[bogi]

Then of course you have lot of headphone DAC/amps where there is only digital volume control.

Of course, including typical portable ones. In such a case HQPlayer volume control is the best option available. Taken from many discussions over the years, it is generally considered to be excellent.

For PCM, HQPlayer volume control is much much better option than DAC chip volume control. With DAC chip based volume control it is not possible to avoid intersample overflows. It's because volume control in DAC chips is performed after the 1st oversampling stage and not at input rate.

For direct DSD, HQPlayer is the only known option to perform digital volume control (without intermediate conversion to PCM).

 

[shwnwllms]

I can only repeat what I already wrote: With lower digital attenuation and adjusting volume level on headamp the sound is more lively with more dynamics, bass impact, fuller instrument colors. When volume level is decreased significantly in digital domain and headamp volume level is set to max, sound becomes liveless, with less impact. This result repeats as patetrn with different DACs, headamps, headphones. I can create a poll somewhere to ask people which of the 2 approaches sounds better or what's the perceived difference.

Curious if you notice any difference between -6dB and -10dB (with regards to dynamics, bass impact, fuller instrument colors etc) or if this is only noticeable in your opinion when strongly digitally attenuating (to something like -30dB from your example)

Also regarding the Holo May, according to these measurements wouldn’t -4dB provide the sweet spot for lowest distortion?

 

[bogi]

Curious if you notice any difference between -6dB and -10dB (with regards to dynamics, bass impact, fuller instrument colors etc) or if this is only noticeable in your opinion when strongly digitally attenuating (to something like -30dB from your example)

When strongly digitally attenuating up to the level that you can run your headamp at full volume. I am not asking you to have the same opinion. Try it self and form your personal listening experience.

Also regarding the Holo May, according to these measurements wouldn’t -4dB provide the sweet spot for lowest distortion?

Note the point I wanted to say and take that graph as an example. There are some small variartions between units and also between channels of the same unit. You don't need necessarily to hear such small differences on your audio gear. Not all DACs show this behavior. Try such things self and form your personal listening experience with your devices.

What I provided is some generalization. You can search for similar measurements on L7audiolab, ASR, in different reviews etc.

 

[genefruit]

Curious if you notice any difference between -6dB and -10dB (with regards to dynamics, bass impact, fuller instrument colors etc) or if this is only noticeable in your opinion when strongly digitally attenuating (to something like -30dB from your example)

Also regarding the Holo May, according to these measurements wouldn’t -4dB provide the sweet spot for lowest distortion?

I see where you're going with the concept but I think one needs to consider unit and measurement variation. That said, interpret the graph how you see fit and adjust accordingly based on what you have after the MAY.

 

[shwnwllms]

I see where you're going with the concept but I think one needs to consider unit and measurement variation. That said, interpret the graph how you see fit and adjust accordingly based on what you have after the MAY.

As stated earlier, that’s exactly what I’ve done.

-3dB is typically the recommended maximum to avoid clipping, however I would still get clipping at -3dB (on poorly mastered recordings) so I used to bump it down to -4dB when I first started using HQP. (Volume dial will turn red in HQP to indicate clipping)

Lately I use -10dB (see @bogi posts for a more detailed explanation on lower distortion with some DACs). This also gives me the added benefit of more attenuation on my preamp (GS-X MK2 which only has a DACT 24-Step attenuator)

In my chain -10dB digital attenuation in HQP actually increased dynamics and led to a fuller more cohesive sound. But I’m running a preamp into a 300B SET with the volume maxed, I’m sure my setup isn’t representative of most use cases.

Reason I asked is I’ve seen both @jlaako & @bogi recommend -10dB at various points so I was just curious if @bogi noticed a difference between -6dB and -10dB or if his example was binary and only noticeable during strong digital attenuation where analog volume is maxed. He answered my question.

 

[bogi]

Reason I asked is I’ve seen both @jlaako & @bogi recommend -10dB at various points so I was just curious if @bogi noticed a difference between -6dB and -10dB or if his example was binary and only noticeable during strong digital attenuation where analog volume is maxed. He answered my question.

Such things are very individual and depend on the whole audio chain and also on listener. When it is about very small differences, some people may hear them and others not. Listening experience is not easily transferrable between people, since we are using different audio chains and different people may be more or less sensitive to different aspects of sound. I didn't tell that very small differences are not audible or that nobody can hear them.

There are more kinds of distortion which can be separately measured. I pointed to intermodulation distortion. Other often measured one is total harmonic distortion (THD). With some DACs also THD at 0dB can be higher than at -6 or -12 dB, but what I have seen was not majority of cases.

Source of distortion is not only DAC device itself, but also downstream equipment (each audio component) adds its own distortion. So you are listening to a composition of those. Source of distortion may be for example also noise spread from computer through digital cables besides of digital signal in analog form, or intermodulation distortion generated in class D power amps out of insufficiently filtered ultrasonic content. Therefore user experience with two different chains which contain the same DAC may be different. Some kind of distortion can mask small differences which would be otherwise audible. Therefore it is not so easy to pre-judicate audibility of small differences on different systems. Since there are many dependencies, one really needs to try on his chain what is audible for him and not to stick on opinion or experience of some other person.

My opinion is that everybody should perform listening tests to know what is audible for him on his chain and setup his equipment based on this on the first place. Theory and measurements should be used to know what may be reasonable to try.

 

[ohcrapgorillas]

Question: how do I determine the best DAC bits setting for my DAC?

Supposedly, for the Gustard R26, the correct bit depth is 15... however, setting the DAC bits to 15 sounds a little less dynamic to me than setting it to 17 or 18 bits.

Does anyone have any advice for how to determine the ideal DAC bits setting by ear? Suggested test tracks? So far 17-18 sound best to me but I'm having a very difficult time distinguishing between those two values. The effect is very subtle. If I could switch DAC bits on the fly, this would be no problem at all, but having to wait 10 seconds and restart the song every time makes it so difficult.

 

[jlaako]

Question: how do I determine the best DAC bits setting for my DAC?

Supposedly, for the Gustard R26, the correct bit depth is 15... however, setting the DAC bits to 15 sounds a little less dynamic to me than setting it to 17 or 18 bits.

Does anyone have any advice for how to determine the ideal DAC bits setting by ear? Suggested test tracks? So far 17-18 sound best to me but I'm having a very difficult time distinguishing between those two values. The effect is very subtle. If I could switch DAC bits on the fly, this would be no problem at all, but having to wait 10 seconds and restart the song every time makes it so difficult.

Only way to really determine the correct setting is by measurements.

Are you using LNS15 or similar noise-shaper at 705.6/768k output rate?

 

[ohcrapgorillas]

Only way to really determine the correct setting is by measurements.

Are you using LNS15 or similar noise-shaper at 705.6/768k output rate?

I have an oscilloscope, but unsure if that's the type of measurement tool required in this case. I'm guessing not, lol.

My problem is that, according to measurements, DAC bits = 15 is correct for the Gustard R26. But like I said, this sounds slightly less dynamic to me than a higher setting, so I'm not so sure.

I am using NS9 at 176.4/192 kHz as my output is limited by the AES interface. Hopefully I'll be upgrading to a Holo Red streamer at some point this year to get up to 705.6/768 kHz.

 

[ohcrapgorillas]

Hi Jussi,

I just updated to the newest version of HQPlayer Client, but I'm noticing a small bug:

1. Play some music via HQPlayer Client
2. Go to HQPlayer Embedded configuration tab, refresh devices. Music will stop playing.
3. Reconnect HQPlayer Client to HQPlayer Embedded, play music again.

The "filter" and "shaper" listed in HQPlayer Client under the Play page are the default values, TPDF and poly-sinc-gauss-long, even tho I am using sinc-M and NS9.

 

[jlaako]

I am using NS9 at 176.4/192 kHz as my output is limited by the AES interface.

I would expect quality to improve quite a bit at 705.6/768k rates...

2. Go to HQPlayer Embedded configuration tab, refresh devices. Music will stop playing.

This will cause HQPlayer server to restart, so your settings are reverted back to the set defaults.

 

[ohcrapgorillas]

I would expect quality to improve quite a bit at 705.6/768k rates...

Same, although unfortunately the LAN streamer built into the Gustard R26 still sounds worse at those rates than the Pi2AES at 176.4/192 kHz. Which is why I want a new streamer.

This will cause HQPlayer server to restart, so your settings are reverted back to the set defaults.

I'm confused. The HQPlayer server restarts when I refresh the devices, but it doesn't revert my settings back to default. The HQPlayer Embedded web interface shows that I'm still using sinc-M for 1x, sinc-L for Nx, and NS9 dithering. Only HQPlayer Client shows the default values.

 

[shwnwllms]

Same, although unfortunately the LAN streamer built into the Gustard R26 still sounds worse at those rates than the Pi2AES at 176.4/192 kHz. Which is why I want a new streamer.

https://up-shop.org/default/up-gate...g-memory-32g-emmc-board-w-vesa-plate-b10.html

 

[suno23]

https://up-shop.org/default/up-gate...g-memory-32g-emmc-board-w-vesa-plate-b10.html

this is a streamer?

 

[shwnwllms]

this is a streamer?

It’s an NAA (using Jussi’s bootable image) which is essentially a “streamer” for HQPlayer