HQ Player

[asdf1000]

6 hours ago, Miska said:

 

If using PCM mode (like in the test), I would set "DAC Bits" to 16 or 17 with LNS15 noise shaper. This keeps linearity in the very flat range, while providing 20+ bits worth of practical dynamic range in audio band.

 

 

Can the UP Board Gateway do PCM ~1.5M with LNS15 with all poly-sinc filters + 2-channel convolution , as a HQP Embedded machine?

 

 

 

[AnotherSpin]

13 hours ago, jimdukey said:

Does anyone listen to music?

Or just the equipment?

Background Music is an Oxymoron.

 

I listen

[asdf1000]

13 hours ago, jimdukey said:

Does anyone listen to music?

Or just the equipment?

Background Music is an Oxymoron.

 

Need the equipment to listen to the music.

 

I guess I could just play Tidal through my iPhone's speakers but I enjoy the music a lot more through my Focal Utopia headphones

 

[StreamFidelity]

10 hours ago, Miska said:

If using PCM mode (like in the test), I would set "DAC Bits" to 16 or 17 with LNS15 noise shaper.

 

@Miska 

I will receive the T+A SDV 3100 HV DAC shortly and have a few questions.

 

For PCM: which number of bits with LNS15 & 768kHz is recommended?

 

The SDV has a built-in NAA (Network Audio Adapter). Is this connection preferable or equivalent to USB?

[Miska]

2 hours ago, StreamFidelity said:

 

@Miska 

I will receive the T+A SDV 3100 HV DAC shortly and have a few questions.

 

For PCM: which number of bits with LNS15 & 768kHz is recommended?

 

Since, AFAIK, it is delta-sigma DAC chip (likely PCM1795), just send it 32-bit data, best material for it's internal DSP.

 

2 hours ago, StreamFidelity said:

The SDV has a built-in NAA (Network Audio Adapter). Is this connection preferable or equivalent to USB?

 

I would use the network connection since it provides galvanic isolation, as long as you don't spoil it with shielded (audiophile) network cables. So just use standard CAT6 UTP.

 

[Miska]

6 hours ago, asdf1000 said:

Can the UP Board Gateway do PCM ~1.5M with LNS15 with all poly-sinc filters + 2-channel convolution , as a HQP Embedded machine?

 

This is quite impossible to say at least regarding convolution, because it so much depends on (real) size of the convolution filters. I've tested upsampling up to 768k with it, but higher rates only as a NAA. Mine is now serving as a NAA in front the of T+A HA 200.

 

[asdf1000]

1 hour ago, Miska said:

 

This is quite impossible to say at least regarding convolution, because it so much depends on (real) size of the convolution filters. I've tested upsampling up to 768k with it, but higher rates only as a NAA. Mine is now serving as a NAA in front the of T+A HA 200.

 

 

Hmmm ok I understand.

 

Without convolution, can your Up Board Gateway do PCM ~1.5M to your Holo Spring 2 L2 , as a HQPe machine  ? With the most CPU intensive poly-sinc filter ?

[patatras]

Hi

I am using HQP with pcm conversion to DSD 512.

My processeur is a xeon D1540 with 8 hearts at 2 ghz.

They work at 75 - 80% with 25% of ram.

 

With many filters and modulators as DSD7, DSDv2 256... I have drops of sound (2s) each about 30s.

I set buffer time to 50 or 100ms and it's the same.

What can I do else ?

Thank's

Pierre

[pcmchild]

20 hours ago, asdf1000 said:

 

Here's a stellar new and very transparent headamp ($300) that should pair nicely with Ares II :

 

https://www.audio “science” review/forum/index.php?threads/smsl-sh-9-thx-headphone-amplifier-review.17519/

 

ASR links don't work here for some reason

I really appreciate your suggestion and that you are trying to help. But I really don't want to go with any other box (and cables) than the PC, dac and headphone. That is one of the reasons (among other reasons) that I went with HQPlayer instead of mscaler.

I don't want network boxes, nothing. Just desktop work PC and dac. 

But again, I thank you.

[ASRMichael]

19 hours ago, Miska said:

I would use the network connection since it provides galvanic isolation, as long as you don't spoil it with shielded (audiophile) network cables. So just use standard CAT6 UTP.

 

Jussi. Can you list the all the shielded cables you have tried? 

[Miska]

1 hour ago, ASRMichael said:

Jussi. Can you list the all the shielded cables you have tried? 

 

Why? I don't need to try STP/SFTP to know what happens with it. It is simple electrical thing.

 

[bogi]

9 hours ago, pcmchild said:

But I really don't want to go with any other box (and cables) than the PC, dac and headphone.

That's understandable. Bigger issue is, that you rarely find DAC + headamp combo with both parts on equal quality level. Often headamp acts more like an add-on (because not all will use it) with lover specs. And it depends on headphones you want to use, because output power as well as output impedance has to match your headphones. I would say there are 3 important things with every headamp:

- the sound quality given by distortion level

- the output power into headphones in relation to headphones sensitivity to drive them correctly

- the impedance match between headphone output impedance and the impedance of headphones itself. It is recommended the ratio be 1:8 or higher. So for example, for typical electrodynamic headphones of impedance 32 Ohm it is not suitable to drive them from 10 Ohm headphone output. With planar magnetic headphones it could be OK (their impedance does not depend on frequency), but they have usually higher driving current requirements, especially to pronounce high quality bass, since they are mostly able to do it from 20Hz (or even less) and the most energy of big bass attack appears here.

 

So it may be not so easy to make a perfect match if you want to avoid compromise solutions. Therefore it is more usual to have separate combo of DAC + headamp. But that is topic rather into the headphone section of this forum.

 

If you would like to use a good combo, you could try RME ADI-2 DAC FS. But AFAIK it is not quite ideal for the PCM to DSD and Linux case if you want that (Miska please correct me if I am wrong), because only DoP (no native DSD) works on Linux and with DoP there is no direct DSD path with RME ADI-2 - the signal path is unconditionally routed into the DSP section. No such issue with Windows or Mac, where native DSD with direct DSD path is supported.

[asdf1000]

13 hours ago, pcmchild said:

I really appreciate your suggestion and that you are trying to help. But I really don't want to go with any other box (and cables) than the PC, dac and headphone. That is one of the reasons (among other reasons) that I went with HQPlayer instead of mscaler.

I don't want network boxes, nothing. Just desktop work PC and dac. 

But again, I thank you.

 

Checkout new iFi Neo iDSD - features headamp, analogue volume control (resistor ladders) and direct conversion of DSD to analogue.

 

All in one box !

 

https://ifi-audio.com/wp-content/uploads/2020/10/Neo-iDSD-manual-Ver1.0.1.pdf

 

[fredg_31]

Hi all,

 

After a few weeks testing and trying to understand hqplayer settings, I'm quite happy with poly-sinc-short-lp and poly-sinc-ext2.

 

I'm using PCM only (old DAC) @192KHz.

 

I'm not confident with the number of bits and dithering settings -> If I use 24 bits, does it change something using no dithering?

 

My DAC (Hegel HD20) is using two AD1955 chips and one AD1896 sample rate converter. I don't know if 24 bits setting is the best choice or if it will be better to change it to a lower value (Important note, I'm using a Weiss INT203 firewire unit to feed the dac).

 

If someone have an idea for my dac, I'll take it with pleasure!

 

Thanks for your help guys.

Fred (Toulouse, France)

[Miska]

1 hour ago, fredg_31 said:

I'm not confident with the number of bits and dithering settings -> If I use 24 bits, does it change something using no dithering?

 

My DAC (Hegel HD20) is using two AD1955 chips and one AD1896 sample rate converter. I don't know if 24 bits setting is the best choice or if it will be better to change it to a lower value (Important note, I'm using a Weiss INT203 firewire unit to feed the dac).

 

Yes, 24-bit setting is correct in this case. And you should use some kind of dithering, for example TPDF or Gauss1, this is needed because the calculated sample values contain more precision than 24-bit integer can hold. Truncating or rounding the values create systematic signal dependent error that is called quantization distortion. Dither is used to randomize this error so that it is not systematic or signal dependent anymore, and thus there is no distortion.

 

[fredg_31]

2 minutes ago, Miska said:

 

Yes, 24-bit setting is correct in this case. And you should use some kind of dithering, for example TPDF or Gauss1, this is needed because the calculated sample values contain more precision than 24-bit integer can hold. Truncating or rounding the values create systematic signal dependent error that is called quantization distortion. Dither is used to randomize this error so that it is not systematic or signal dependent anymore, and thus there is no distortion.

 

Thanks Miska, for your reply and for your work, HQPlayer is really a game changer in my system, I loved Daphile, but Roon + HQplayer is on another level!

 

TPDF or Gauss1 will be better than NS9 in my case? I hear a difference, but I can't tell wich one is better, it seems to change between 2 albums.

 

NS9 is more precise but sometime harsh

TPDF, Gauss1 or Shaped are more relaxing, but less precise

 

[Miska]

13 minutes ago, fredg_31 said:

Thanks Miska, for your reply and for your work, HQPlayer is really a game changer in my system, I loved Daphile, but Roon + HQplayer is on another level!

 

TPDF or Gauss1 will be better than NS9 in my case? I hear a difference, but I can't tell wich one is better, it seems to change between 2 albums.

 

NS9 is more precise but sometime harsh

TPDF, Gauss1 or Shaped are more relaxing, but less precise

 

Any of those is fine, depending on which one you prefer. NS9 will give some extra dynamic range in audio band. But at 24-bits the difference in that respect is small.

 

[ASRMichael]

9 hours ago, Miska said:

 

Why? I don't need to try STP/SFTP to know what happens with it. It is simple electrical thing.

 

Imagine if someone said upsampling is a waste of time?! Which I’ve seen this comment on various forums. To which I’ve encouraged many people to go & trial HQPlayer. Maybe you should follow my lead. I’d be very happy to send you a Sablon cable to trial? 

[Miska]

55 minutes ago, ASRMichael said:

Imagine if someone said upsampling is a waste of time?! Which I’ve seen this comment on various forums. To which I’ve encouraged many people to go & trial HQPlayer. Maybe you should follow my lead. I’d be very happy to send you a Sablon cable to trial? 

 

I said STP/SFTP cables spoil galvanic isolation of ethernet because it connects grounds of two devices to each other. One of the points of using ethernet is to get the galvanic isolation...

 

Always use UTP cable for connecting audio devices. What ever brand of choice, as long as it conforms to ethernet cabling standard specifications.

 

[LoryWiv]

On 11/14/2020 at 3:54 PM, pcmchild said:

Delete.

 

 

[asdf1000]

6 hours ago, ASRMichael said:

16 hours ago, Miska said:

 

Imagine if someone said upsampling is a waste of time?!

 

Ground loops is an engineering thing, like Miska's original reply... i.e. can be measured.

 

Same as upsampling...

[AnotherSpin]

17 hours ago, ASRMichael said:

Imagine if someone said upsampling is a waste of time?! Which I’ve seen this comment on various forums. To which I’ve encouraged many people to go & trial HQPlayer. Maybe you should follow my lead. I’d be very happy to send you a Sablon cable to trial? 

 

I use Ethernet cables between mac mini and the NUC NAA endpoint. I have tried and compared several UTP cables and several STP (different from Supra, Audioquest). UTP cables in general sound natural in comparison, while STP types seem to "add" something in the top and bottom areas. This can make an initial impression, but bring certain irritation with resulted fatigue in the long run. Finally, I got Meicord UTP cables, which sounds both expansive and natural to me.

[Zauurx]

Try this (for the final part - before streamer) :
Cheapest but excellent !!
Micro SlimRun Cat6, UTP, Pure Bare Copper Wire, 32AWG

[Gavin1977]

21 hours ago, ASRMichael said:

Imagine if someone said upsampling is a waste of time?! Which I’ve seen this comment on various forums. To which I’ve encouraged many people to go & trial HQPlayer. Maybe you should follow my lead. I’d be very happy to send you a Sablon cable to trial? 

Depends upon the implementation of DAC - prior to NOS & actually understanding what a 'DSD Direct' DAC was I would have said the same thing.

[bipet]

21 hours ago, Miska said:

I said STP/SFTP cables spoil galvanic isolation of ethernet because it connects grounds of two devices to each other. One of the points of using ethernet is to get the galvanic isolation...

 

Always use UTP cable for connecting audio devices. What ever brand of choice, as long as it conforms to ethernet cabling standard specifications.

This is interesting to me. I have a whole load of different cables in my network, mostly Cat 6 or 7, probably. Some Audioquest, lots of Amazon basiscs etc. I suppose it might be worth checking the last run, so the cable from the switch to the streamer is UTP and not STP/SFTP??