Holo audio Cyan DAC+Headamp Impressions thread

[giordy60]

Miska, have the measurements dell'holo spring?

[Rick82]

Hi everyone, i'm new here.

This new Holo dac seems very interesting, someone can give some listening impressions without using HQPlayer but with the "built in" conversion from PCM to DSD? 

Maybe also differences between OS and NOS mode?

Thanks!

[Telstar2]

On ‎17‎/‎12‎/‎2017 at 5:33 AM, Panelhead said:

The 28 bit ladder for PCM should be impressive

It's not real 28 bit, not even 24 with its SNR specs.

[zorntel]

Given Jussi's measurements comparing the Cyan between DSD128, DSD256, and DSD512 input I'm curious if anyone on the list (including Jussi) who actually has a Cyan has upsampled identical files to these rates in HQplayer and done any comparative listening tests.  I've done this with both the T+A and the IFI Black label DACs and DSD512 is astounding compared to the lower rates.  I also know people with the Holo Spring and they also prefer DSD512.  Measurements are fine and welcome, but when you're an audiophile listening is what matters in the end.  Just curious.

 

Robert

[lentille]

Co-ax versus USB versus I2s

 

For those of you also using dedicated streamers, I have done enough A to B listening  between Co-ax and USB now to feel comfortable settling on the coax input for my pcm version Cyan. I use a carbon fibre cable which overcomes some of the potential issues with coax.

 

I have another DAC/ headamp/pre-amp which does true direct DSD and my DSD library is small compared to my PCM library which is why I went with the PCM version of the Cyan. 

 

My streamer does not have I2s so others may want to order comment on how that compares to USB. 

 

Vincent 

 

[barrows]

On 2/9/2018 at 1:40 AM, Miska said:

 

Curiously the USB interface is not fully up through USB power because the DAC is not seen by computer unless powered up...

 

I like the way my DSC1 works with Amanero board, because regardless if the DAC is powered up or not, the USB interface is always fully up and running as long as it is connected to computer. So sequence in which things are started doesn't matter. If DAC is off, one can still play music to it, but of course it just doesn't do anything, being just null sink.

Maybe the interface is isolated and the masterclock is coming form the clean DAC side, sounds like an advantage to me.

[barrows]

On 2/12/2018 at 2:13 PM, Telstar2 said:

It's not real 28 bit, not even 24 with its SNR specs.

Huh, it is a DSD DAC, single bit really.  As far as S/N goes, no DAC I am aware achieves better than 21 actual performance via its analog outputs.  Of course, no one's audio system in their room can even do 20 bits...  Unless you are willing to damage your hearing on peaks.

[semente]

5 hours ago, barrows said:

Huh, it is a DSD DAC, single bit really.  As far as S/N goes, no DAC I am aware achieves better than 21 actual performance via its analog outputs.  Of course, no one's audio system in their room can even do 20 bits...  Unless you are willing to damage your hearing on peaks.

 

Very quiet room noise floor at 20dB, maximum SPL of large speakers at 115dB/1m, that's around 90dB at the listening spot.

But good luck finding a recording with such a wide dynamic range.

[barrows]

3 hours ago, semente said:

 

Very quiet room noise floor at 20dB, maximum SPL of large speakers at 115dB/1m, that's around 90dB at the listening spot.

But good luck finding a recording with such a wide dynamic range.

Good example, and that 90 dB is about 15 bits.  They were pretty much spot on that the 16 bits of RBCD is adequate if we are just talking about playback.  But of course some headroom is nice, and, if you are mixing in digital then you are going to want to capture at 24 bits to have room for level adjustments.

If it is not windy, my room is very quiet at night, you can literally hear a pin drop (rural area). 

[semente]

3 minutes ago, barrows said:

Good example, and that 90 dB is about 15 bits.  They were pretty much spot on that the 16 bits of RBCD is adequate if we are just talking about playback.  But of course some headroom is nice, and, if you are mixing in digital then you are going to want to capture at 24 bits to have room for level adjustments.

If it is not windy, my room is very quiet at night, you can literally hear a pin drop (rural area). 

My current home is also in a pretty quiet area but when I was living in a busy street things were pretty dim.

 

Some recordings of unamplified music also take place at night, noise floor isn't just a problem for listening.

[barrows]

No doubt.  My girlfriend is a musician, and she usually records her stuff in homes, or modest home studios.  As these spaces are not sonically isolated like really good pro studios, she always records starting around 11:30 PM, and on 'til dawn. 

BTW, for people who are worried about what goes on in their system way down in the -90 dB to -120 dB area, you really should get an idea of the level at -90 dB in your system.  Get one of the Stereophile test CDs which has the -90 dB 1kHz test tone, and play back that tone in your system at your normal reference volume level-it can be quite surprising how low in level -90 dB is.

[semente]

4 minutes ago, barrows said:

Get one of the Stereophile test CDs which has the -90 dB 1kHz test tone, and play back that tone in your system at your normal reference volume level-it can be quite surprising how low in level -90 dB is.

 

Here's an online alternatively:

 

http://www.audiocheck.net/audiotests_dynamiccheck.php

[Telstar2]

9 hours ago, barrows said:

Huh, it is a DSD DAC, single bit really.  As far as S/N goes, no DAC I am aware achieves better than 21 actual performance via its analog outputs.  Of course, no one's audio system in their room can even do 20 bits...  Unless you are willing to damage your hearing on peaks.

Uh? The DSD part must be but the PCM isnt a ladder like the bigger brother Spring?

Anyway, there exist DACs with true 24bit performance (MSB, Metrum...something with Sabre depending on implementation) but the rest of the chain isn't. The main culprit being the amplification (assuming one goes DAC direct, otherwise there is ZERO chance to even get 130db SNR.

 

Still, any db of less noise matters. The hear can listen below noisefloor, about 40dB below IIRC an AES study I have read.

[ted_b]

16 minutes ago, Telstar2 said:

Uh? The DSD part must be but the PCM isnt a ladder like the bigger brother Spring?

The Cyan comes in two flavors, PCM R2R dac or DSD R2R dac.  Spring has both, and also uses a "backup" set of ladders to linearly compensate if ladders are off slightly.

[barrows]

12 minutes ago, Telstar2 said:

Anyway, there exist DACs with true 24bit performance

Incorrect.  There are DACs which can achieve 24 bit performance in the digital domain (I have one), but they do not achieve that performance at their analog outputs.  The very best DACs achieve at most 21 bit performance at their analog outputs.  which is still incredibly good, 21 bits is around 130 dB.  

To understand how low in level these signals are, I recommend that you play the aforementioned -90 dB tone in your system, I have a feeling that you will be surprised.

[Telstar2]

8 minutes ago, barrows said:

Incorrect.  There are DACs which can achieve 24 bit performance in the digital domain (I have one), but they do not achieve that performance at their analog outputs.  The very best DACs achieve at most 21 bit performance at their analog outputs.  which is still incredibly good, 21 bits is around 130 dB.  

To understand how low in level these signals are, I recommend that you play the aforementioned -90 dB tone in your system, I have a feeling that you will be surprised.

24 bit is 144db SNR, I know that. And i stand corrected with MSB and Metrum measurements at hand. Johnson's noise comes into play.

 

I'm gonna play that tone with 0 attenuation (with my usual software i go from -18dB to -24dB depending on recording). My current speakers are inefficient (88dB) and the amp low gain (15) so I think I wont hear anything, but i'll do that  

 

[barrows]

You will need to play the tone at your normal reference volume level to get an understanding of how audible it is in a real world  listening situation.

Yep, the analog domain is the limiting factor.  But I do think that having the digital domain performance really high matters, even if the analog stage cannot resolve it.  I recently upgraded from an ESS 9018 based DAC to an ESS 9038 based one, the improvement in resolution is quite obvious, even though, by the numbers alone, it would appear that the 9018 is as good as is necessary...

Still have not heard a Holo DAC yet though.  Kinda waiting to see if Audirvana+ adds a DSD capable volume control.

[Telstar2]

On ‎06‎/‎03‎/‎2018 at 5:09 PM, barrows said:

You will need to play the tone at your normal reference volume level to get an understanding of how audible it is in a real world  listening situation.

I'll do. ASAP.

[WenYee]

 

My DAC + ROCK64 as NAA

[Panelhead]

On 3/6/2018 at 10:09 AM, barrows said:

You will need to play the tone at your normal reference volume level to get an understanding of how audible it is in a real world  listening situation.

Yep, the analog domain is the limiting factor.  But I do think that having the digital domain performance really high matters, even if the analog stage cannot resolve it.  I recently upgraded from an ESS 9018 based DAC to an ESS 9038 based one, the improvement in resolution is quite obvious, even though, by the numbers alone, it would appear that the 9018 is as good as is necessary...

Still have not heard a Holo DAC yet though.  Kinda waiting to see if Audirvana+ adds a DSD capable volume control.

  The ESS9038 dac must have better noise levels. The ESS9018 is already lower noise than any dac can deliver.

  My favorite test of a dac is seeing a 1K test tone played back at -90.3 dB. First with 16 bit file and then with 24 bit. The drop is noise level shows resolution. The very best go from -128 dB to -160 dB. That is over 21 bit resolution. I have seen ESS9018 based dacs deliver these results. 

  I do not have a dac with a 9038 chip. But Suspect it will be an improvement over my LHL Pulse Infinity. It delivers 21 bit resolution and sits unused in the closet. 

  There is much more to it than resolution. 

[barrows]

3 hours ago, Panelhead said:

  The ESS9038 dac must have better noise levels. The ESS9018 is already lower noise than any dac can deliver.

  My favorite test of a dac is seeing a 1K test tone played back at -90.3 dB. First with 16 bit file and then with 24 bit. The drop is noise level shows resolution. The very best go from -128 dB to -160 dB. That is over 21 bit resolution. I have seen ESS9018 based dacs deliver these results. 

  I do not have a dac with a 9038 chip. But Suspect it will be an improvement over my LHL Pulse Infinity. It delivers 21 bit resolution and sits unused in the closet. 

  There is much more to it than resolution. 

Indeed.  That is the beauty of DIY, as I can change many parameters to get the DAC to sound the way I want it.  What I cannot change is the ultimate resolution of the DAC chip though, so starting with the highest resolution there is a good place to start. I know some people deride ESS Sabre DACs as being "clinical", but this is not really the case at all.  Implementation can make a huge difference in DAC sound quality, and the ESS offers so much resolution that it will reveal everything about the implementation: do good implementation and you can have spectacular sound.  

For a commercial product, the Ayre QX-5 DAC (ESS 9038), for example, is not "clinical" sounding.

I also optimize the build for DSD playback, and then oversample all PCM to DSD 128, i prefer the sound of DSD with this DAC.

[ted_b]

Barrows, are you referring to the sound of the Cyan?  I am confused as to why these latest posts are about SABRE chips and DIY.  This is a Holo Cyan (R2R dac) impressions thread. 

[barrows]

Hi Ted, sorry, did not mean to derail...  Just a quick direct response to George.

 

BTW, I would love to hear more impressions from Cyan users.  But I suspect they might be too busy listening to music,  

 

Here is a question on the Cyan DSD: my understanding is that the Spring has the double ladder configuration for the PCM converter, which increases resolution, but the Cyan PCM does not.  I would suspect that the DSD converter (a version of Jussi's DSC-1 approach) would not have a doubled resistor array because doing so would not offer the same advantage it does for a R2R PCM converter.  So is the converter architecture for DSD in the Spring and Cyan DSD the same? 

[Bimmer100]

I am testing a new firmware for the Cyan at this time, and will update info in a bit. And we are accepting orders for the PCM Cyan Again! BTW.

 

On an off note, i've not tested it before but recently have been using the susvara's with the Cyan headamp and it's excellent.  I was not sure how demanding these headphones would be, but they are around 85/100 steps on the cyan for loud volumes, and 90/100 for extremely loud.  My iFi Pro iCan drives them around same volume levels at about 12 o clock and 1 o clock on 18+gain 

 

Just an interesting note for those who wonder how capable the Cyan Headamp (PCM) is capable of.  I'm using them on HP LO btw.

 

on a side note, we are seriously considering building a Headamp that is quite powerful for a newer product in the future. able to drive anything with ease and do it extremely well. 

 

 

[GUTB]

I’ll note that the SRC in the Cyan DSD is very good. There isn’t a large difference between playing PCM or using HQPlayer to upsample to DSD 512.