HQ Player

[Miska]

On 9/13/2018 at 12:35 AM, dean70 said:

On an unrelated issue with pcm out, there is a snippet that plays from a previous track after stopping and starting a new playlist (like the contents of the last buffer) that only occurs with pcm out. This has been occurring over the last few versions.

 

I've been trying to reproduce this, without success. I wonder what setting could be triggering this...

 

Are you using something like ASIO driver on Windows? It could be also ASIO driver not clearing it's buffers between stop and restart.

 

[Yviena]

@Miska Hmm do you think HQplayer could take advantage of the new tensor cores in the 2000 series?

 

Hopefully it will be able to DSD512 on my other DAC.

[Miska]

24 minutes ago, Yviena said:

@Miska Hmm do you think HQplayer could take advantage of the new tensor cores in the 2000 series?

 

Hopefully it will be able to DSD512 on my other DAC.

 

No, Tensor cores are only for AI and not useful for anything else.

 

But if they have not crippled pure computational functionality, 2000 series could be potentially nice.

 

[tboooe]

2 hours ago, juanitox said:

every times appears a new filter there is a Wow effect on me ,

but at the end only one can rul them all

P/S/S?

Poly since short?

[tboooe]

2 hours ago, juanitox said:

every times appears a new filter there is a Wow effect on me ,

but at the end only one can rul them all

P/S/S?

Poly sinc short?

[joelha]

Jussi,

 

First and most of all, congratulations on developing the poly-sinc-ext2 filter. It's the best I've heard in your or any digital player.

 

While the more experienced users of HQPlayer may know the answer to these questions, I'll ask anyway.

 

What's the best way to know when a new version of HQPlayer becomes available?

 

Do your filters still operate, if selected, when there's no upsampling employed?

 

Thanks for a truly great player.

 

Joel

[Miska]

23 minutes ago, joelha said:

What's the best way to know when a new version of HQPlayer becomes available?

 

Either following my Twitter account, or the Facebook group. Twitter account has some other things in addition to HQPlayer, so a bit more volume there. While Facebook group doesn't really have anything else (being dedicated to HQPlayer only).

 

25 minutes ago, joelha said:

Do your filters still operate, if selected, when there's no upsampling employed?

 

Some filters allow 1:1 rate conversions, which makes sense only for apodizing filters where you get the filter impulse response remade even without rate conversion. And naturally also some filters also allow down-conversion, so one can play for example DXD (352.8 kHz) content even if DAC doesn't support such rate.

 

[joelha]

Thanks for the very quick reply, Jussi. 

 

Joel

[Theobetley]

What is your twitter handle Jussi?

[Miska]

6 hours ago, Theobetley said:

What is your twitter handle Jussi?

 

https://twitter.com/signalyst

 

There are also links to both Twitter and Facebook on left pane of the HQPlayer Desktop product page.

[dean70]

12 hours ago, Miska said:

 

I've been trying to reproduce this, without success. I wonder what setting could be triggering this...

 

Are you using something like ASIO driver on Windows? It could be also ASIO driver not clearing it's buffers between stop and restart.

 

 

It does appear to be related to the ASIO driver. Choosing the lowest latency setting will minimize the duration of what is left over in the buffer (using Tascam US-366 device). I am sure this also occurs with my DAC in PCM mode (also ASIO driver) - will need to confirm

[craighartley]

16 hours ago, Miska said:

 

poly-sinc-short is trying to be as few taps as possible while still having high attenuation (same as poly-sinc) and not too slow roll-off. At the moment poly-sinc is between the -short and -xtr, but closer to -short. I've thought about poly-sinc-long, which would be otherwise same, but longer/steeper than regular poly-sinc. -xtr is not called -long because it changes some other parameters too, like stop-band attenuation. The new poly-sinc-ext2 is closer to -xtr but a bit shorter, while having otherwise completely different design.

 

So at the moment, poly-sinc could be closest to what you are after.

 

 

Miska,

So far I’ve always gone back to poly-sinc (I listen to a lot of Classical and well recorded acoustic jazz and I feed everything as DSD 512 via NAA to a T&A DAC8 DSD), but I might stick with the new poly-sinc-ext2. I can’t really get my head round the varying factors that might explain why I might slightly prefer this to plain poly-sinc. But I’d be really interested to try a poly-sinc-long filter!

 

How do you find the cf between poly-sinc and poly-sinc-ext2 when playing to the T &A DAC8 at DSD 512?

 

Craig

[nicoff]

On 9/15/2018 at 1:51 AM, Le Concombre Masqué said:

1 won't debate if "correct inefficiencies in my room" is the right wording but YES

2. 3. NO & NO but MAYBE 4. (I can't and don't do 512 but convolution takes its toll on computing resources). If you specify what you have at the moment maybe someone convolving and upsampling to 512 could benchmark 

 

Thank you for the reply.

[Allan F]

Jussi,

Could you please describe the "new control interface methods to switch between time display modes" introduced in the latest version of HQ Player.

[Miska]

36 minutes ago, Allan F said:

Could you please describe the "new control interface methods to switch between time display modes" introduced in the latest version of HQ Player.

 

It is switch between the time modes you have in HQPlayer Desktop main window. With values 0 = "Time", 1 = "Remain" and 2 = "Total Remain". So now this can be changed through hqp-control (--set-display and --get-display options).

 

[paulmckwan]

On 9/15/2018 at 11:59 PM, Miska said:

 

Especially in this kind of case you should set "DAC Bits" to 24 and use some dither like TPDF.

 

Truncation always causes distortion, while dithering doesn't. Dithering to 24-bit adds white noise at -144 dBFS level. So if you'd hear dither it would sound like tape hiss or like FM radio between stations. If dither interferes with something in upstream, there is something seriously wrong with the upstream component.

 

I also wonder what is the proper setting of bit depth when output to streamer through ethernet NAA, which is then output to DAC through USB or SPDIF.

Streamer / ethernet seems can handle 32/192 PCM, but SPDIF suppose is 24/192 max, and often we can't sure how streamer perform the 32 to 24 conversion. So if we use streamer's SPDIF output, shall we better set 24 bit in HQP, and leave it at 32 is fine for using streamer's USB output ?

 

I use SPDIF output from my streamer to Devialet, and in some tracks, I can hear minor difference between 24 and 32 bit settings. 

[Allan F]

5 hours ago, Miska said:

 

It is switch between the time modes you have in HQPlayer Desktop main window. With values 0 = "Time", 1 = "Remain" and 2 = "Total Remain". So now this can be changed through hqp-control (--set-display and --get-display options).

 

 

Thank you for the prompt reply, Jussi, but I am confused. Where do I set the value to  "1"  so that "Remain" becomes the default?  I tried changing the value of the field "period_time" at the bottom of the settings.xml file to period-time="1", but when I rebooted it did not change the time mode from the default. In an abundance of caution, not knowing what that field was for, I changed its value back to "0".

[Miska]

5 hours ago, Allan F said:

Thank you for the prompt reply, Jussi, but I am confused. Where do I set the value to  "1"  so that "Remain" becomes the default?  I tried changing the value of the field "period_time" at the bottom of the settings.xml file to period-time="1", but when I rebooted it did not change the time mode from the default. In an abundance of caution, not knowing what that field was for, I changed its value back to "0".

 

It currently always defaults to "Time" and this cannot be configured anywhere. But you can make a startup script that calls hqp-control after HQPlayer is started to change the mode.

 

I can look into storing this as part of the UI state store in future versions...

 

Edit: I first thought this was about HQPlayer Embedded, but then realized it was about Desktop...

 

[Miska]

5 hours ago, paulmckwan said:

I also wonder what is the proper setting of bit depth when output to streamer through ethernet NAA, which is then output to DAC through USB or SPDIF.

Streamer / ethernet seems can handle 32/192 PCM, but SPDIF suppose is 24/192 max, and often we can't sure how streamer perform the 32 to 24 conversion. So if we use streamer's SPDIF output, shall we better set 24 bit in HQP, and leave it at 32 is fine for using streamer's USB output ?

 

I use SPDIF output from my streamer to Devialet, and in some tracks, I can hear minor difference between 24 and 32 bit settings. 

 

It should be set to that of S/PDIF or the DAC behind it, which ever is lowest. Otherwise the lowest bits get just rudely truncated at some point without proper dither.

 

For example:

 - NAA with S/PDIF connected to 16-bit TDA1541A DAC: set DAC Bits to 16

 - NAA with S/PDIF connected to 24-bit PCM1704 DAC: set DAC Bits to 24

 - NAA connected through USB to Metrum Musette DAC: set DAC Bits to 16

 - NAA connected through USB to an iFi DAC: set DAC Bits to 24

 

For the two first cases, since S/PDIF is unidirectional, HQPlayer cannot detect what kind of resolution the DAC behind it has. Also some of the S/PDIF interfaces lie that they have 32-bit resolution while they actually don't.

 

For the third case, the USB interface lies that the DAC is 32-bit while it is in fact 16-bit. And for the fourth case, the USB interface lies that the DAC is 32-bit while it is in fact 24-bit. So you need to tell the reality to HQPlayer.

 

[Le Concombre Masqué]

21 hours ago, craighartley said:

 

Miska,

So far I’ve always gone back to poly-sinc (I listen to a lot of Classical and well recorded acoustic jazz and I feed everything as DSD 512 via NAA to a T&A DAC8 DSD), but I might stick with the new poly-sinc-ext2. I can’t really get my head round the varying factors that might explain why I might slightly prefer this to plain poly-sinc. But I’d be really interested to try a poly-sinc-long filter!

 

How do you find the cf between poly-sinc and poly-sinc-ext2 when playing to the T &A DAC8 at DSD 512?

 

Craig

cf?

I'll do the effort and take the risk to word an explanation:

poly sinc... used to always go back to it too... guess I did not think of it as fancy enough; thanks for reminding it : turned out to be the best filter for 24/96 Minimalist Dreams by Katia & Maria Labeque for which mqa was so so while being at least as satisfying as mqa for 24/96 Buana Vista for which mqa was wonderful.

 

Becomes my go to filter for 24/96

 

My criteria, maybe very personal, are density speed details, timbre, comfort ;

ie the ripped files from the SACD of So by Peter Gabriel satisfy me (ASDM5, CIC)

 

With 44.1, ext 2, to my ears, vs poly sinc, raises speed and details while lowering density to a more credible lightness ; through details it also raises specific sound signature ie of a piano ; leads me to think it raises timbral accuracy but it could be a fool's gold effect and, ie, Arrau's piano (Final sessions 3, Debussy) was on the verge of discomfort while I could hardly like any other filter for I got hooked to spatial and sound signature details via ext2.

 

ext 2 is nevertheless my goto filter for 44.1

 

Comparisons are complicated further by the choice of SPL. I measured no more than  89 C for the Arrau with an iPhone app that usually measures 91/92 where Iike to seat, a recent piano concert included. Nevertheless, lowering the SPL for the Arrau might not be a heresy.

 

I'm still not sure about mqa, if it's good for 192 or just too unpredictable in terms of timbral accuracy. Below 192, i's too dense and slow while often seducing in terms of comfort and details

 

 

 

 

 

[Le Concombre Masqué]

@Miska

I understand that filters are described in the manual and that it's great than everybody can pick one suiting his/her tastes/system/music etc

However could we see graphs, à la SRC or iZotope RX (delirious exemple joined), of what happens for a bunch of popular filters (sinc, sinc short, mqa, xtr, ext2, minringFIR, CF) facing 44 96 192 source files with 2 scenarios for each sample rate : "idea"l and "desperately in need for apodizing" ?

[craighartley]

29 minutes ago, Le Concombre Masqué said:

cf?

I'll do the effort and take the risk to word an explanation:

poly sinc... used to always go back to it too... guess I did not think of it as fancy enough; thanks for reminding it : turned out to be the best filter for 24/96 Minimalist Dreams by Katia & Maria Labeque for which mqa was so so while being at least as satisfying as mqa for 24/96 Buana Vista for which mqa was wonderful.

 

Becomes my go to filter for 24/96

 

My criteria, maybe very personal, are density speed details, timbre, comfort ;

ie the ripped files from the SACD of So by Peter Gabriel satisfy me (ASDM5, CIC)

 

With 44.1, ext 2, to my ears, vs poly sinc, raises speed and details while lowering density to a more credible lightness ; through details it also raises specific sound signature ie of a piano ; leads me to think it raises timbral accuracy but it could be a fool's gold effect and, ie, Arrau's piano (Final sessions 3, Debussy) was on the verge of discomfort while I could hardly like any other filter for I got hooked to spatial and sound signature details via ext2.

 

ext 2 is nevertheless my goto filter for 44.1

 

Comparisons are complicated further by the choice of SPL. I measured no more than  89 C for the Arrau with an iPhone app that usually measures 91/92 where Iike to seat, a recent piano concert included. Nevertheless, lowering the SPL for the Arrau might not be a heresy.

 

I'm still not sure about mqa, if it's good for 192 or just too unpredictable in terms of timbral accuracy. Below 192, i's too dense and slow while often seducing in terms of comfort and details

 

 

Sorry, 'cf' is standard English abbreviation for 'compare', or in this context 'comparison', but you've done that anyway in your reply, so thanks.

 

I should have said that I listen (mostly?) to high--res downloads (88.2 to 352.5), although I am also upsampling 44.1 (and low-res AAC radio streams). And I generally use the AMSDM& 512+fs modulator.

Quote

 

[Le Concombre Masqué]

4 hours ago, craighartley said:

 

Sorry, 'cf' is standard English abbreviation for 'compare', or in this context 'comparison', but you've done that anyway in your reply, so thanks.

 

I should have said that I listen (mostly?) to high--res downloads (88.2 to 352.5), although I am also upsampling 44.1 (and low-res AAC radio streams). And I generally use the AMSDM& 512+fs modulator.

strange, in French we have cf too, and coming from the same latin locution conferre but the meaning is more "go see and check"

 

poly sinc definitely makes sense with hires ; can't tell about ext2 and n48 hires

[tedwoods]

6 hours ago, Le Concombre Masqué said:

With 44.1, ext 2, to my ears, vs poly sinc, raises speed and details while lowering density to a more credible lightness ; through details it also raises specific sound signature ie of a piano ; leads me to think it raises timbral accuracy but it could be a fool's gold effect and, ie, Arrau's piano (Final sessions 3, Debussy) was on the verge of discomfort while I could hardly like any other filter for I got hooked to spatial and sound signature details via ext2.

 

ext 2 is nevertheless my goto filter for 44.1

You are spot on!

ext 2 really caught me off guard! 

The combination of the updated DSD5 modulator and ext 2, is the best sound I ever got!

Cheers, @Miska

[Allan F]

8 hours ago, Miska said:

It currently always defaults to "Time" and this cannot be configured anywhere. But you can make a startup script that calls hqp-control after HQPlayer is started to change the mode.

 

I can look into storing this as part of the UI state store in future versions...

 

Edit: I first thought this was about HQPlayer Embedded, but then realized it was about Desktop...

 

I am sorry, Jussi, but now I am even more confused. I must be missing something, but I don't understand what is new about this feature. I have always been able to choose between "Time", "Remain", and "Total Remain" in the upper right corner of the HQPlayer interface on my desktop.

 

When I read your initial reply, I mistakenly thought that you were referring to a new setting in Ales' HQPDcontrol app. I emailed him to ask him if he had issued a new version of his app to incorporate this new feature. Of course, he replied that he had not and would look into how to implement it.

 

What is hqp-control that you refer to? I see hqp-control.exe (189KB) in the program folder for HQPlayer, but I don't know what it does and clicking on it just causes a brief flash on the screen.