HQ Player

[Yviena]

4 hours ago, Miska said:

 

 

I'm thinking about way to let you tell which cores you want to get allocated. Extending the environment variable is easy, doing a nice GUI for it is not so straightforward.

 

On another note it seems that i need to use the environment variable to disable corepinning in the latest version, or HQplayer will crash if i'm doing manual assigment, when stopping starting songs, it quickly displays a cmd console before closing.

[Ernie]

I am wondering what is the best way to implement the real time audio input function of HQPlayer to feed the new amazon hdmusic service directly into it. Roon is not an option for now.

[pcmchild]

I have all my music in redbook PCM 44.1K WAV files.

My goal is to convert them all to 705K or higher (44.1 multiple) PCM WAV files.

WIll a future HQP version be able to do the conversion and play the upscaled files? Sorry, if it has been discussed already, I could not find it.

I read HQP Pro does offline conversion, but it is overkill with so many features, and the price reflects that. I don't need much options, really just an open file and upscale button (with multiple parameter 4x, 8x, 16x, 32x,...).

Offline would allow ideal conversion without having to concede on time optimizations.

[Miska]

14 minutes ago, Ernie said:

I am wondering what is the best way to implement the real time audio input function of HQPlayer to feed the new amazon hdmusic service directly into it. Roon is not an option for now.

 

I guess output from the desktop app is most suitable at the moment, either through loopback device or hardware. I was yesterday wondering if Bluesound Node2i would work, and doesn't seem to work yet:

https://support1.bluesound.com/hc/en-us/community/posts/360035058834-Amazon-Music-HD

 

I have AES output from MacMini, connected to my HQPlayer Embedded server, that is most generic way for these kind of things so far.

 

[asdf1000]

3 minutes ago, Miska said:

I guess output from the desktop app is most suitable at the moment, either through loopback device or hardware. I was yesterday wondering if Bluesound Node2i would work, and doesn't seem to work yet:

https://support1.bluesound.com/hc/en-us/community/posts/360035058834-Amazon-Music-HD

 

@The Computer Audiophile had it working via an update today:

 

https://audiophilestyle.com/forums/topic/57226-article-amazon-music-hd-launches/?do=findComment&comment=990930

 

So should open up the opportunity of Node 2i and miniDSP USBStreamer ?

 

 

 

[Miska]

6 minutes ago, pcmchild said:

I have all my music in redbook PCM 44.1K WAV files.

My goal is to convert them all to 705K or higher (44.1 multiple) PCM WAV files.

WIll a future HQP version be able to do the conversion and play the upscaled files? Sorry, if it has been discussed already, I could not find it.

I read HQP Pro does offline conversion, but it is overkill with so many features, and the price reflects that. I don't need much options, really just an open file and upscale button (with multiple parameter 4x, 8x, 16x, 32x,...).

Offline would allow ideal conversion without having to concede on time optimizations.

 

Not with file output, but it of course can do conversion on the fly at playback time, no need to store conversion result in a file. Especially because such conversion is very light weight so can be done even on a less powerful computer.

 

[Miska]

3 minutes ago, asdf1000 said:

https://audiophilestyle.com/forums/topic/57226-article-amazon-music-hd-launches/?do=findComment&comment=990930

 

Interesting, because their response is just 9 hours old... I understood it displays the content, but cannot play...

 

Amazon Echo Link seems to be limited to max 96k.

 

6 minutes ago, asdf1000 said:

So should open up the opportunity of Node 2i and miniDSP USBStreamer ?

 

Yes, lack of automatic input rate switching may be a bit annoyance for hires content. On Linux, the RME ADI-2 Pro is so far most flexible on input front because automatic rate switching works.

 

I have not yet checked if RME ASIO driver supports rate change reporting properly. If it does, that would open the functionality on Windows too.

 

[pcmchild]

32 minutes ago, Miska said:

 

Not with file output, but it of course can do conversion on the fly at playback time, no need to store conversion result in a file. Especially because such conversion is very light weight so can be done even on a less powerful computer.

 

Thank you very much for the reply. But light weight? Rob Watts says the mscaler function can't be done real time (not even near) in a PC...

Although I have my doubts on that, I am surprised you say it is light weight.

Also I thought HQP was doing some optimizations (deviation from ideal function in order to be time efficient). The mscaler does optimizations (deviations from ideal). Offline can avoid any non-ideal optimizations. But if there are none...

Of course one thing is the number os samples we use for the sync function, the more samples the higher the output resolution (real bits per sample), I think. Which also causes a few seconds delay with mscaler.

 

[Miska]

15 minutes ago, pcmchild said:

Thank you very much for the reply. But light weight? Rob Watts says the mscaler function can't be done real time (not even near) in a PC...

 

That's just so immense bullshit... MScaler upsamples to just 16x, here we are usually talking about upsampling to 256x or 512x and in addition running complex modulators and maybe digital room correction while doing so.

 

15 minutes ago, pcmchild said:

Also I thought HQP was doing some optimizations (deviation from ideal function in order to be time efficient).

 

My optimizations are about how to best utilize hardware features, such as AVX vector units inside CPU, to achieve maximum processing speed for the same algorithm. Never trading any quality for optimizations, only finding fastest way to do the same thing. HQPlayer has much better precision than MScaler.

 

My approach is to NOT optimize by trading quality. Instead if your computer is too slow to run the algorithms, you need to buy faster one.

 

15 minutes ago, pcmchild said:

Offline can avoid any non-ideal optimizations.

 

Since there are no non-ideal optimizations, HQPlayer 4 Pro has exactly same algorithms and precision as HQPlayer 4 Desktop and HQPlayer 4 Embedded.

 

[pcmchild]

Thank you again for the explanation. I am trying to understand as I read conflicting statements. So thank you for your patience.

For example I thought perfect interpolation requires an infinite length (time) sinc, so infinite processing and infinite delay. Of course that is not possible, but doesn't a higher number of samples used on the sync, increase the output resolution (real bits of each sample)? If that was the case, more processing would result in better quality. Or is it not noticeable?

[Ernie]

7 hours ago, Miska said:

 

I guess output from the desktop app is most suitable at the moment, either through loopback device or hardware. I was yesterday wondering if Bluesound Node2i would work, and doesn't seem to work yet:

https://support1.bluesound.com/hc/en-us/community/posts/360035058834-Amazon-Music-HD

 

I have AES output from MacMini, connected to my HQPlayer Embedded server, that is most generic way for these kind of things so far.

 

 

Thanks Miska..I might try loopback. But your note makes me think I might be able to run an optical out from my mac powerbook to the imac that is running HQP

[maya]

On 9/15/2019 at 7:48 PM, Miska said:

 

Have you tried running OS from Intel Optane M.2? That is as optimal as it can get.

 

All my Debian, Fedora and Ubuntu Server installations are on either SATA SSD or M.2 SSD. Especially with M.2 I like that HQPlayer Embedded is up and running in ~5 seconds after pressing power button. And shuts down in one second after pressing power button.

 

We use msata ssd all along, fast boot and shutdown.

We have a custom made heavy aluminum chassis to accommodating both motherboards  for HQP and NAA, seems to be the first one to make this!

We found HQPe ram boot from Extreme Pro USB stick although it  needs less than one minute, much longer than msata SSD after removing USB stick the background becoming so dark and much more musical which we have never experienced with HQPlayer ! 

Actually the operating speed is not the thing we want most. 😊 

[Miska]

13 hours ago, pcmchild said:

For example I thought perfect interpolation requires an infinite length (time) sinc, so infinite processing and infinite delay. Of course that is not possible, but doesn't a higher number of samples used on the sync, increase the output resolution (real bits of each sample)? If that was the case, more processing would result in better quality. Or is it not noticeable?

 

It is not quite that simple. Theoretically yes, if you want to reconstruct up to exactly Nyquist (fs/2) freqency and you are also OK with infinitely bad time domain response.

 

Practically, if we begin with source content, it is certainly not relevant and proper up to Nyquist. In fact, depending on ADC/mastering tools, there can be quite a bit of aliasing band at the top of the frequency band. You wouldn't want to reproduce that distortion part. That's why we have apodizing upsampling filters, for example in HQPlayer, to clean it up. In addition, even with RedBook, there is some amount of margin between wanted frequency band (let's assume 20 kHz here) and Nyquist frequency (22.05 kHz). So we have 2.05 kHz wide transition band available between wanted frequency band and Nyquist.

 

So in first place you need to consider what was used to produce the content you are trying to reconstruct, because the source content is certainly far from perfect. So instead of assuming perfect content up to Nyquist, you need to consider imperfect content up to Nyquist and what you can make out of that.

 

In HQPlayer you can run 16 million tap filter up to 2048x rates, if you want to.

 

[Miska]

4 hours ago, maya said:

We found HQPe ram boot from Extreme Pro USB stick although it  needs less than one minute, much longer than msata SSD after removing USB stick the background becoming so dark and much more musical which we have never experienced with HQPlayer ! 

Actually the operating speed is not the thing we want most. 😊 

 

Optane gives you persistent RAM-like SSD... That's the whole idea of Optane.

 

[Miska]

7 hours ago, Ernie said:

Thanks Miska..I might try loopback. But your note makes me think I might be able to run an optical out from my mac powerbook to the imac that is running HQP

 

That could also work. Down side is that you need to switch sampling rate manually when it changes. But it is not necessarily too bad, since most content is 44.1k anyway.

 

[Yviena]

49 minutes ago, Miska said:

 

It is not quite that simple. Theoretically yes, if you want to reconstruct up to exactly Nyquist (fs/2) freqency and you are also OK with infinitely bad time domain response.

 

 

Isn't the pre-ringing from long-filters which is said to affect time domain response also at ultra-sonic frequencies which we can't hear anyway, so how can it affect how we hear.

[pcmchild]

2 hours ago, Miska said:

 

It is not quite that simple. Theoretically yes, if you want to reconstruct up to exactly Nyquist (fs/2) freqency and you are also OK with infinitely bad time domain response.

 

Practically, if we begin with source content, it is certainly not relevant and proper up to Nyquist. In fact, depending on ADC/mastering tools, there can be quite a bit of aliasing band at the top of the frequency band. You wouldn't want to reproduce that distortion part. That's why we have apodizing upsampling filters, for example in HQPlayer, to clean it up. In addition, even with RedBook, there is some amount of margin between wanted frequency band (let's assume 20 kHz here) and Nyquist frequency (22.05 kHz). So we have 2.05 kHz wide transition band available between wanted frequency band and Nyquist.

 

So in first place you need to consider what was used to produce the content you are trying to reconstruct, because the source content is certainly far from perfect. So instead of assuming perfect content up to Nyquist, you need to consider imperfect content up to Nyquist and what you can make out of that.

 

In HQPlayer you can run 16 million tap filter up to 2048x rates, if you want to.

 

Thank you very much for the detailed reply, I understand now.

[Miska]

11 hours ago, Yviena said:

Isn't the pre-ringing from long-filters which is said to affect time domain response also at ultra-sonic frequencies which we can't hear anyway, so how can it affect how we hear.

 

For RedBook content, right at the edge between audible frequencies and ultrasonic frequencies, at about 22.05 kHz. Which some people in fact can directly hear.

 

But there are multiple aspects related to this. Spread over time ("blur"), transition/rise time, timing accuracy, spectral correctness of transient, ringing, and imaging/aliasing. Many of these have inverse/conflicting relationship, so biggest challenge is to make something that works well in all aspects simultaneously. This is what I prefer, although I provide filters with different weightings since different people are sensitive to different aspects of the sound.

 

One example of complete opposite extremes in this respect are MQA's upsampling filters and Chord's upsampling filters.

 

[Yviena]

2 hours ago, Miska said:

 

For RedBook content, right at the edge between audible frequencies and ultrasonic frequencies, at about 22.05 kHz. Which some people in fact can directly hear.

 

But there are multiple aspects related to this. Spread over time ("blur"), transition/rise time, timing accuracy, spectral correctness of transient, ringing, and imaging/aliasing. Many of these have inverse/conflicting relationship, so biggest challenge is to make something that works well in all aspects simultaneously. This is what I prefer, although I provide filters with different weightings since different people are sensitive to different aspects of the sound.

 

One example of complete opposite extremes in this respect are MQA's upsampling filters and Chord's upsampling filters.

 

I see, so source sample rates of 48, 88.2/96khz, or newer ADC that correctly bandwidth limit (if it exists) will not have this problem correct?

 

Also in your opinion what filter would you say is the most balanced regarding what you mentioned above?

[Miska]

39 minutes ago, Yviena said:

I see, so source sample rates of 48, 88.2/96khz, or newer ADC that correctly bandwidth limit (if it exists) will not have this problem correct?

 

48k sampling rate still has the ringing at 24 kHz which is too close. When you move to hires, like 88.2/96k sampling rate, ringing moves to 44.1/48 kHz frequency. Other aspects still matter, but less. Note that decimation filter problems that get embedded into the source content disappear when the filter has nothing to remove. So if you use sampling rate of 176.4k or higher, you likely get rid of most problems. And using 352.8 kHz rate, even with best wide band microphones, properly steep decimation filter doesn't touch any actual content at all. With DSD, this problem doesn't exist at all, since only very gentle analog antialias filter is needed, if even that.

 

With commonly used DPA and Sennheiser microphones and classical music, musical content can be detected up to at least 65 kHz. With jazz, rock and blues type of music it should go higher, but finding such content for analysis that would have been recorded with wide band microphones at >= 176.4k sampling rate is not so easy. (most newly recorded true hires is classical)

 

[Lio_B]

I have a question about HQPDesktop 3 and USB output priority:

 

I have a mac mini, connected via USB to my integrated ampli (with DAC inside) Roon HQP3, all works fine.

 

My HQP settings are:

 

 

 

Or:

 

But when I want to hear an internet sound or VLC sound, I must shut of HQP, it's impossible to give priority to an over media.

 

Even if I choose my USB on VLC setting.

 

My mac audio settings: I choose my USB for sound output, but when Roon's running it move to an over output in settings !

 

 

Is there a solution without shutting HQP ?

 

Thanks

 

[Account Closed]

Jussi,

 

I see you have new filters in the latest Pro release. Are they coming to Desktop.

 

Thanks,

 

Bob

[Miska]

3 hours ago, bobflood said:

I see you have new filters in the latest Pro release. Are they coming to Desktop.

 

Yes, next release of Desktop and Embedded!

 

[Miska]

3 hours ago, Lio_B said:

But when I want to hear an internet sound or VLC sound, I must shut of HQP, it's impossible to give priority to an over media.

 

This is intended. Usually one wouldn't want to bring in the OS mixer and sample rate converter into the picture.

 

Idea is that you have some other audio device set as OS default audio output for unimportant sounds like YouTube and such. And then a DAC specifically for music that is not the OS default audio device. IOW, have HQPlayer have exclusive control over the audio output device when used. If you want to release the audio output device, just close HQPlayer with Cmd-Q first. It takes only couple of seconds to start again. But you definitely don't want to involve the OS mixer/rate converter for high quality audio.

 

[Account Closed]

1 hour ago, Miska said:

 

Yes, next release of Desktop and Embedded!

 

Thanks