HQ Player

[Bertel]

7 minutes ago, Miska said:

ASDM7EC -> ADSM7ECv2

 

Seems I have been missing something - are there new versions of the modulators coming? Why? I mean, how will they differ from the previous EC, what's the focus and benefit? Interesting! :-)

[Bertel]

1 hour ago, StreamFidelity said:

How do you get the analog into the digital? So the way from the turntable via cinch cable to a USB audio interface or something else?

 

30 minutes ago, Miska said:

For digital RIAA you also need some pickup pre-amp that doesn't have RIAA correction built-in.

I use Phaedrus Audio‘s Groove Sleuth, together with their Phlux cartridge (Stereo Lab Software does the RIAA correction and more for me). I use the Benchmark ADC1 USB as A/D converter. Certainly doing all of this in the digital domain,instead of analog components, has helped getting my experience of vinyl reproduction to a new (IMHO superior) level, at moderate expense in comparison.

 

[Bertel]

20 minutes ago, StreamFidelity said:

 

This is very interesting. Does the "line out" go into the "mic-in" of the computer?

i don‘t use the Micro but the Groove Sleuth II. Its output goes into the ADC1 which outputs over USB to my PC where I do the recording with Audacity, and RIAA correction then with Stereo Lab.

 

3 hours ago, Miska said:

Technically, with ADI-2 Pro, you could even run DSD256 -> RIAA -> DSD256.

 

Now this has me intrigued! I can replace the ADC1 and keep the current process (need to figure out though whether output to PC over USB is possible, or preferrable alternatives) and still do the DSD256->RIAA->DSD256 in realtime when I feel like spinning a record on my Technics SP-10 Mk3 😄

[Bertel]

11 minutes ago, Miska said:

Why not in realtime? ;)

Because I do a lot more in Stereo Lab than just the RIAA correction, e.g. noise reduction, specific adjustments for the PHLUX cartridges etc. Well, to be precise I do needle drops at 2/3 speed to minimize inner groove distortion and spin up in digital domain. Plus I just want to rip the vinyl once to digital, no hassle with physical media for me ;-)

[Bertel]

Gents (and especially @Miska of course 😉) - can I please check back quickly:

 

I just received my Topping D90SE, which I am going to run alongsde my Holo May KTE to compare both for a longer period. The D90SE uses an ESS ES9038PRO chip.

 

Can't recall what the latest was on optimal settings for this chip - do I remember correctly that it was

Output = PCM

DAC bits = 20

Dither = LNS15

Max sample rate = 1.536M

Adaptive Output rate = checked

SDM = none (so no DoP)

Buffer = default

 

I think I remember that PCM was recommended over DSD or the ESS' internal processing, along with the above parameters. Is that also true for the PRO version of the 9038?

 

[Bertel]

26 minutes ago, bogi said:

With SDM (DSD) output from HQPlayer you can skip complete oversampling in DAC

Oh. i must have mixed thatup with some other DAC implementation then (likely Chord I suppose) - thought it was the other way around, PCM was direct and DSD further modulated.

 

Many thanks @bogi, I‘ll then happily use the exact same configuration as with my Holo May KTE, i.e. everything to DSD256 with ASDM7ECv2 and poly-sinc-ext2. Makes it even more interesting to compare May and D90SE.

[Bertel]

20 hours ago, Miska said:

Recommended is to use DSD512 or even better DSD1024 (which is officially supported by ES9038Pro) with ASDM5EC/ASDM5ECv2 modulator. For DSD1024 you cannot do this in realtime, but you can do ASDM5 at DSD1024. This gives you minimal processing through the chip without rate conversions.

Many thanks @Miska !

 

I can‘t do DSD1024 at the moment, the Topping D90SE accepts DSD1024 only over IIS, its USB input is limited to DSD512. (Might try DSD1024 at a later point in time, have read that although advertised the D90SE currently doesn‘t work with DSD1024 over IIS, have contacted their support to clarify).

 

Do I understand correctly that for DSD512 indeed ASDM5ECv2 is recommended and preferred over ASDM7ECv2 and AMSDM7 512+fs? Can I ask why? (I have tried all three, I think I can hear some differences but not sure yet which I‘d prefer, so would want to understand the technicalities first)

[Bertel]

17 minutes ago, Miska said:

Yes, that's correct. It is because ESS has totally different kind of DSD D/A conversion filter than other chips, it is not nearly as steep as other converters leading to more ultrasonic noise left-over if you use steep modulator. This assuming typical analog reconstruction filter. If a DAC manufacturer would include a steeper analog reconstruction filter, then 7th order modulators would be more recommended.

 

Ah, I see - many thanks @Miska for the explanation!

 

Well, I guess for exactly this reason the D90SE has a "fast roll off corrected minimum phase" filter (labelled F-7 in the below graph from the website) to get rid of that ultrasonic noise - but that's for PCM. There are four filters for DSD, the description just calls out frequencies of 47kHz (that's the default), 50kHz, 60kHz and 70kHz. I have kept the default of 47kHz, which I assume to be the steepest, but I have no further details...

[Bertel]

16 minutes ago, Miska said:

Those are PCM oversampling filters and totally unrelated.

 

All those DSD filters are equally steep, just corner point frequency is different.

Yes, that's why I said "but that's for PCM" 😉

 

I just looked at the data sheet for the ES9038PRO (that Mouser provides here: https://www.mouser.de/datasheet/2/1082/ES9038PRO_Datasheet_v3_6-1923511.pdf) and on page 21 found ESS' recommendation "Note: 47.44k filter should only be used for PCM data. Recommended settings for DSD data are 50k, 60k and 70k". The filter profiles are shown on page 54 (attached here).

 

So I guess I'll rather select the 50kHz filter in the D90SE, and go with the recommended ASDM5ECv2 modulator in HQPlayer.

[Bertel]

Strange - I changed my mainboard and CPU (Gigabyte Z390 Aorus Master and i9-9900K to Gigabyte Z690 Aorus Elite DDR4 and i9-12900K) and now get an error message when starting HQPlayer4 Desktop (see attached). BIOS is set to Load Optimized Default and XMP1 Profile as recommended.

 

@Miska, can you help please, what does that mean, and what can I do to fix?

[Bertel]

15 minutes ago, Miska said:

With version 4.15.1?

Ah, apologies - no, that was an older version (don‘t recall which, some 4.14.X).

 

Have installed the latest 4.15.1, had still massive stutters (every 2-3 seconds) with Multicore DSP checked - now have it grayed out (with Adaptive Outpur Rate checked) and all works well.

 

Now checking HQplayer OS with Embedded, which was stuttering as well. Had the x64gen version, if I understand correctly will need to try the x64amd version

[Bertel]

1 hour ago, Bertel said:

Now checking HQplayer OS with Embedded, which was stuttering as well. Had the x64gen version, if I understand correctly will need to try the x64amd version

Confirmation that with the x64md version of HQPlayer OS with Embedded everything is working perfectly 👍

[Bertel]

@Miska quick question please: When sending 705.6kHz or 768kHz PCM in 32 bit to the Topping D90SE and its ES9038PRO, using poly-sinc-ext2 - is TPDF dither or LNS15 noise shaper recommended? Or a different dither or noise shaper? I understand that the noise shapers have an advantage when using <24 bit, but not sure what works best at 32 bit in terms of dynamic range and distortion?

[Bertel]

With all reports of failure, stuttering etc., just want to say: also 4.15.4 is working perfectly fine for me, with brilliant sound. But I‘m only doing 32bit 786kHz with poly-sinc-ext2 and NS15 to my Topping D90SE (which I meanwhile clearly prefer over my Holo May KTE), using a 12th gen i9-12900K with 64GB RAM, directly connected over USB using ASIO.

[Bertel]

23 minutes ago, Miska said:

 

You get most consistent performance out of that D90SE DAC with ASDM5ECv2 at DSD256... Remember to set the DSD filter to 50 kHz.

 

Following Archi‘s measurements (see here: http://archimago.blogspot.com/2021/10/measurements-review-topping-d90se-dac.html ) I‘m not using the D90SE with DSD, as it measurably performs substantially worse with DSD compared to PCM (also fully consistent with my subjective experience). Using 32bit instead of 24bit with PCM, as you suggested, for me clearly performs/sounds best.

[Bertel]

11 minutes ago, Miska said:

 

I measured it, results here:

https://community.roonlabs.com/t/topping-d90se-hqp-settings/183551/6?u=jussi_laako

 

It is certainly most consistent performer at DSD256. PCM has host of issues, varying from noise floor modulation to poor low level linearity.

 

Thank you, very interesting! Will need to thoroughly compare it to Archi's measurements, from my understanding they indicated something different, hope I can find a consistent bottom line...

 

Do you? I mean, can you see some alignment of your measurements with Archi's? Would be suprised if they differ so substantially.

[Bertel]

Hmm, hard for me to understand why Amir (ASR), Wolf (I7lab), Archi and others measure it so differently/superior…

 

EDIT: well, they do 16/44.1 up to 24/384 max, maybe that‘s why?

[Bertel]

14 minutes ago, Miska said:

Certain input signals triggers it misbehave, like in this sweep test at 705.6k rate when input frequency goes over 10 kHz (by the way, same happens with all DSD rate inputs too):

 

 

@Miska, in the measurements you showed here: https://community.roonlabs.com/t/topping-d90se-hqp-settings/183551/6?u=jussi_laako , you showed the 1kHz graph at DSD256 with ASDM5ECv2, which looks pretty clean. With your above statement, are you saying that the 10kHz measurement (which is not shown at the Roonlabs link) looks like the above also for DSD256 with ASDM5ECv2...? While 1kHz is fine, 10kHz is wrecked like this?

[Bertel]

On 2/11/2022 at 9:36 AM, Miska said:

 

When it misbehaves, the noise floor bumps up.

 

 

It takes me about a day to make my normal measurement set. I have a multitone result for DSD256 stored:

 

I bought the SMSL M500 MkII and received it on Friday, but it has strange issues with occasional electrical noise (low level stuff that sounds like quantisation noise or so, plus high pitched squeal at low volume at the end of tracks at fade out), although I feed it DSD512 only - guess that‘s what you mean with „it misbehaves“… 

 

@Miska have you tested and measured any ESS9038Pro based DAC yet which works without issues and quirks and plays DSD512 without any flaws? I like the very clean and transparent sound a lot and definitely want such a DAC in my collection, but haven‘t found properly engineered units yet it seems…

[Bertel]

2 hours ago, Miska said:

It works, but as I mentioned before, it needs to stick to those 16x rates since boot-up not to misbehave. You can also use it with DSD256 if DSD512 misbehaves. What is it connected to? Some NAA? Mine is running mostly from iFi Zen Stream, so max DSD256.

Use 5th order modulator.

 

Thanks. As said, it doesn't work in my case - I start up with DSD512, use nothing else, and still after a couple tracks it starts to 'misbehave'. Will try DSD256 just to check whether it makes a difference. Also have a SMSL SU-9n coming, which I hope is basically the same as M500 Mk II minus headphone amp and MQA, and the latter is what I hope will get rid of also this misbehavior ( as @copy_of_a you had reported that it at least doesn't have the 16x issue)

 

I connect my DACs to a Signature Rendu SE with its NAA in play. Am playing DSD512 with ASDM5ECv2 and poly-sinc-ex2 which runs just fine, no glitches or anything on server side.

[Bertel]

On 2/14/2022 at 3:54 PM, Miska said:

There are more 9028PRO based DACs. And I think I will soon'ish get another one. For some reason lot of more well known names use 9028PRO.

 

Likely exaSound is one of those that work fine but it is too expensive for me. I have just the now old e28 model.

Well, interestingly enough, I just saw that exaSound will introduce their new S82 next week at Florida Audio Expo - and it‘s using the ESS9038Pro 😉 (Like the S88, which John Atkinson measured pretty well, but I don’t need 8 channels…) Will keep an eye on that, especially once measurements become available. Could be just what I‘m looking for.

[Bertel]

12 hours ago, Miska said:

Autonegotiation should appear ON, otherwise it is hinting that there may not be the necessary negotiation happening between nodes.

Out of curiosity I looked this up on my system as well, and found a confusing (for me at least) situation:

 

Auto-negotiate is enabled both in the Mikrotik switch and on my Intel X520-DA2 NIC, but ethtool -a for the NIC reports Autonegotiate to be off… Any hint where I could look or what to do to resolve this? I sometimes (once or twice a week) have a few network dropouts when HQPlayer Embedded is playing to the NAA, playback stops once or twice for a few seconds, then resumes automatically, so I‘d assume this is the reason.

 

[Bertel]

7 hours ago, Miska said:

 

For sonicOrbiter OS, you get updates through SGC. If you are on 4.1.1 you should be fine for most cases. Most recent 4.2.0 adds some fixes to deal with certain buggy XMOS firmwares, this may or may not be applicable to you depending on your DAC.

 

But functionally, there's not much difference since 4.1.1.

 

I typically upgrade directly through the sonicorbiter.com web view of my Signature Rendu. Did that again just now, a few packages were updated, but the NAA app is still on version 4.1.3. So no 4.2.0 yet in the upgrade process through sonicorbiter.com. Not that this would be a problem, just sharing 😉

[Bertel]

2 hours ago, Carlos269 said:

Was Bertel’s issue resolved? What was the root-cause or fix for it? I’m having the exact same issue. As I said before it worked flawlessly for months but all of the sudden this noise is driving me crazy. One other data point, if I don’t use my setup for a few days the issue is not there when I first start playing music but after about 15 minutes of playing it shows up and does not go away unless I shut everything down and de-energize the dac and disconnect the  USBcable.

 

Well, the issue I had is likely due to the implementation of the ESS9038PRO chip, as Miska had pointed out. I also tried the SMSL SU-9n which behaves exactly the same (fine with DSD256, noise with DSD512, as well as not being able to play 48k at the right speed). I now have the VMV D1se which works fine and doesn't have these issues.

 

As your Lampi doesn't use the ESS chip, I don't think there's much to learn and transfer from my example. Everything up to and including DSD1024 plays fine with my Holo May KTE, so I don't think there's any influence of the upstream part of my chain (optical, Signature Rendu, USB) to this ESS chip issue.

[Bertel]

@Miska, not sure whether this has been discussed before, Search didn't find me anything:

 

I am going to use FIR filters provided by @mitchco for adjusting the FR of some of my headphones. There's a filter for every PCM sample rate (44.1kHz, 48kHz, 88.2kHz etc. up to 384kHz). I know that I just could pick e.g. the 384kHz filter and let HQPlayer do the downsampling to match the respective source rate, but I want to use the specific matching filters for each rate.

 

Roon and @mitchco's own Hang Loose Convolver allow loading the ZIP file which contains all sample rate specific filters as WAV files, so that automatically in the process the corresponding filter file which matches the source rate is selected - as far as I see that's currently not possible with HQPlayer, right? Is it possible to have that implemented? Or is there maybe a way to script this process e.g. in the Matrix Pipeline setup or so?