophile was often using the 3DC measurement as a prof of low noise DAC. It is from an undithered 997Hz sine at -90.31dBFS. With 16bits, the signal should appear (on a scope) as the 3DC levels of the smallest symmetrical sign magnitude digital signal. Here is the result with the DM13:
Houston, we have a problem. What we see here is asymmetry of the output that I've never seen before. This view means the output is far from the requested -90.31dBFS, and actually I measured it at -95.37dBFS. So that is a lack of linearity by 5dB, and that is huge. It does not get much better if I set the output at 5Vrms because the deviation is still a massive 4dB!
Fact is the DM13 looses linearity as early as -70dBFS (by 0.7dB), while we are used to see linearity nailed down to -120dBFS these days! This is equivalent to what I saw with the Marantz CD-73, which had a 14bits DAC. So this is really bad results for the DM13.
So we have the culprit here, I believe. There is processing of the signal which deteriorates it at low levels and that is too bad.
Other measurements (not shown):
Note that the player does not apply de-emphasis filter if needed,
Fiio DM23 - Testing the drive
What would be good measurements if the drive would not properly read a slightly scratched CD, or one that was created at the limits of the norm? The below tests reply to these questions.
The drive was able to consistently continue playing, and without generating typical digital clicks, with dropouts of up to 3mm. The interpolation effect remained hidden to my hears when it kicked-off and the DM13 did not fail to maintain a constant flow with up to 4mm dropouts. It is the maximum I can measure, so that is near perfection. The Fiio had no issue with variable linear velocity and/or track pitch, as well as with HF detection. This is a crazy good drive!
Fiio DM13 - Measurements (Optical Out)
I've seen several of you reviewing CD players using their digital outputs, in case the results could be improved from an external DAC.
The lack of precision of the clock means it was nearly impossible to get a clean FFT as windowing errors kept on repeating because there's no PLL with the digital receiver when I run this measurement. Reducing the FFT length and without averaging, I got the below (999.91Hz @0dBFS):
We find again the distorsion we already saw, with a set of odd harmonics that are not on the test CD. This means there is digital modification also with the digital output. This is not "bit perfect" by far.
By the way, our 3DC levels are also impacted, this is again asymmetry and therefore lack of linearity in the digital domain:
I never saw that before. This drives me to believe there is processing before the signal reaches the DAC and also before it reaches the output. What for? I don't know, but this is too bad, even if I could not hear any negative impact. I guess there is some resampling via an ASRC or something because the player has also a Bluetooth option (well, not mine) and maybe that's to simplify the processing downstream. Again I don't know, but it's not good.
EDIT: I did not think before about running the below test, due to the imprecision of the clock. But this is a bandwidth analysis from periodic white noise done with 600+ averages and smoothed. It provides an interesting result (Linear Frequency Scale from 20Hz to 20kHz):
I compared with a CD Player, the Denon DCD-3560, both from their digital outputs (optical). The Denon (purple trace) shows what we expect, a straight line in digital domain. But the Fiio (blue trace) exhibits the typical ringing of a low resolution digital filter. This is what we had more than 40 years ago, with the first 2x oversampling digital filters. This means there's an interpolator, most probably a sample rate convertor, performing a resampling (even if it is resampling from 44.1kHz to 44.1khz) in band limited, of the digital signal, and prior to send it to the digital output. This is not desired.
Conclusion
The Fiio DM13 is extremely good at reading a scratched CD and following a track without failing. It's the best I measured in that area. It's the only player able to read correctly with 3mm dropouts on the disc and only few click with 4mm, amazing.
It is therefore too bad to see such digital processing and degradation of the signal. I think this is a software error and I hope a firmware can fix that. I will update if that's the case.
In the meantime, I can't say the flaws shown can be heard. They remained hidden to my hears. But this little player missed a very high mark because of some undesired processing. Too bad.
I left a message on Fiio’s forum to request if it’s ok to use their recent firmware update which theoretically applies only to the Bluetooth version. If they confirm, I’ll update and cross my fingers that it solves this processing issue.
Cheers
PS: Message to Fiio engineering: Please fix that software error with digital attenuation calculation, and you have the best CD Player/Transport of all times (besides clock, but that will be for the next iteration/super high-end version of this player). Pretty please?
This is a review and detailed measurements of the Fiio DM13 portable CD player.
Fiio DM13 - Presentation
I'm not sure I need to provide an extensive presentation as everything is available on the vendor's website. In a nutshell, this is a portable CD player that plays on batteries and be used to rip your CDs too.
The one I got is the non-bluetooth version, and I like the fact it offers unbalanced and balanced outputs, as well as a digital one via a a single 3.5mm jack output. Note that a 3.5mm to RCA coax converter is provided and you'll need to buy another one if you want to use the optical out.
I bought a 4.4mm to 3.5mm balanced cable from Fiio, to perform the below measurements.
I must admit it was real fun to be back to use a pseudo "Discman". That reminded me of when I was young. I could appreciate a good resistance to shocks. I had pleasure listening to it and I could not spot obvious flaws.
Fiio DM13 - Measurement (Analog out)
I performed all measurements as I described them in the post “More than we hear”, and as I reported them for the Onkyo C-733 review. Over time, this will help comparing the devices I reviewed.
The Fiio DM13 outputs a high 5Vrsm from balanced outputs (I did not test unbalanced). There was a slight channel imbalance of 0.17dB which is decent. The balanced outputs are noninverting.
All the below measurements are from balanced output with volume down to 4.2Vrms, to kind of align with Amir's standard. But the measurements are the same at full 5Vrms output, it made no differences.
Let's start with the standard 999.91Hz sine @0dBFS (without dither) from the test CD:
The Fiio DM13 does a little better than its published specs with 0.0015% of THD+N (0.002% published). The resolution is indeed 16bits.
That said, note these spikes of odd harmonics (H3, H5, H7, ...). That reminded me of some issue we already discussed here when trying the burn a test CD with Windows Media Player. The distorsion you see here does no look like coming from the conversion itself... So, let's investigate.
I usually add the same measurement at -6dBFS:
Ah, again the same distorsion profile (odd harmonics slowly decreasing) at a lower level too, that is very uncommon. With this, we can suspect a digital change of the signal prior to conversion, some sort of level equalization as we saw with WMP. Using shaped dither, we can lower the noise floor to analyze deeper.
Below is the same 999.91Hz @0dBFS but with shaped dither (hence the double wave of low level noise you see starting at 6kHz):
Well, here we reveal an effect of processing the digital signal before conversion. Too bad because I think otherwise the DAC in use here is very good (unless this is a function of it, but I doubt).
Let's park this issue for the moment and move on to the good news. First, this is an very silent player:
No visible components from mains, this is perfect and too rare. You might argue with the previous view that was showing a 50Hz spike, but that was my error, the PC in use for measurements was not on batteries and I got lazy to run the measurement again, sorry.
Let's continue with the good news as the bandwidth is very flat:
You can see the channel mismatch (0.17dB) which is not a concern. The roll off at the end is an effect of the oversampling filter in use.
Talking oversampling filter, this is how it works:
The filter attenuates by -100dB maximum, which is decent but we see noise shaping from the conversion. So a lot of noise is created outside audio band, which is common with DSD (SACD) for instance. There might be an upsampling to DSD happening before conversion.
Multitone (1/10 decade) shows a happy CD player, not having issue to clear 16bits of data:
The Jitter test is relatively good:
We can see some unwanted artifacts but they will remain hidden into music.
Started with the Teac VRDS-20 review, and on your request + support to get it done (more here), I'm adding an "intersample-overs" test which intends to identify the behavior of the digital filtering and DAC when it come to process near clipping signals. Because of the oversampling, there might be interpolated data that go above 0dBFS and would saturate (clip) the DAC and therefore the output. And this effect shows through distorsion (THD+N measurement up to 96kHz):
| Intersample-overs tests Bandwidth of the THD+N measurements is 20Hz - 96kHz | 5512.5 Hz sine, Peak = +0.69dBFS | 7350 Hz sine, Peak = +1.25dBFS | 11025 Hz sine, Peak = +3.0dBFS |
| Teac VRDS-20 | -30.7dB | -26.6dB | -17.6dB |
| Yamaha CD-1 | -84.6dB | -84.9dB | -78.1dB |
| Denon DCD-900NE | -34.2dB | -27.1dB | -19.1dB |
| Denon DCD-SA1 | -33.6dB | -27.6dB | -18.3dB |
| Onkyo C-733 | -88.3dB | -40.4dB | -21.2dB |
| Denon DCD-3560 | -30.2dB | -24.7dB | -17.4dB |
| Marantz CD-73 | -50.3dB | -52.0dB | -27.7dB |
| Fiio DM13 (-3dB attenuation output) | -57.5dB (-91.6dB) | -58.2dB (-78.1dB) | -58.4dB (-81dB) |
I kept some references and will keep the same for other reviews, so you can quickly compare. The results of the Fiio DM13 mean the oversampling filter has 3dB+ headroom when outputting 4Vrms. This is very good but means there is digital attenuation. That said, even a full output (5Vrms), there is still 2dB headroom. So I would again suspect the presence of an SRC used to perform digital attenuation, or something equivalent. For once I've put in parenthesis the THD (without noise) because the noise shaping is added into the calculation and would look like beginning of clipping but it's not.
Stereophile was often using the 3DC measurement as a prof of low noise DAC. It is from an undithered 997Hz sine at -90.31dBFS. With 16bits, the signal should appear (on a scope) as the 3DC levels of the smallest symmetrical sign magnitude digital signal. Here is the result with the DM13:
Houston, we have a problem. What we see here is asymmetry of the output that I've never seen before. This view means the output is far from the requested -90.31dBFS, and actually I measured it at -95.37dBFS. So that is a lack of linearity by 5dB, and that is huge. It does not get much better if I set the output at 5Vrms because the deviation is still a massive 4dB!
Fact is the DM13 looses linearity as early as -70dBFS (by 0.7dB), while we are used to see linearity nailed down to -120dBFS these days! This is equivalent to what I saw with the Marantz CD-73, which had a 14bits DAC. So this is really bad results for the DM13.
So we have the culprit here, I believe. There is processing of the signal which deteriorates it at low levels and that is too bad.
Other measurements (not shown):
- IMD AES-17 DFD "Analog" (18kHz & 20kHz 1:1) : -87.7dB
- IMD AES-17 DFD "Digital" (17'987Hz & 19'997Hz 1:1) : -90.8dB
- IMD AES-17 MD (41Hz & 7993Hz 4:1): -118.1dB
- IMD CCIF (19kHz & 20kHz 1:1) : -90.6dB
- Dynamic Range : 98dB
- Crosstalk: -126dB (100Hz), -109dB (1kHz), -89dB (10kHz)
- Pitch Error : 19'995.88Hz (19'997Hz requested) ie +0.0003% (56ppm)
Note that the player does not apply de-emphasis filter if needed,
Fiio DM23 - Testing the drive
What would be good measurements if the drive would not properly read a slightly scratched CD, or one that was created at the limits of the norm? The below tests reply to these questions.
| Test type | Technical test | Results |
| Variation of linear cutting velocity | From 1.20m/s to 1.40m/s | Pass |
| Variation of track pitch | From 1.5µm to 1.7µm | Pass |
| Combined variations of track pitch and velocity | From 1.20m/s & 1.5µm to 1.40m/s & 1.7µm | Pass |
| HF detection (asymmetry pitch/flat ratio) | Variation from 2% to 18% | Pass |
| Dropouts resistance | From 0.05mm (0.038ms) to 4mm (3.080ms) | Up to 3mm. |
| Combined dropouts and smallest pitch | From 1.5µm & 1mm to 1.5µm & 2.4mm | Pass |
| Successive dropouts | From 2x0.1mm to 2x3mm | Pass |
The drive was able to consistently continue playing, and without generating typical digital clicks, with dropouts of up to 3mm. The interpolation effect remained hidden to my hears when it kicked-off and the DM13 did not fail to maintain a constant flow with up to 4mm dropouts. It is the maximum I can measure, so that is near perfection. The Fiio had no issue with variable linear velocity and/or track pitch, as well as with HF detection. This is a crazy good drive!
Fiio DM13 - Measurements (Optical Out)
I've seen several of you reviewing CD players using their digital outputs, in case the results could be improved from an external DAC.
The lack of precision of the clock means it was nearly impossible to get a clean FFT as windowing errors kept on repeating because there's no PLL with the digital receiver when I run this measurement. Reducing the FFT length and without averaging, I got the below (999.91Hz @0dBFS):
We find again the distorsion we already saw, with a set of odd harmonics that are not on the test CD. This means there is digital modification also with the digital output. This is not "bit perfect" by far.
By the way, our 3DC levels are also impacted, this is again asymmetry and therefore lack of linearity in the digital domain:
I never saw that before. This drives me to believe there is processing before the signal reaches the DAC and also before it reaches the output. What for? I don't know, but this is too bad, even if I could not hear any negative impact. I guess there is some resampling via an ASRC or something because the player has also a Bluetooth option (well, not mine) and maybe that's to simplify the processing downstream. Again I don't know, but it's not good.
EDIT: I did not think before about running the below test, due to the imprecision of the clock. But this is a bandwidth analysis from periodic white noise done with 600+ averages and smoothed. It provides an interesting result (Linear Frequency Scale from 20Hz to 20kHz):
I compared with a CD Player, the Denon DCD-3560, both from their digital outputs (optical). The Denon (purple trace) shows what we expect, a straight line in digital domain. But the Fiio (blue trace) exhibits the typical ringing of a low resolution digital filter. This is what we had more than 40 years ago, with the first 2x oversampling digital filters. This means there's an interpolator, most probably a sample rate convertor, performing a resampling (even if it is resampling from 44.1kHz to 44.1khz) in band limited, of the digital signal, and prior to send it to the digital output. This is not desired.
Conclusion
The Fiio DM13 is extremely good at reading a scratched CD and following a track without failing. It's the best I measured in that area. It's the only player able to read correctly with 3mm dropouts on the disc and only few click with 4mm, amazing.
It is therefore too bad to see such digital processing and degradation of the signal. I think this is a software error and I hope a firmware can fix that. I will update if that's the case.
In the meantime, I can't say the flaws shown can be heard. They remained hidden to my hears. But this little player missed a very high mark because of some undesired processing. Too bad.
I left a message on Fiio’s forum to request if it’s ok to use their recent firmware update which theoretically applies only to the Bluetooth version. If they confirm, I’ll update and cross my fingers that it solves this processing issue.
Cheers
PS: Message to Fiio engineering: Please fix that software error with digital attenuation calculation, and you have the best CD Player/Transport of all times (besides clock, but that will be for the next iteration/super high-end version of this player). Pretty please?
Last edited:
:
