HQ Player

[GoldenOne]

16 hours ago, Miska said:

 

No need for high sample rate inputs, just 64-bit source at 44.1k input is enough:

You can also compare the 512+fs and the regular one this way.

Thanks for this. The file I was using WAS 64 bit, but I think the issue is that the FFT was factoring in some of the silence from HQP at the start/end of the file.

Usually I use RASA's 'remove digital silence' option but I hadn't considered that this wouldn't function if the source is DSD and therefore digital silence isn't really a thing.

I need to trim the files to cut out the silence and re-check.

Additionally you mentioned I was not using the best filters, which filters would you recommend and how come Sinc-M/Sinc-L would be a limiting factor?

[camott]

On 12/4/2023 at 1:28 PM, Miska said:

 

 

From just looking at these measurements, benefit of 512+fs in the audible band appears to be *far* more significant than the minor differences in the ultrasonics. Confused as to why one would ever use regular modulator ... what am I missing?

 

[Jud]

55 minutes ago, camott said:

 

From just looking at these measurements, benefit of 512+fs in the audible band appears to be *far* more significant than the minor differences in the ultrasonics. Confused as to why one would ever use regular modulator ... what am I missing?

 

 

From Miska a couple of days ago:

 

Quote

Mostly class-D is sensitive to ultrasonic noise, because it works like "ADC". When the input exceeds the switching frequency which is like sampling rate, aliasing down to lower frequencies happens.

 

[camott]

7 minutes ago, Jud said:

 

From Miska a couple of days ago:

 

 

 

Yep I was aware of that, but now that I am looking at an actual plot, I don't see that much difference in ultrasonic noise, whereas there is a big difference in audible band noise.

[bogi]

10 hours ago, camott said:

 

From just looking at these measurements, benefit of 512+fs in the audible band appears to be *far* more significant than the minor differences in the ultrasonics. Confused as to why one would ever use regular modulator ... what am I missing?

 

Maybe I formulated it too strictly but the principle (filter bandwith at given output rate vs source content bandwidth) is explained on the previous page:

 

[Atriya]

On 6/15/2023 at 8:36 PM, Strikertango55 said:

 NAA output clNetEngine::Execute(): clNetEngine::Disco(): clSocket::SendTo(): sendto(): Unknown error

This is the error msg I'm getting

 

I get this same error, and then HQPlayer crashes when any audio is played. I do see the NAA device in Output Device Settings-> Device and it is selected. But when I close the Settings window I get this error.

 

The NAA is a RoPieeeXL device that's in the same WiFi network as my laptop running HQPlayer.

 

On the same setup, Roon Bridge works fine.

[Apollo]

23 minutes ago, Atriya said:

 

I get this same error, and then HQPlayer crashes when any audio is played. I do see the NAA device in Output Device Settings-> Device and it is selected. But when I close the Settings window I get this error.

 

The NAA is a RoPieeeXL device that's in the same WiFi network as my laptop running HQPlayer.

 

On the same setup, Roon Bridge works fine.

 

 

Maybe already checked and answered, but have you made the right selection in HQPlayer Configuration:

 

If your DAC is connected and powered on (and it huld!) , then you get 2 options to choose from.

The one that needs to be selected, is the one with 'DAC' in it.

[satshanti]

I while back I asked about possible under 500 Euro upgrades to my Topping D30 DAC. A recommendation for a Cayin RU7 was given, and I acted on that. I received it about a week ago and started playing around with it on DSD256. The sound was very good, much better and less glaring and harsh than the Topping, a natural, organic sound that was very appealing. As I burned it in, it opened up more and more. As it's a mobile DAC, it does need some adapters for home use, but it's definitely excellent value for money.

 

I would have kept it, but it unexpectedly gave up the ghost one day, no longer powering up when plugging it in. Cayin offered to replace it, but I didn't want to wait that long and, as I now had a taste of what's possible with a better DAC, bought myself a SMSL D300, delivered next day. That  one is even better, not surprising as it costs almost twice as much. It has a smooth, almost analog sound, no harsh highs, great mids and the cleanest bass I've experienced so far. I might still get another RU7 for mobile use though, as it made a very good impression. 

 

So, after that bit of sharing of my experiences, I do have a question and it's about 48k-based DSD. I don't have a top level PC and I like the sinc-short/medium/long filters the best. They sound the most natural and have the greatest micro-details. But because my Topping couldn't run 48k DSD, I couldn't run any of these filters with 96k and 192k material. I took a chance with the RU7 and it did support it. I read ( I think it was Bogi?) that the D300 supports 48k DSD, so I was looking forward to using those better filters with my hi-res files.

 

However, my experience is that of these 3 DACs only the RU7 genuinely plays 48k DSD on my system. Both the Topping and the SMSL play it, but at a slightly lower speed, presumably at 11290 in stead of at 12288 MHz. With some music you hardly notice it, but with most music, my ears can't bear it for long. It just sounds off. (HQP shows the right rate BTW (12288) but of course all 3 DACs do not, as one would expect) I've not seen any mention of this phenomenon online. What could be the cause of it? Could it be the Rpi NAA? Anyone else have this experience? I've seen it mentioned that the Topping should be switched to 48k based PCM first, and then to DSD. I've tried that with both DACs but it doesn't make a difference. Any thoughts? 

[Miska]

12 hours ago, camott said:

From just looking at these measurements, benefit of 512+fs in the audible band appears to be *far* more significant than the minor differences in the ultrasonics. Confused as to why one would ever use regular modulator ... what am I missing?

 

Those audio band differences are also far beyond what is allowed by laws of physics in analog world.

 

So if we take a different kind of look...

 

Since the content available for purchase is generally at most 24-bit, we can first start from purely theoretical point of view for 24-bit source and look how -120 dB tone on 24-bit source looks like. Remember that 24-bit already exceeds what is possible due to laws of physics in analog world:

 

In practice, we can pass through 24-bit source bit-perfect through this conversion and back with either modulator.

 

And then if we feed some real hires music through:

Sorry for the mistake in label, this source is 2L DXD, 352.8/24 FLAC.

 

So the actual measurable analog world differences are in the ultrasonic range. But in any case, when you do DSD512 or more, you have the choice! There are cases where the 512+fs ones can be beneficial and give you extra peace of mind. Such as when you use HQPlayer's digital volume control, or for example do a down-mix of multichannel to stereo in HQPlayer (be it DSD or PCM source).

 

These 512+fs modulators in particular give you massive extra headroom when doing mixes of for example multichannel/mult-track DSD256 or DXD sources to DSD512/DSD1024 output.

 

So with 512+fs modulators you can do like -120 dB volume setting with HQPlayer's digital volume control without any SNR loss (in digital domain).

 

In the end, the limitations are posed by physics in analog world.

 

[Miska]

8 minutes ago, satshanti said:

However, my experience is that of these 3 DACs only the RU7 genuinely plays 48k DSD on my system. Both the Topping and the SMSL play it, but at a slightly lower speed, presumably at 11290 in stead of at 12288 MHz. With some music you hardly notice it, but with most music, my ears can't bear it for long. It just sounds off. (HQP shows the right rate BTW (12288) but of course all 3 DACs do not, as one would expect) I've not seen any mention of this phenomenon online. What could be the cause of it? Could it be the Rpi NAA? Anyone else have this experience? I've seen it mentioned that the Topping should be switched to 48k based PCM first, and then to DSD. I've tried that with both DACs but it doesn't make a difference. Any thoughts? 

 

This is typical, many DACs don't know how to switch the clocks when input is DSD. It is a bug in the USB interface firmware, a kind of bug that needs extra code not to switch the clock, since they switch clocks correctly for PCM sources. But unfortunately the original XMOS sample code has many bugs, including this one, and many of those "cheaper" companies just use it pretty much as-is with all it's bugs...

 

So the DAC ends up using clock intended for the 44.1k family for DSD always, including 48k DSD family content. This is why there is the specific "48k DSD" check box where you can tell HQPlayer that you are absolutely sure that the 48k-base DSD actually works with your DAC...

 

[Atriya]

4 hours ago, Apollo said:

 

 

Maybe already checked and answered, but have you made the right selection in HQPlayer Configuration:

 

If your DAC is connected and powered on (and it huld!) , then you get 2 options to choose from.

The one that needs to be selected, is the one with 'DAC' in it.

 

No, I see only one option for Device, but it shows the correct information (correct name of NAA machine, and the device connected to it via USB).

[camott]

6 hours ago, Miska said:

So the actual measurable analog world differences are in the ultrasonic range.

 

Ok thanks for the additional plots. I think I get it now! 

[bogi]

6 hours ago, satshanti said:

I read ( I think it was Bogi?) that the D300 supports 48k DSD, so I was looking forward to using those better filters with my hi-res files.

I answered in D300 thread, since IMO it is better to collect D300 specific info at one place.

6 hours ago, satshanti said:

That  one is even better, not surprising as it costs almost twice as much. It has a smooth, almost analog sound, no harsh highs, great mids and the cleanest bass I've experienced so far.

I like it too! 👍
I have this test CD

 

1. Marcus Miller - Outro Duction
2. Marcus Miller - Power
3. Brian Bromberg - Caravan
4. Brian Bromberg - Come Together (Lennon / McCartney)
5. Milt Hinton - Joshua Fit the Battle of Jericho
6. Michel Jonasz - Le lac gele
7. Michel Jonasz - Le temps pass e
8. Patricia Barber - Regular Pleasures
9. Roger Waters - Three Wishes
10. Dead can Dance - Yulunga (Spirit Dance)
11. Elph vs Coil - Caged Birds

 

I didn't find a shop offering it currently, but you can find those individual tracks on streaming services. The bass is ... uff! Really nicely pronounced on D300. Clean, effortlessly detailed, without any distortion on my setup (my headamp was designed for yet more power hungry HiFiman HE-6).
 

[bogi]

6 hours ago, satshanti said:

I've seen it mentioned that the Topping should be switched to 48k based PCM first, and then to DSD. I've tried that with both DACs but it doesn't make a difference. Any thoughts?

That trick worked for me with the oldest Topping E30, but I don't own it already 2 years. That time I wrote and used a script for that switching, one click on taskbar icon made the switch .

 

On Topping E50 switching to/from 48k based DSD rates works without issues (and not only my unit, confirmed on Slovak hifi forum).

 

But experiences may vary from model to model and firmware versions may influence it too.
On SMSL D300 like I mentioned above and just described in D300 thread I experienced no issues with 48k based DSD in my setup.

[skipspence]

Hi guys, 

Kinda unusual question here, but I guess can gain more attention in this topic than in Immersive Audio, anyway Chris can move it if it's OT here... 

Can anyone share their Matrix settings for HQP 2 channel downmix of the new wonderful Peter Gabriel Dolby Atmos TrueHD BD release ie Inside-Mix? 

Thanks in advance and sorry again if wrong thread here for this) 

[bogi]

1 hour ago, skipspence said:

Can anyone share their Matrix settings for HQP 2 channel downmix of the new wonderful Peter Gabriel Dolby Atmos TrueHD BD release ie Inside-Mix?

This is for 5.1 to stereo suggested by Miska:

https://audiophilestyle.com/forums/topic/19715-hq-player/?do=findComment&comment=585860

 

For 7.1 I don't remember to see HQPlayer matrix pipeline setup, but you could create it using dB values from this table:

 

[skipspence]

1 hour ago, bogi said:

This is for 5.1 to stereo suggested by Miska:

https://audiophilestyle.com/forums/topic/19715-hq-player/?do=findComment&comment=585860

 

For 7.1 I don't remember to see HQPlayer matrix pipeline setup, but you could create it using dB values from this table:

 

Thanks so much, but doing from second table seems to be the hard tusk for me, since I can't get right config for 7.1 from 12.1 channel layout...

Maybe someone can advise me further, thanks in anvance) 

[skipspence]

3 hours ago, bogi said:

This is for 5.1 to stereo suggested by Miska:

https://audiophilestyle.com/forums/topic/19715-hq-player/?do=findComment&comment=585860

 

For 7.1 I don't remember to see HQPlayer matrix pipeline setup, but you could create it using dB values from this table:

 

Tried to setup 7.1 to 2.0 according to Miska's suggestion for 5.1 from your post having set BL-BR to 7-8 and that's what I got, no idea if it's right or not, but it plays at least:) 

[timeslip]

Did something change in the last 2w with HQP 4 or 5?  I am running it on my mac mini m1 to a holo red via NAA and I suddenly get stutters.

[Miska]

12 hours ago, skipspence said:

Thanks so much, but doing from second table seems to be the hard tusk for me, since I can't get right config for 7.1 from 12.1 channel layout...

 

That table posted by @bogi looks like 12.1 to stereo settings. Although it looks a bit strange and not normalized gain. But you can try that one as well, maybe as a different profile.

 

[Miska]

25 minutes ago, timeslip said:

Did something change in the last 2w with HQP 4 or 5?  I am running it on my mac mini m1 to a holo red via NAA and I suddenly get stutters.

 

Not really. Especially for v4. Have you tried with local playback to figure out whether it is CPU load issue or a network issue?

 

[timeslip]

4 minutes ago, Miska said:

 

Not really. Especially for v4. Have you tried with local playback to figure out whether it is CPU load issue or a network issue?

 

 

I will have to reconfigure another machine to test.  I haven't made any changes other than macos updates on the m1 mac.  I was playing around with the settings and now I get the error message "requested filter not possible with this rate combination 44100 / 1536000, stop"

[Miska]

16 minutes ago, timeslip said:

I will have to reconfigure another machine to test.  I haven't made any changes other than macos updates on the m1 mac.  I was playing around with the settings and now I get the error message "requested filter not possible with this rate combination 44100 / 1536000, stop"

 

You are asking for fixed 1.536M output rate (48k family), but the filter you have selected cannot convert between rate families.

 

[timeslip]

What setting did I mess up for this in trying to fix the stutter issue?

[Miska]

2 minutes ago, timeslip said:

What setting did I mess up for this in trying to fix the stutter issue?

 

That error message is due to change in filter. But I suspect the issue may be that you get higher load because of the fixed output rate request? Does the situation change if you enable "Adaptive output rate" and use the filter you've used before?