• 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 and Measurements of Holo Audio Cyan DSD DAC

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,381
Location
Seattle Area
This is a review and detailed measurements of Kitsune Hifi's Holo Audio Cyan DAC with DSD option. It is on kind loan from a member. The Cyan comes either in PCM R2R with headphone amp or just DSD (you can also get both). The cost for the DSD version is USD $1099. I am assuming you have to pay shipping on top of that.

While the pedigree of off-the-shelf DIY case cannot be erased, the Cyan DAC adds enough touches to make it look nicer than the crowd:

HOLO Audio DSD Cyan Review.jpg

I am especially fond of the dot matrix white LEDs. Not a good choice in the bedroom likely but otherwise, it is bright and legible.

As noted, the headphone amp option is deleted from the DSD version but oddly the volume control and jacks are left there. I guess it saved them manufacturing cost to make another front panel without those holes. From what I recall, even the headphone circuits are in there but just disabled. The reason being that the DSD stream doesn't allow digital volume control change so the designer opted to eliminate the headphone out option just the same. You have to be pretty convinced of value of DSD DAC to want to lose this option.

Note that PCM playback is still available. The Cyan DAC internally converts PCM to DSD prior to playback.

I am pleased to see balanced output which in my view is mandatory in any DAC above $300:

HOlo Audio Cyan DSD Back Panel Review.jpg

In addition to usual inputs we also have an I^S input using HDMI connector.

For my testing I focused on USB input alone.

And oh, I forgot to say the Cyan is one heavy beast. The corners were damaged as such by the time I got it, not having been double boxed. You can see the nick in the top right. Also note that nice looking copper color sides. Wish they were visible from the front though.

Measurements
I started with my usual dashboard view, first with RCA output and PCM signal at 44.1 kHz:
HOlo Audio Cyan RCA Output Measurements.png


The Cyan DAC completely meets its THD+N spec. Performance improves a bit using XLR outputs:

HOLO Audio Cyan XLR Balanced Output Measurements.png


Tha mains leakage is eliminated as one would expect from balanced output.

Note that in both cases, the output is well below what we like to see. RCA out should be 2 volts. Balanced should be 4 volts. By keeping the output voltage lower, the Cyan DAC avoids driving its output stage hard and perhaps gets better distortion ratings than it would otherwise. Going with the XLR output, the Cyan DSD DAC lands in tier 2 of our DAC ratings which is actually quite good for "boutique"/custom DAC:

HOLO Audio Cyan XLR Balanced Output SINAD Measurements.png


Now, this being a DSD DAC, one may complain that rating it as a PCM DAC is unfair. Putting aside the fact that most of your content is PCM, let's address that by feeding it a 1 kHz tone at DSD64:

HOLO Audio Cyan XLR Balanced Output DSD64 Tone Measurements.png


The noise floor goes down as does the magnitude of the the second harmonic error. Note that the output level remains too low so that is not related to PCM to DSD conversion.

The dynamic range unfortunately doesn't come close to matching its spec:

HOlo Audio Cyan USB Input Dynamic Range Measurements.png


Oops. One of those should say XLR but seeing how the two are similar, I am not going to remeasure to confirm. :) I am unclearly how they got to 121 dB number. Perhaps they used an On/Off scheme as opposed to AES-17 recommendation where a low amplitude tone is used to make sure the unit doesn't mute. There are a lot of vagaries in how dynamic range/signal to noise ratio is measured so I won't fault them as much as I should.

Frequency response shows roll off below audible band:

HOlo Audio Cyan USB Input Frequence Response Measurements.png


We are down about 1 dB at 20 kHz. The younger members may worry about that but the rest of us can't hear it and will live happily ever after!

White noise shows slow filter which seems to be in vogue these days:

HOLO Audio Cyan RCA Output Filter Response Measurements.png


Square wave response shows asymmetrical ringing:

HOLO Audio Cyan XLR Balanced Output Square Wave Measurements.png


I thought it was clipping but it doesn't seem so.

Handcrafting custom DACs with precision is hard. Our intermodulation distortion versus level shows this clearly:

HOLO Audio Cyan USB Input Intermodulation Distortion Measurements.png


The chip-based AKM DAC in the Topping DX3 Pro has far, far better performance. The sloping down section of the curve is normally noise dominated. For distortion to make it peak it means it is fairly high level. That is what we see in at least three segments of the curve. The final value is quite low and hence the reason the Cyan DAC looked good in the Dashboard. Fair bit of fidelity is lost as you lower the digital samples.

We also have significantly higher noise floor although some of that may be due to PCM to DSD conversion.

Another place where a spotlight is put on custom DACs is linearity test:

HOLO Audio Cyan USB Input Linearity Error Measurements.png


We see very odd rise in level error below -100 dB in balanced and then in unbalanced.

Multitone results are excellent though:
HOLO Audio Cyan USB Input Multitone Measurements.png


Usually I see a rise in noise floor at low frequencies but here, it is quite low and remains so.

High bandwidth spectrum of 1 kHz tone is also unusually clean, showing very little noise shaping:
HOLO Audio Cyan XLR Balanced Output 1 kHz Spectrum Measurements.png


Conclusions
Whenever I test these custom DACs, I get prepared for poor showing. Here though, the Holo Audio Cyan DSD seems to avoid outright failures like we see in many R2R DACs. Measurements are quite respectable in many regards and the faults likely not audible.

That said, I still don't see the appeal of spending $1,100 and getting worse performance than a $99 DAC. There is nothing I see here that is euphonic in nature. It is a DAC with less perfection than chip-based DACs.

If you are itching to get a custom DAC, then the HOLO Audio Cyan DSD DAC is a much better choice than many others.

------------
As always, questions, comments, recommendations, etc. are welcome.

Money was donated for gas to go and get some good fish chips in town. Need a few more coins to pay for the meal itself though. 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).
 

Attachments

  • HOlo Audio Cyan USB Input Dynamic Range Measurements.png
    HOlo Audio Cyan USB Input Dynamic Range Measurements.png
    16.3 KB · Views: 482

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
3,425
Likes
3,980
Location
SoCal
So it is using an R2R ladder to convert DSD to analog? How does it work? Perhaps it is converting DSD to PCM before sending it to the R2R?
 

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,626
Likes
10,202
Location
North-East
This is a review and detailed measurements of Kitsune Hifi's Holo Audio Cyan DAC with DSD option. It is on kind loan from a member. The Cyan comes either in PCM R2R with headphone amp or just DSD (you can also get both). The cost for the DSD version is USD $1099. I am assuming you have to pay shipping on top of that.

While the pedigree of off-the-shelf DIY case cannot be erased, the Cyan DAC adds enough touches to make it look nicer than the crowd:


I am especially fond of the dot matrix white LEDs. Not a good choice in the bedroom likely but otherwise, it is bright and legible.

As noted, the headphone amp option is deleted from the DSD version but oddly the volume control and jacks are left there. I guess it saved them manufacturing cost to make another front panel without those holes. From what I recall, even the headphone circuits are in there but just disabled. The reason being that the DSD stream doesn't allow digital volume control change so the designer opted to eliminate the headphone out option just the same. You have to be pretty convinced of value of DSD DAC to want to lose this option.

Note that PCM playback is still available. The Cyan DAC internally converts PCM to DSD prior to playback.

I am pleased to see balanced output which in my view is mandatory in any DAC above $300:


In addition to usual inputs we also have an I^S input using HDMI connector.

For my testing I focused on USB input alone.

And oh, I forgot to say the Cyan is one heavy beast. The corners were damaged as such by the time I got it, not having been double boxed. You can see the nick in the top right. Also note that nice looking copper color sides. Wish they were visible from the front though.

Measurements
I started with my usual dashboard view, first with RCA output and PCM signal at 44.1 kHz:
View attachment 23209

The Cyan DAC completely meets its THD+N spec. Performance improves a bit using XLR outputs:

View attachment 23210

Tha mains leakage is eliminated as one would expect from balanced output.

Note that in both cases, the output is well below what we like to see. RCA out should be 2 volts. Balanced should be 4 volts. By keeping the output voltage lower, the Cyan DAC avoids driving its output stage hard and perhaps gets better distortion ratings than it would otherwise. Going with the XLR output, the Cyan DSD DAC lands in tier 2 of our DAC ratings which is actually quite good for "boutique"/custom DAC:

View attachment 23211

Now, this being a DSD DAC, one may complain that rating it as a PCM DAC is unfair. Putting aside the fact that most of your content is PCM, let's address that by feeding it a 1 kHz tone at DSD64:

View attachment 23212

The noise floor goes down as does the magnitude of the the second harmonic error. Note that the output level remains too low so that is not related to PCM to DSD conversion.

The dynamic range unfortunately doesn't come close to matching its spec:

View attachment 23214

Oops. One of those should say XLR but seeing how the two are similar, I am not going to remeasure to confirm. :) I am unclearly how they got to 121 dB number. Perhaps they used an On/Off scheme as opposed to AES-17 recommendation where a low amplitude tone is used to make sure the unit doesn't mute. There are a lot of vagaries in how dynamic range/signal to noise ratio is measured so I won't fault them as much as I should.

Frequency response shows roll off below audible band:

View attachment 23215

We are down about 1 dB at 20 kHz. The younger members may worry about that but the rest of us can't hear it and will live happily ever after!

White noise shows slow filter which seems to be in vogue these days:

View attachment 23216

Square wave response shows asymmetrical ringing:

View attachment 23221

I thought it was clipping but it doesn't seem so.

Handcrafting custom DACs with precision is hard. Our intermodulation distortion versus level shows this clearly:

View attachment 23217

The chip-based AKM DAC in the Topping DX3 Pro has far, far better performance. The sloping down section of the curve is normally noise dominated. For distortion to make it peak it means it is fairly high level. That is what we see in at least three segments of the curve. The final value is quite low and hence the reason the Cyan DAC looked good in the Dashboard. Fair bit of fidelity is lost as you lower the digital samples.

We also have significantly higher noise floor although some of that may be due to PCM to DSD conversion.

Another place where a spotlight is put on custom DACs is linearity test:

View attachment 23218

We see very odd rise in level error below -100 dB in balanced and then in unbalanced.

Multitone results are excellent though:
View attachment 23220

Usually I see a rise in noise floor at low frequencies but here, it is quite low and remains so.

High bandwidth spectrum of 1 kHz tone is also unusually clean, showing very little noise shaping:
View attachment 23222

Conclusions
Whenever I test these custom DACs, I get prepared for poor showing. Here though, the Holo Audio Cyan DSD seems to avoid outright failures like we see in many R2R DACs. Measurements are quite respectable in many regards and the faults likely not audible.

That said, I still don't see the appeal of spending $1,100 and getting worse performance than a $99 DAC. There is nothing I see here that is euphonic in nature. It is a DAC with less perfection than chip-based DACs.

If you are itching to get a custom DAC, then the HOLO Audio Cyan DSD DAC is a much better choice than many others.

------------
As always, questions, comments, recommendations, etc. are welcome.

Money was donated for gas to go and get some good fish chips in town. Need a few more coins to pay for the meal itself though. 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).

Thank you, Amir! Any way to test it with DSD256 or DSD512 input? The older sibling Holo Spring DAC seems to like higher DSD rates, producing less distortion and lower jitter when fed these. I suspect Cyan may too.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,381
Location
Seattle Area
Thank you, Amir! Any way to test it with DSD256 or DSD512 input?
I only have DSD test tones in DSD64. Let me see if I can get some higher rate ones....
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,381
Location
Seattle Area
So it is using an R2R ladder to convert DSD to analog? How does it work? Perhaps it is converting DSD to PCM before sending it to the R2R?
It is not R2R. I think it just upsamples to high enough rate to then quantize to one bit. I did not see any description of how it works. Maybe someone else knows.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,381
Location
Seattle Area
Here is DSD 128:

1551999313383.png


Not sure why the one channel got worse. Unfortunately the owner wants the unit back so I have no time to troubleshoot....
 

Schackmannen

Active Member
Forum Donor
Joined
Aug 6, 2018
Messages
167
Likes
225
Location
Tucson, Arizona
Can someone explain how the dynamic range is lower than the THD+N? I thought by definition DNR > SINAD or is there a difference between dynamic range and signal to noise ratio?
 

Ilkless

Major Contributor
Forum Donor
Joined
Jan 26, 2019
Messages
1,757
Likes
3,437
Location
Singapore
DACs like these are purchased for the narrative and provenance - non-acoustic reasons without any empirically-verifiable relevance to sound. Except for people who conflate the non-acoustic with the acoustic using folk theories and intuition.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,381
Location
Seattle Area
Can someone explain how the dynamic range is lower than the THD+N? I thought by definition DNR > SINAD or is there a difference between dynamic range and signal to noise ratio?
The bandwidth of dynamic range test is higher so it includes more noise.
 

Schackmannen

Active Member
Forum Donor
Joined
Aug 6, 2018
Messages
167
Likes
225
Location
Tucson, Arizona
The bandwidth of dynamic range test is higher so it includes more noise.
That explains it :) . So DNR is the same as SNR as long as it's measured with the same bandwidth? And what's the bandwidth used for the dynamic range test?
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,381
Location
Seattle Area
That explains it :) . So DNR is the same as SNR as long as it's measured with the same bandwidth? And what's the bandwidth used for the dynamic range test?
From memory, it is 90 kHz. DNR is used here because it measures the dynamic range relative an already active signal. SNR can be on/off.
 

mikemaverick

Member
Joined
Feb 24, 2019
Messages
10
Likes
2
According to the designer he used AES-17 filter therefore out of band noise is excluded from his result
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
I think Holo Audio at least are very good at what they make. Compared to 'other' R2R/multibit seller.

Whether they are worth such a premium is another question, but seeing the SINAD is quite close to the once much favoured NAD M51 is something that is pretty respectable for such a boutique product! :)
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
PS @amirm. The reason for the lower output is the DSD mode of the chip. In PCM it should be nominal 2V/4V unbalanced balanced.
In DSD, the output is 6dB lower supposedly due to SACD/DSD limitation of sorts. From @Miska:

DSD 0 dB is specified as 50% modulation index that translates to -6 dB PCM level in direct conversion. And short term peaks are specified to be max +3 dB which is 75% modulation index and translates to -3 dB PCM level in direct conversion.
 

BYRTT

Addicted to Fun and Learning
Forum Donor
Joined
Nov 2, 2018
Messages
956
Likes
2,452
Location
Denmark (Jutland)
Here is DSD 128: .....

.....Not sure why the one channel got worse. Unfortunately the owner wants the unit back so I have no time to troubleshoot....
So higher is not necessarily better. Interesting.

Maybe coincidence but whenever one channel is exactly a tiny 1 micro volt lower than the other channel SINAD improves few dB better plus THD+N digits improves too, for unballanced its CH1, for ballanced its CH2, for DSD 64 its CH2 and for DSD128 its CH1. In general it doesn't look like lower voltage is better because then DSD128 should shine over DSD64.

1000.png
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
Maybe coincidence but whenever one channel is exactly a tiny 1 micro volt lower than the other channel SINAD improves few dB better plus THD+N digits improves too, for unballanced its CH1, for ballanced its CH2, for DSD 64 its CH2 and for DSD128 its CH1. In general it doesn't look like lower voltage is better because then DSD128 should shine over DSD64.

View attachment 23232
Then the conclusion is this DAC's distortion profile is highly dependent on level? Since the IMD vs level plot also shows a lot of variations.
 
Top Bottom