HOLO Audio MAY DAC

[Ozan Bolat]

1 hour ago, Miska said:

 

No, I think both are very similar in this respect. Important thing to note about ADI-2 is that it has been around with three different DAC chips! My ADI-2 Pro and ADI-2 Pro AE are the first generation.

 

then you mean Spring 3 is much too expensive and waste of money in comparison, especially since it does not offer ADC and other valuable options?

[Miska]

4 minutes ago, Ozan Bolat said:

then you mean Spring 3 is much too expensive and waste of money in comparison, especially since it does not offer ADC and other valuable options?

 

No? Spring 3 has similar pricing as ADI-2 Pro. Then you need to decide which one you like. Especially since you cannot anymore get new ADI-2's with AKM DAC chips...

 

I have Cyan, Spring, Spring 2 and Spring 3. And two RME ADI-2 Pro's. And I don't feel like wasting any money.

 

One of my ADI-2 Pro's is now serving input side with it's ADC and digital inputs.

 

[Cogito]

9 hours ago, Miska said:

 

I tried this track too, from Qobuz, with following settings:

 

1) PCM: poly-sinc-gauss-long + LNS15 at 20-bit, 1.536M output rate, HQPlayer volume set to -9 dBFS (to compensate Holo's 6 dB gain difference between PCM and DSD)

2) DSD: poly-sinc-gauss-long + ASDM7EC at 12.288M output rate, HQPlayer volume set to -3 dBFS

 

Holo Spring 3 + Ferrum Oor headphone amplifier.

 

I have to say that it is hard to tell these two cases different with this track. The snap is not particularly hard sounding. Biggest difference I notice is that with DSD output there's more "air" in the sound. The space and depth is better with DSD. With PCM it is more closed-in, "on your face", the snap is little bit drier.

 

Try listening to echo of the snap and other sounds, how the echo fades off. And how well the direct sound separates from the echo.

 

Kind of interesting and analog-sounding result is with poly-sinc-xtr-short-mp. Space becomes notably more condensed, also the snap sounds different and more like finger snap.

 

But I'd say (2), to me, subjectively gives best insight to what is on the recording.

 

is this a valid test to determine the differences between PCM and DSD outputs of HQP?

Quboz is streaming this album at 24-bit 96kHz resolution. If Quboz has created 24-bit PCM directly from the tapes, you are loosing 4-bit resolution in HQP PCM output settings and another 1bit resolution in attenuation.

 

To my ears, DSD sounds little softer overall than 24-bit PCM played thru’ Spring.  Elvis’ voice seems to have little more texture in PCM.  

[MarkusBarkus]

13 hours ago, firedog said:

I just listened to a 24/96 version of that track and it sounds like finger snaps to me.

...ditto. 24/96 from Qobuz. Sounds like finger snaps to me... *but* all snaps are not the same.
 

This one sounds like it was made with the index finger curled in.

You can try it at home. Index finger extended, is higher freq than curled, though snap-force has an effect too. Snap on, people...

[Ozan Bolat]

1 hour ago, MarkusBarkus said:

...ditto. 24/96 from Qobuz. Sounds like finger snaps to me... *but* all snaps are not the same.
 

This one sounds like it was made with the index finger curled in.

You can try it at home. Index finger extended, is higher freq than curled, though snap-force has an effect too. Snap on, people...

If you want to train you'll find nice finger snaps on Peggy Lee's version of Fever and those are pretty well documented : https://www.peggyleediscography.com/p/LeeResearchFever.php . Elvis retained this orchestration idea but to my ears he did not have as good finger snappers at hand (pun intended) and resorted to castanets

[Ozan Bolat]

next time someone comes into Elvis he or she should ask him

[Ozan Bolat]

On 9/24/2021 at 1:30 AM, ted_b said:

You tell us. There is no bigger DSD fan than me but on the May I love PCM staying PCM and playing at 32fs. Ymmv

what are your favorite filters ? I reassessed PCM on my Spring 3 after burning it with a feed of SDM exclusively and to my surprise I love PCM too now while out of the box SDM sounded much better. If there's a rational different from a change of my mood, I'd point that SDM is more maths inside the computer, less physicality inside the Holo

[KenMoreira]

2 minutes ago, Ozan Bolat said:

what are your favorite filters ? I reassessed PCM on my Spring 3 after burning it with a feed of SDM exclusively and to my surprise I love PCM too now while out of the box SDM sounded much better. If there's a rational different from a change of my mood, I'd point that SDM is more maths inside the computer, less physicality inside the Holo

What is sdm

[Extreme_Boky]

21 hours ago, Ozan Bolat said:

hMM, think the things you listed for potential burn in benefit appear pretty common to both paths, don't they? However reassessing PCM 1536 vs DSD 256 on the Spring 3, I feel like now writing : it depends ; while I was 100% pro SDM before burn in. Most comfortable claim would be change of mind or special mindset this morning but can we rule out that by nature the R2R path is more susceptible to burn in while that burn in has not been R2R specific? 

 

Both the PCM and the DSD DA conversion can be implemented by using resistor networks. 

 

PCM uses an R-2R resistor network; DSD uses a summing resistor network. 

 

With DSD, the incoming DSD serial stream needs to be converted to parallel stream first, followed by summing resistor network - which has an analog signal at its output.

 

It seems that May does indeed have two separate DAC "streams" that separately handle PCM and DSD. 

 

I had a look inside the power supply unit last week. Time-permitting, I'll be opening the DAC unit next week - what I do not understand at this stage is how Jeff utilised the same analog stage for both DAC streams, i.e. how are the signals from each DAC (PCM and DSD) routed to the same analog output stage... 🤯

[Ozan Bolat]

3 hours ago, KenMoreira said:

What is sdm

Sigma Delta Modulation, call it DSD if you wish

[Miska]

10 hours ago, Cogito said:

is this a valid test to determine the differences between PCM and DSD outputs of HQP?

Quboz is streaming this album at 24-bit 96kHz resolution. If Quboz has created 24-bit PCM directly from the tapes, you are loosing 4-bit resolution in HQP PCM output settings and another 1bit resolution in attenuation.

 

Yes, it is the only correct way.

 

No, I'm not losing any resolution. I'm adding resolution and correcting low-level linearity error of the R2R section.

 

Output produced this way has much more resolution than the 24-bit original. Even in simplistic terms:

96000 * 24 = 2.304 Mbps

1536000 * 20 = 30.72 Mbps

 

 

OTOH, you notice there's clearly audible tape hiss on the old recording. Meaning that it has maybe 8-bit worth of dynamic range...

 

Here is Spring 3 playing 24-bit 1 kHz -120 dBFS test tone:

 

And here the same test tone, but with DAC Bits set to 20 and LNS15 noise-shaper:

 

You can see that all the distortion is gone and no changes in the noise floor. I could even drop DAC Bits to 16 and it would still look the same.

 

If you just send 24-bit data there, you are losing resolution.

 

[Miska]

6 hours ago, Extreme_Boky said:

I had a look inside the power supply unit last week. Time-permitting, I'll be opening the DAC unit next week - what I do not understand at this stage is how Jeff utilised the same analog stage for both DAC streams, i.e. how are the signals from each DAC (PCM and DSD) routed to the same analog output stage... 🤯

 

You can usually set the latches in high impedance mode. So you can have two conversion sections connected in parallel, the other one is just inactive when the other one is operational. Or you could of course use a relay or a semiconductor switch.

 

[mushi]

Hey guys, for the past few months something has started to happen with my holo may: the power section is clearly heard after switching on, as is the noisy transformer through which the current flows. I'm not sure, but it seems to me that the holo was quite quiet at the beginning, it was only in the last couple of months that i started noting this issue.

 

In the evening, after connecting holo to the power, i can hear from the listening position (2.5 m) how the power section works. And only once so far, it happened that for half a minute the holo power section started to rattle really loud and then calmed down, it bothered me...

 

Are your holos silent?

 

[GoldenOne]

16 minutes ago, mushi said:

Hey guys, for the past few months something has started to happen with my holo may: the power section is clearly heard after switching on, as is the noisy transformer through which the current flows. I'm not sure, but it seems to me that the holo was quite quiet at the beginning, it was only in the last couple of months that i started noting this issue.

 

In the evening, after connecting holo to the power, i can hear from the listening position (2.5 m) how the power section works. And only once so far, it happened that for half a minute the holo power section started to rattle really loud and then calmed down, it bothered me...

 

Are your holos silent?

 

This is caused by DC offset on your mains, which saturates the transformer core and causes it to hum.

 

Its annoying, but not harmful. Just means the transformers are slightly less efficient. But the ones in the may are specced for 400% of what it draws anyway. 

 

I get dc offset every day from about 9am to 11am. Not sure why but its there every day. Everything I have with a big transformer just starts humming :P

 

You can buy something like the humX or ifi dc blocker to address the issue

[firedog]

https://avahifi.com/products/humdinger-dc-line-blocker

 

I can testify that this works.

[mushi]

Wow! Thanks guys, i'll definitely try one of those.

 

At this moment i can try ifi dc blocker. Also found that emotiva cmx-2 works for some people, but i need eu plugs.

 

[Zauurx]

You can try this :

 

Or this :

Simply... Putting distance between electrical appliances that can pollute each other.
(electric field decreases with the square of the distance)

[firedog]

2 hours ago, Zauurx said:

You can try this :

 

Or this :

Simply... Putting distance between electrical appliances that can pollute each other.
(electric field decreases with the square of the distance)

If it's DC on the line that won't help. It can even be coming from a source outside his home.

[Zauurx]

Yes, but everyone has an extension cord and it doesn't cost much to test and put some (electrical) distance between source and amp.

I use this DC-Blocker, upstream of my audio sources:
https://www.atlhifi.com/shop/fully-assembled-devices/combined-module-dc-blocker-emi-rfi-common-mode-filter-assembled-in-case/

 

[GoldenOne]

4 hours ago, Zauurx said:

You can try this :

 

Or this :

Simply... Putting distance between electrical appliances that can pollute each other.
(electric field decreases with the square of the distance)

An extension cord/physical distance would only potentially help emi. 

 

But it could make ground/rf noise worse due to... Well... Much more wiring to pick up the noise :P

 

And will not help a dc offset. A dc offset has to be addressed with a dc blocker. (or ignored if its minor enough) 

[fas42]

A simple tweak with using the extension cord method is to tightly twist the cable in the direction that its internal construction already has such - could improve the rejection of pickup, and, stabilise the internal materials a touch more, so that there is less scope for movement of the construction parts.

[toddrhodes]

Color me a little confused here. I've always understood PCM to be 6 db louder than DSD, all else equal. 

 

I use HQ Player and I generally upsample to DSD 256. Using a volume range on HQP of -55 to -4 dB, I find a comfortable listening level at about -28. "Loud" on most tracks is -20. "Wake up my wife" is -15. 

 

I decided to try out NOS PCM mode (and I've also tried upsampled PCM via HQP as well, no change) and "loud" is way up at -8 dB on the HQP volume control. DSD output is substantially louder than PCM and I'm not quite sure why.

 

I can ask this in the HQP software thread if need be, but wondered if anyone else had anything similar to report with May + HQP. 

[GoldenOne]

1 minute ago, toddrhodes said:

Color me a little confused here. I've always understood PCM to be 6 db louder than DSD, all else equal. 

 

I use HQ Player and I generally upsample to DSD 256. Using a volume range on HQP of -55 to -4 dB, I find a comfortable listening level at about -28. "Loud" on most tracks is -20. "Wake up my wife" is -15. 

 

I decided to try out NOS PCM mode (and I've also tried upsampled PCM via HQP as well, no change) and "loud" is way up at -8 dB on the HQP volume control. DSD output is substantially louder than PCM and I'm not quite sure why.

 

I can ask this in the HQP software thread if need be, but wondered if anyone else had anything similar to report with May + HQP. 

Double check your device volume itself isn't <100.

 

For some reason ALSA / Linux sets a lot of devices to 45/100 by default. 

Many dacs the hardware addressable volume won't do anything. But some like the may do have dsp vol control on the USB implementation so it will indeed make it quieter. 

 

In Windows check the device volume is set to 100.

For Linux check device volume too. (I just told roon to use 'device volume' and set it to 100 and that sorted it). 

I had the same thing in the terminator plus but the hardware vol control doesn't do anything on it so it didn't actually make anything quieter

[toddrhodes]

I will give it a shot - the HQP PC that feeds my NAA is a headless setup so I need to get logged into it with a terminal to see if that could be it. Definitely seems reasonable!

[toddrhodes]

Well this is even weirder... I took a look at the settings and just saw outputs for S/PDIF (motherboard) and monitor sound. Both were at 100%. So I went and plugged May in directly to the PC and viewed the USB output both in Windows and in the Holo ASIO driver control panel - everything set to 100%.

 

Then I took a listen and it's definitely back to what I'd expect - PCM is about 6 dB louder than DSD. Listening to the same album I had on in DSD - ...and Justice for All - and had it at -18 dB on DSD, and it's *loud* now in PCM at -28 dB or thereabouts.

 

So for right now, either my adjustments (I played with the sliders, setting them back to 100) "took" or something is up with Zen Stream in NAA PCM mode, I think?

 

Either way, direct from HQ Player PC --> PCM --> May is not quieter than DSD.

 

Thanks for the nudge @GoldenOne!