Article: Holo Audio Spring Level 3 Kitsune Tuned Edition R2R DAC Review

[Superdad]

8 hours ago, Forehaven said:

 

 

Those are some serious compact woofers!  Congrats.

[Forehaven]

LOL!  Thanks!

[anroj]

On ‎3‎/‎30‎/‎2017 at 6:40 AM, ted_b said:

(I will add this to the review body when Chris figures out why the author, me, can no longer edit  , a  growing pain of the otherwise wonderful site upgrade )

 

So, it is exciting to give you feedback on this new beta firmware for the Singxer SU-1.  A few caveats:

1)  This firmware is brought to you by Kitsune Hifi, and is going to be exclusively for Kitsune SU-1 customers.

2)  This firmware, named 2.20, is ONLY targeted to be used with the Holo dac

3)  This firmware is still not ready for prime time, as it currently needs work to handshake properly with Roon and others.  Please don't bug Tim about it or it will get more behind, and reviewers like me will get fewer of these exclusive testing opportunities, something Tim is hoping will further the production release, not delay it.

4)  Tim or I will announce as soon as this thing is ready.

 

All that being said, I am now getting native DSD512, from HQPlayer, via I2S and the SU-1/Holo combo!  In my case I am first using the Windows Holo ASIO driver on my Windows NAA.  I have not tested it with the microRendu yet, as that unit is not yet certified with the SU-1 and currently requires DoP when using the microRendu in NAA mode.

 

I will do my reviewer work tonight and over the weekend; i.e compare sq of DSD512 playback via I2S (NAA to SU-1 to Holo) vs USB (NAA to Holo).

Thanks, Ted. I am also considering the Spring DAC Level III, so the Singxer (Kitsune version) would solve my issue.

[Superdad]

On 3/30/2017 at 6:40 AM, ted_b said:

1)  This firmware is brought to you by Kitsune Hifi, and is going to be exclusively for Kitsune SU-1 customers.

2)  This firmware, named 2.20, is ONLY targeted to be used with the Holo dac

 

Hi Ted:

 

Maybe you can ask Tim to name the new firmware something other than 2.20.  Here is why I ask that:

1) The current firmware choices are 2.0 and 2.02, with the only difference between them being the left-right channel assignment for DSD output.

2) Given the above, calling it 2.20 is confusing, especially since

3) The SU-1 I just received came with 2.02 FW loaded--and it is the wrong one for channel assignment when used with the Holo Spring!

 

Tomorrow I will be flashing my SU-1 (got to take it to my son's Windows computer as the XMOS DFU system does not support macOS) to 2.0.  

 

Now I assume that the new "2.2" being worked on will have the channel assignment of 2.0 since you say it is intended only for Spring DAC users.  But I think you can start to see why I thing the choice of naming it 2.2 is unfortunate. 

 

Best,

--Alex

[ted_b]

Yes, 2.2 has correct DSD channels.  I can't imagine they will change the version number, but who knows.

[ted_b]

I need to update my impressions of the Su-1 since spending more time with the 2.2 firmware (DSD512) and with doing some tweaking of my Windows HQplayer server (which I'll describe in the "massive improvement" thread, thanks mostly to lmitche's ideas of local hdd music, better SATA cables and better connection to the cpu via a cheap M.2 SATA adapter). 

 

I earlier said the SU-1 (Windows NAA ASIO -> JCAT USB card -> TotalDAC1 USB cable -> SU-1 -> Nordost SilverScreen HDMI cable -> Holo Kistune DAC I2S input) was slightly better than going direct JCAT card to Holo USB input.  Well, I am going to change that to significant qualitative change, and not only for DSD512 (for everything DSD) but also comparing my HQPlayer PCM upsampling (44k base to 352, 48k base to 384).  Way more detail, cleaner edges, better image density, less splashy highs.  I may assume it is simply the better clocks in the SU-1, but something is happening where you won't be able to convince me to sell the SU-1...it's staying put and is my go-to for the Holo. 

[tboooe]

Thank you @ted_b.  So is the only difference the firmware? I am curious about your comments on the better clocks in the SU-1.  You were already using the HDMI output from the SU-1 previously when you stated that the I2S output was only a little better than the USB correct?  If so unless the firmware somehow changed the performance of the SU-1 clocks it seems to that is constant between your previous experience and today.  The improvement must be down to the firmware only since you report big improvements even when playing PCM which were doing before?? 

[ted_b]

2 minutes ago, tboooe said:

Thank you @ted_b.  So is the only difference the firmware? I am curious about your comments on the better clocks in the SU-1.  You were already using the HDMI output from the SU-1 previously when you stated that the I2S output was only a little better than the USB correct?  If so unless the firmware somehow changed the performance of the SU-1 clocks it seems to that is constant between your previous experience and today.  The improvement must be down to the firmware only since you report big improvements even when playing PCM which were doing before?? 

Tboooe, hi.  No, I also think my server improvements made the SU-1 superiority more obvious.  In any case, it is now my go-to.  I may try things like the Intona on it soon (from NAA to SU-1).

[ted_b]

For those of you who are thinking about using an LPS-1 (or other ps) as the power supply for the SU-1, Tim (KitsuneHiFi) has decided to sell a conversion kit that includes the "molex"-to-female dc cable and the replacement back panel board for the dc input (replaces the IEC portion), all for like $30.

 

https://kitsunehifi.com/product/singxer-su-1-dc-power-conversion-kit-kitsunehifi-2-1mmx5-5mm/

[pgbcincy]

On 4/19/2017 at 8:29 AM, ted_b said:

https://kitsunehifi.com/product/singxer-su-1-dc-power-conversion-kit-kitsunehifi-2-1mmx5-5mm/

 

Thanks Ted, much better looking than my white PVC plate!   Any updates on SU-1 Firmware 2.2 testing?

TIA,

Paul 

[mikicasellas]

On 7/10/2016 at 8:29 AM, ted_b said:

 

Unfortunately, also, the fact that I haven't heard either dac (and mentioned T+A just above). Chris has been getting the dacs lately, not me for some reason. If someone wants to send me any of the dacs discussed here, I will listen for a week or two and return, all at my cost.

 

I also consider comparing two or more musical dacs (easy to compare to poor ones) somewhat akin to comparing favorite foods or the beauty of Margot Robbie vs Kate Upton. It's possible, but more likely that personal preferences are the biggest factor.

 

If your library of great favorite music is mostly or all PCM, then I would investigate the Holo Audio, the Chord Hugo/2Qute or Mojo (yes, Mojo), and the Yggdrasil (based on other comments), or if you have the coin, the MSB Analog. If you prefer to upsample to DSD bit rates (or collect lots of music that is recorded or transferred in DSD), the exaSounds and the more expensive Fore Audio (I mentioned above) should be in your eval list. Clearly the T+A too. But to say one is empirically better than the other? Depends on need, budget, sonics (none sound absolutely identical, some or slightly warmer, some more dynamic, etc). And most importantly, some just jive better with your equipment. The Holo Audio ability to do DSD512 with the microRendu is a real value to me.

 

As Is aid in the review, this Holo Audio Spirng KitsuneHiFi Edition dac is the first one in my room that does PCM and DSD equally well and to my liking enough to buy it. Some of the dacs above do one or the other as well (Chords are ever so slightly warmer colors for PCM; exaSound Windows USB driver produces zero pops and clicks in DSD transitions).

 

If I could have one wish for the HSKE it would be to have an external power supply so I could roll my own with some of the fabulous ones out there. However, I found nothing lacking in the ps, per se (but how do I know). Just a wish, that's all.

 

 

It would be very interesting to know how HOLO SPRING L3 in NOS mode sounds compared to T+A DAC8 DSD upsampled at DSD512 with the same audio file (PCM or DSD), which is better in that specifically comparison?

 

That for me would be the key decision...

 

 

[ted_b]

Agreed...as I said, if someone wants to send me a T+A then I'm open to paying for shipping.

[ted_b]

Edit: The fight would be over DSD512 (and quite possibly the T+A would have advantages, who knows) but I have yet to hear anyone speak highly of the T+A's PCM capability, so they upsample everything to DSD512 (and why not!).  Me and my Holo, I stay in PCM for the most part as I find 352 for 44k-based stuff and 384k for 48k stuff (PCM-to-PCM) sounding slightly more robust than PCM-to-DSD, or at least a tie.  It's a real benefit that PCM-to-PCM is much easier cpu-wise, so I simply upsample all DSD to DSD512 and that way my i7-6700k plus GPU can do all filters/all modulators to DSD512 without having to worry about taxing my server to PCM-to-DSD512.  Not needed (but possible..simply replace less tolerant filters with their -2s variant).

[mikicasellas]

1 hour ago, ted_b said:

Edit: The fight would be over DSD512 (and quite possibly the T+A would have advantages, who knows) but I have yet to hear anyone speak highly of the T+A's PCM capability, so they upsample everything to DSD512 (and why not!).  Me and my Holo, I stay in PCM for the most part as I find 352 for 44k-based stuff and 384k for 48k stuff (PCM-to-PCM) sounding slightly more robust than PCM-to-DSD, or at least a tie.  It's a real benefit that PCM-to-PCM is much easier cpu-wise, so I simply upsample all DSD to DSD512 and that way my i7-6700k plus GPU can do all filters/all modulators to DSD512 without having to worry about taxing my server to PCM-to-DSD512.  Not needed (but possible..simply replace less tolerant filters with their -2s variant).

 

 

So when you say "I stay in PCM capability" you mean NOS right?

 

And when you upsample PCM 44k to 352 and 48 to 384 & DSD to 512, are you doing that from the DAC or HQP? that part was not clear for me.

 

Mike

[mikicasellas]

I can get the T+A DAC8 DSD for US 3,100 delivered  & the HOLO L3/SU-1 combo US 2,900 Plus Shipping

 

About the same even little more with HOLO/SU-1....

[ted_b]

Mike,

NOS always.  HQplayer does the work.

[Superdad]

19 hours ago, ted_b said:

Me and my Holo, I stay in PCM for the most part as I find 352 for 44k-based stuff and 384k for 48k stuff (PCM-to-PCM) sounding slightly more robust than PCM-to-DSD, or at least a tie.

 

Not to mention that the output level of the Spring when fed PCM seems to be about 10dB hotter than when fed DSD (both from HQP).  At least that is how it is via the I2S input from the Singxer SU-1.

 

 

[ElviaCaprice]

On ‎4‎/‎6‎/‎2017 at 1:47 PM, ted_b said:

I earlier said the SU-1 (Windows NAA ASIO -> JCAT USB card -> TotalDAC1 USB cable -> SU-1 -> Nordost SilverScreen HDMI cable -> Holo Kistune DAC I2S input) was slightly better than going direct JCAT card to Holo USB input.  Well, I am going to change that to significant qualitative change, and not only for DSD512 (for everything DSD) but also comparing my HQPlayer PCM upsampling (44k base to 352, 48k base to 384).  Way more detail, cleaner edges, better image density, less splashy highs.  I may assume it is simply the better clocks in the SU-1, but something is happening where you won't be able to convince me to sell the SU-1...it's staying put and is my go-to for the Holo. 

 

I wouldn't doubt it's better with the extra additional good clocking going on in the SU-1.  But what would happen if you added a SoTM super clocking Ultra component instead (like the TX)?  Now it's a more fair fight between the two and it just may be that the SU-1 has the inferior clocking?

  Let alone where the ISO Regen will fall in this?

[Superdad]

43 minutes ago, ElviaCaprice said:

 

I wouldn't doubt it's better with the extra additional good clocking going on in the SU-1.  But what would happen if you added a SoTM super clocking Ultra component instead (like the TX)?  Now it's a more fair fight between the two and it just may be that the SU-1 has the inferior clocking?

  Let alone where the ISO Regen will fall in this?

 

Careful about mixing things up when talking about clocking.  While USB and processor clocks (13.0MHz, 24.0MHz, 25.0MHz, etc.) do have an effect (was very clear when we compared and chose to use a Crystek 575 in the ISO REGEBN), those are not at all the same as the far more critical audio master clocks (22.5792MHz and 24.576MHz and multiples of) used by I2S and S/PDIF and the DAC.

I've been seeing a lot of clock-obsessed folks here and on other forums conflating the two types, and that is not right.

 

By the way, @scan80269 discovered that the Spring actually uses crystals (X) and not crystal oscillators (XO), and supplies its own oscillator circuits closely surrounding the crystals.  Certainly speaks to Mr. Zhu's engineering prowess to do that.

And it is not clear--without confirmation from the designer--if feeding the Spring I2S/DSD (over LVDS/HDMI) does indeed entirely bypass the DAC's own master clocking.  Might or might not.  Depends upon how he designed it.

 

I am with Ted in preferring the SQ with the Spring fed by the SU-1 (and that combo is VERY revealing of differences in the USB devices feeding it--which only says something either good or bad about the SU-1).  But I admit that I have not spent much time comparing Spring's USB input versus SU-1/I2S since receiving the final production ISO REGEN boards on Wednesday.  So not only can I not directly answer your final question above, I will leave it to others to make that comparison and judgement next month.

 

--Alex C.

 

 

[AmusedToD]

On 4/22/2017 at 2:36 AM, Superdad said:

 

Careful about mixing things up when talking about clocking.  While USB and processor clocks (13.0MHz, 24.0MHz, 25.0MHz, etc.) do have an effect (was very clear when we compared and chose to use a Crystek 575 in the ISO REGEBN), those are not at all the same as the far more critical audio master clocks (22.5792MHz and 24.576MHz and multiples of) used by I2S and S/PDIF and the DAC.

I've been seeing a lot of clock-obsessed folks here and on other forums conflating the two types, and that is not right.

 

By the way, @scan80269 discovered that the Spring actually uses crystals (X) and not crystal oscillators (XO), and supplies its own oscillator circuits closely surrounding the crystals.  Certainly speaks to Mr. Zhu's engineering prowess to do that.

And it is not clear--without confirmation from the designer--if feeding the Spring I2S/DSD (over LVDS/HDMI) does indeed entirely bypass the DAC's own master clocking.  Might or might not.  Depends upon how he designed it.

 

I am with Ted in preferring the SQ with the Spring fed by the SU-1 (and that combo is VERY revealing of differences in the USB devices feeding it--which only says something either good or bad about the SU-1).  But I admit that I have not spent much time comparing Spring's USB input versus SU-1/I2S since receiving the final production ISO REGEN boards on Wednesday.  So not only can I not directly answer your final question above, I will leave it to others to make that comparison and judgement next month.

 

--Alex C.

 

 

 

@Superdad, thank you for the clarifications.

 

I intend to bite the bullet on the Holo Spring DAC and wonder if I should stretch to a SU1 in order to exploit the I2S connection. However, since I am using the Sotm sms200Ultra (with the superior clock), I would like to know whether connecting the sms200Ultra directly to the Holo will sound better than going through the SU1. I understand the Holo features a rather good and galvanically isolated USB input, and perhaps the Sotm directly into Holo via USB cable (or the Uptone adapter) should be the better choice. I am afraid the SU1 might "degrade" the signal quality since it probably uses an inferior clock. Or am I wrong?

 

[tboooe]

8 hours ago, AmusedToD said:

However, since I am using the Sotm sms200Ultra (with the superior clock)

 

8 hours ago, AmusedToD said:

I am afraid the SU1 might "degrade" the signal quality since it probably uses an inferior clock. Or am I wrong?

Question...how do you know the SOTM clocks are the superior to the ones used in the SU-1?  I dont think any information has been provided yet.  If I am wrong, can you pls point to where this information is?

[Superdad]

14 hours ago, AmusedToD said:

 

@Superdad, thank you for the clarifications.

 

I intend to bite the bullet on the Holo Spring DAC and wonder if I should stretch to a SU1 in order to exploit the I2S connection. However, since I am using the Sotm sms200Ultra (with the superior clock), I would like to know whether connecting the sms200Ultra directly to the Holo will sound better than going through the SU1. I understand the Holo features a rather good and galvanically isolated USB input, and perhaps the Sotm directly into Holo via USB cable (or the Uptone adapter) should be the better choice. I am afraid the SU1 might "degrade" the signal quality since it probably uses an inferior clock. Or am I wrong?

 

 

Sorry, but you are doing exactly what I was complaining about: Conflating the AUDIO clocks of the SU-1 with the USB clocks of the streamer source.  Different worlds and different importance.

[AmusedToD]

On 13.6.2017. at 11:07 AM, tboooe said:

 

Question...how do you know the SOTM clocks are the superior to the ones used in the SU-1?  I dont think any information has been provided yet.  If I am wrong, can you pls point to where this information is?

 

Well, much has been said about these clocks on another thread: "A novel way to massively improve...."

 

I don't know if they are superior as I can't measure that. I haven't even seen the SU-1, let alone heard it. But I do have the sms200ultra and I know it sounds fantastic, much better in fact than the plain sms200, and the only difference between the two is in the clock.

[AmusedToD]

On 13.6.2017. at 5:17 PM, Superdad said:

 

Sorry, but you are doing exactly what I was complaining about: Conflating the AUDIO clocks of the SU-1 with the USB clocks of the streamer source.  Different worlds and different importance.

 

Ok, but I am just trying to understand the difference and how it affects sound quality. You said the SU-1 benefits from good USB sources. I just need to know whether feeding the USB input of the Holo Spring with a fine source like the SOTM sMS-200Ultra will sound as good or better than going from the same source to a SU-1 and then via HDMI to the Holo Spring i2S input.

 

And does the clock in the sms200Ultra matter at all if it's connected to a device like the SU-1 which will reclock it. 

[Superdad]

6 hours ago, AmusedToD said:

 

Ok, but I am just trying to understand the difference and how it affects sound quality. You said the SU-1 benefits from good USB sources. I just need to know whether feeding the USB input of the Holo Spring with a fine source like the SOTM sMS-200Ultra will sound as good or better than going from the same source to a SU-1 and then via HDMI to the Holo Spring i2S input.

 

Virtually all USB inputs on DDCs and DACs benefit from being fed with improved signal integrity, impedance match, and for reasons readily heard but still being explored, a source with a good clock.  Such clock is not the audio-related clock, rather just the USB (or other processor data clock).  That's what you get when running from a streamer with a great clock (such as the SMS-200ultra or the forthcoming Sonore ultraRendu), or a USB regenerator with a great clock (such as the SOtM tX-USBultra or the UpTone ISO REGEN).

 

But the Holo Spring DAC benefits from the use of the SU-1 DDC because, when driving the Spring via I2S (over HDMI cable using LVDS--low voltage differential signaling--drivers) the SU-1's superior AUDIO FREQUENCY clocks (Crystek 575s at 22.5792MHz and 24.576MHz) take over as the master clock for the DAC (the Spring does not reclock that).

 

More directly to your question: While your SMS-200ultra is a fine source, it will neither turn the Spring's USB input into a world-class input (it is fine as typical XMOS USB input boards go), nor will you be bypassing the quite average audio clocks of the DAC (Holo admits they put the money elsewhere in the product--it is not a high priced unit and compromises have to be made somewhere; as most here know, I own and enjoy the Spring for what it is).

 

So yes, even with your new streamer, the Holo Spring should sound better driven by the SU-1 through its I2S input.

 

Quote

And does the clock in the sms200Ultra matter at all if it's connected to a device like the SU-1 which will reclock it. 

 

Not really appropriate to say that the SU-1 is "reclocking" what it receives via USB.  An entire USB PHY/processor>I2S>isolator>LVDS>output conversion occurs (with S/PDIF and AES also output options)--and both a cheap data clock for the XMOS processor and the Crystek CCHD-575s for the rate clocking are involved.

 

So while yes, there are skeptics about the relevance of source data clocking, what you are asking about is the same for either direct to DAC or through the DDC.  To more clearly state that: A low jitter I2S/DSD signal into the DAC chips (or discrete R2R ladders in the case of the Holo Springs) is always the end goal.

Upstream clocks seem to put a fingerprint on the result, but the audio-rate clocks are the most crucial to focus on.  And the SU-1's are far better than what is in the Spring.

 

Hope that helps a little.

--Alex C.