Jump to content
IGNORED

HQplayer 4 and Mac-mini M1


Bushikai

Recommended Posts

@Miska how far is too far ?

 

I'm thinking of :

 

Having a dedicated Desktop setting for DSD sources with 6 types of convolution .wav pairs (Katz, BK, Trained listeners etc) 44.1 based

+ Having a dedicated Desktop setting for PCM sources with my previous design (base .txt, VBA as .wav 48 based, tailored (or not) per disk my Cello+.txt)

 

or

 

Having a dedicated Desktop setting for DSD sources with 6 types of convolution .wav pairs (Katz, BK, Trained listeners etc) 44.1 based

+Having a dedicated Desktop setting for PCM sources with 6 types of convolution .wav pairs (Katz, BK, Trained listeners etc) 44.1 AND 48  based + a seventh entry, say with BK convolution, as .wav + my Cello+.txt whenever I have to correct the mastering, not just try to figure out the mastering monitors settings

 

or

 

Just having 6 types of convolution .wav pairs (Katz, BK, Trained listeners etc) 44.1 based (for the sake of minimising the load with DSD sources), thus even for PCM 48 based, + a seventh entry, say with BK convolution, as .wav + my Cello+.txt whenever I have to correct the mastering, not just try to figure out the mastering monitors settings

 

Link to comment
10 hours ago, Kalpesh said:

@Miska how far is too far ?

 

I'm thinking of :

 

You can have as many profiles as you like of course. But I don't think it is necessary to have 44.1k and 48k family settings. In optimal case you would have your convolution filters for 352.8k rate only.

 

At the time of processing, it doesn't matter much from convolution filter point of view, because the convolution filter is pre-processed to be suitable for the source content before any playback begins.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
1 hour ago, Miska said:

 

You can have as many profiles as you like of course. But I don't think it is necessary to have 44.1k and 48k family settings. In optimal case you would have your convolution filters for 352.8k rate only.

 

At the time of processing, it doesn't matter much from convolution filter point of view, because the convolution filter is pre-processed to be suitable for the source content before any playback begins.

 

Thank you.

352.8 passed the test (64 to 512 SDM) ; thus I will create a unique set of convolution filters at 352.8 for any source.

 

As of what happens at time of processing, having a single .wav (per channel) made the huge difference of 512 OK vs only 128 OK with a mix of .wav and .txt ; so I'm not sure I ever correctly understood/interpreted  "the convolution filter is pre-processed to be suitable for the source content before any playback begins" !

 

Link to comment
2 hours ago, Miska said:

 

You can have as many profiles as you like of course. But I don't think it is necessary to have 44.1k and 48k family settings. In optimal case you would have your convolution filters for 352.8k rate only.

 

At the time of processing, it doesn't matter much from convolution filter point of view, because the convolution filter is pre-processed to be suitable for the source content before any playback begins.

 

@Miska can you please explain Is there any advantage for using 352.8k vs 384k? Also does it mean I don’t need to enable “hf expansion” at either sampling rate?

 

thanks

 

Deric

Link to comment
2 hours ago, Kalpesh said:

As of what happens at time of processing, having a single .wav (per channel) made the huge difference of 512 OK vs only 128 OK with a mix of .wav and .txt ; so I'm not sure I ever correctly understood/interpreted  "the convolution filter is pre-processed to be suitable for the source content before any playback begins" !

 

.txt, or other parametric EQ are processed as IIR biquads. Not as convolution filter. You can combine both convolution filters and those parametric EQs, but they are processed separately. Only wav are convolution filters. Every other plugin is something different (wav goes to convolution plugin and txt goes to iir plugin).

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
7 hours ago, Miska said:

 

AFAIK, nobody is selling 384k PCM, but 352.8k PCM is common (DXD). 352.8k is also integer multiple of DSD content available. And it is also integer multiple of 44.1k RedBook which is the most common format. It's response also covers enough audio bandwidth (176.4 kHz) for any imaginable music recording, so you don't need HF Expansion.

 

otoh, 24/96 is the trending format (for what I listen to at least), 48 the base format in REW : here are 2 reasons to go 48 based but if it doesn't affect SQ or create undue load, let's go 352.8 !

Link to comment
12 hours ago, Miska said:

 

.txt, or other parametric EQ are processed as IIR biquads. Not as convolution filter. You can combine both convolution filters and those parametric EQs, but they are processed separately. Only wav are convolution filters. Every other plugin is something different (wav goes to convolution plugin and txt goes to iir plugin).

 

Jussi, I have a tremendous respect for you and I have purchased  4 or 5 HQP licences so far so please allow me to express a serious disappointment :

 

I have posted numerous claims that my performance with DSD sources was really bad

I have explained how I do DRC, I'm pretty sure I can dig out at least one of your posts where you acknowledge and approve my use of .txt 

I have lost hours looking for a solution, including searching for a beefier Mac in my budget

And at the present time I type while my MMM1 plays DSD 64 source +convolution @ 256 (+ in total wifi mode since the ethernet cables are now plugged in the Studio) and my predicament is 

do I keep the Studio ? do I cancel the sale of my MMM1 that I just sold after lowering the price to 390 € ?

 

I purchased the Studio because I was limited to ASDM7 @ 64 with DSD sources ; I would have never bothered knowing my MMM1 plays DSD 64 +convolution @ 256, that all I had to do was use .wav convolution instead of .txt based correction.

 

+ I have little dropouts @512 with the Studio since I installed Roon, even with Roon closed and after reboot...

Link to comment
1 hour ago, Kalpesh said:

I have explained how I do DRC, I'm pretty sure I can dig out at least one of your posts where you acknowledge and approve my use of .txt 

 

Yes, it is fine.

 

1 hour ago, Kalpesh said:

 do I keep the Studio ? do I cancel the sale of my MMM1 that I just sold after lowering the price to 390 € ?

 

I don't know, only you can test your setup and see what can possibly do what you are looking for. And only you can decide what kind of hardware you want to use.

 

If there's no Mac that can do what you want, there may be a PC setup with GPU that can do it.

 

1 hour ago, Kalpesh said:

+ I have little dropouts @512 with the Studio since I installed Roon, even with Roon closed and after reboot...

 

I always run Roon on my ancient i5 NUC, under Debian Linux. Nothing else on that machine. But I rarely use it. Only if I happen to want to listen something from Tidal.

 

Quote

please allow me to express a serious disappointment

 

Hardware manufacturers are correct address for that. I don't sell hardware...

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
6 hours ago, Kalpesh said:

otoh, 24/96 is the trending format (for what I listen to at least), 48 the base format in REW : here are 2 reasons to go 48 based but if it doesn't affect SQ or create undue load, let's go 352.8 !

 

HQPlayer will adapt the convolution filters to what ever sampling rate the source content is. 384k as filter rate doesn't make sense, since it doesn't match any (?) commercially available content.

 

With REW, you don't need to unnecessarily convert the sampling rate agnostic parametric EQ's to sampling rate bound convolution filters. So with REW I always just export the filters natively as .txt (instead of converting to .wav).

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
21 minutes ago, Miska said:

 

Yes, it is fine.

 

 

I don't know, only you can test your setup and see what can possibly do what you are looking for. And only you can decide what kind of hardware you want to use.

 

If there's no Mac that can do what you want, there may be a PC setup with GPU that can do it.

 

 

I always run Roon on my ancient i5 NUC, under Debian Linux. Nothing else on that machine. But I rarely use it. Only if I happen to want to listen something from Tidal.

 

 

Hardware manufacturers are correct address for that. I don't sell hardware...

 

I have not been straight enough : your assistance I usually praise failed here. You were the right person to point out that being limited to ASDM 7 64 output only from 64 source with a MMM1  is seriously wrong and help me sort out what the issue was. Especially since I posted a lot about my way to do correction and my performance issues. Now we know it was because I was using convolution plugin for my wav file + iir plugin for the .txt, that was the cause of too high a burden, it's a HQPlayer usage cause, not something to be dealt on the hardware side. 

Link to comment
13 minutes ago, Miska said:

 

 

With REW, you don't need to unnecessarily convert the sampling rate agnostic parametric EQ's to sampling rate bound convolution filters. So with REW I always just export the filters natively as .txt (instead of converting to .wav).

 

That's the advice I followed but with the combination of a .wav for VBA. 

Link to comment
15 minutes ago, Miska said:

 

It is perfectly fine. It is precisely something to be dealt with hardware side. You just need fast enough computer. Or you need to lower/reduce what you are asking your computer to compute.

 

"Too high burden" just means you are asking too much from your hardware, it doesn't mean it would be generally the case. If you have limited computing resources, you need to adapt your computing request to something it can handle. Or you need more computing resources.

 

You could have more success with latest top of the line Intel or AMD CPUs. Mac CPUs are per core more than 50% shy of those...

 

Your answers are really disappointing...

I have now (or before Roon) reached : 7EC 512 output including with DSD 64 source on Mac platform ; that's good enough for me, I'm not interested in Intel or AMD

I could have enjoyed (and be happy and not bother upgrading) 7EC 256 output including with DSD 64 source with MMM1 since Light variant inception, provided you gave in time the assistance you gave me with the Studio

Link to comment
23 minutes ago, Kalpesh said:

I have now (or before Roon) reached : 7EC 512 output including with DSD 64 source on Mac platform ; that's good enough for me, I'm not interested in Intel or AMD

I could have enjoyed (and be happy and not bother upgrading) 7EC 256 output including with DSD 64 source with MMM1 since Light variant inception, provided you gave in time the assistance you gave me with the Studio

 

I don't understand what you are disappointed at? I have only ever told you what works for me.

 

And you have been all the time trying to run something else than what I have said works for me. Then you figured out yourself that you are asking too much from the Studio... Rest is your decision what you are happy with.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
47 minutes ago, Miska said:

 

I don't understand what you are disappointed at? I have only ever told you what works for me.

 

And you have been all the time trying to run something else than what I have said works for me. Then you figured out yourself that you are asking too much from the Studio... Rest is your decision what you are happy with.

 

I'm disappointed you did not bother help me find out why I was stuck with ASDM7 @ 64 output from 64 source with a MMM1 while I can now, simply with a new scheme for room correction, reach 7EC Light @ 256.

What I'm asking from the Studio is not the point ; the point is that you did not bother helping me get what I could get from a MMM1

can't you just say, sorry, I did not figure out that using 2 plugins instead of one for eQ could be the cause ?

Link to comment

😳🙄

____________________________________________________

Mac Mini, HQPlayer | iFi Zenstream (NAA) | Intona 7055-B | Singxer SDA-6 pro | Vincent SV237 | Buchardt S400 | SPL Phonitor One | Beyer DT1990pro | Avantone Pro Planar II
Desktop: Audirvana Origin | Intona 7054 | SMSL M500MKII | Pro-Ject Stereo Box S | Aperion Novus B5 Bookshelf | Lehmann Rhinelander | Beyer DT700proX

Link to comment

To reset the mood positively :

 

To my ears, Studio vs MMM1 8 Go RAM is worth it,

the sound is definitely classier, once you have dug it, there's no coming back (because of 512 vs 256 or less memory swapping, power supply ???)

Listening to Qobuz is a much much smoother experience

 

The culprit for interrupts was not Roon but an aborted installation of Rogue Amoeba's Audio Hijack. I aborted when asked to circumvent security to install their Audio Content Extractor in Kernel : not worth it IMO. Nevertheless I had to follow the manual uninstall procedure for removing ACE driver and Launch deamons. I take the time to explain for I had Audio Hijack installed on the MMM1 (back then I did not receive as terrifying messages from the system nor followed the physical restart procedure now required). So I can't rule out that some limitations and dropouts with the MMM1 was sometimes due to A H's Audio Content Extractor. 

 

Link to comment
5 hours ago, Kalpesh said:

What I'm asking from the Studio is not the point ; the point is that you did not bother helping me get what I could get from a MMM1

 

How!?

 

In addition, just for fun, imagine yourself trying to answer about 100 questions per day, do software development, and pretty much everything else involved.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
49 minutes ago, Miska said:

 

How!?

 

 

In addition, just for fun, imagine yourself trying to answer about 100 questions per day, do software development, and pretty much everything else involved.

How!?

That's called Support

 

And maybe you should do yourself a favour and organise a wiki, a FAQ, what have you, for the redundant questions you answer to zillions times, and create tickets/flag for new issues coming and then nurture the wiki, FAQ, what have you,

Link to comment
18 minutes ago, Kalpesh said:

How!?

That's called Support

 

So I would travel to your house, and we would try together with your settings to see if some setting combinations work or don't. So I would be just holding your hand. But for that, I would need to increase price of HQPlayer by 100x at least.

 

18 minutes ago, Kalpesh said:

And maybe you should do yourself a favour and organise a wiki, a FAQ, what have you, for the redundant questions you answer to zillions times, and create tickets/flag for new issues coming and then nurture the wiki, FAQ, what have you,

 

You could consider this forum as such?

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...