Jump to content
IGNORED

HQ Player


Recommended Posts

I believe that no functional differences exist between HQPlayer Desktop and Embedded, except the latter currently runs on Unix/Linux only. Can someone clarify this?

 

From playback point of view there is no difference. The main difference is that Embedded doesn't have any GUI but instead it has a standard player control interface for other applications to use (different from the one offered by Desktop and used by Roon/Muso).

 

Also, will HQPlayer Embedded ever be available for Windows?

 

No Windows and no Mac at the moment, neither of the two are really embedded platforms without GUI and neither has the required infra for running Embedded version. I don't say never, but very unlikely to happen in it's current incarnation.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
This is a really complex question, because there are so many variables, and I add new optimizations over time and things like CUDA change the overall structure significantly. In the end, to say X can do Y, I would need to buy X and test how it performs Y in practice. I have some rough hunch though, but there are still many variables, and really hard to say how it compares to Z.

 

If you don't have CUDA, in many cases at the moment, optimal number of cores is roughly N*2+2 where "N" is number of audio channels and then pick as fast clock speed, memory bus bandwidth and large cache as possible. If you have CUDA, you can do with N+2 cores and maybe get higher clock speed. Since there are not many 6-core CPUs, for stereo case you may need to go with either 4 or 8 cores; so if you use Roon, go for 8 cores, if not you may get quite a bit more clock speed at quad core (i7-6700K runs at 4 GHz base). With something like i7-5960X or Xeon E5v3 you get twice as wide RAM bus though, doubling the RAM access bandwidth plus large caches.

 

Where the bottleneck is largely depends on the particular HQPlayer settings.

 

This is probably pretty good price/performance ratio in addition to i7-6700K:

Intel® Xeon® Processor E5-1650 v3 (15M Cache, 3.50 GHz) Specifications

 

Or fast dual-socket options (but it's not cheap anymore by any means):

Intel® Xeon® Processor E5-2637 v3 (15M Cache, 3.50 GHz) Specifications

 

For Mac users, Pro is also available as 6-core.

 

16-core machine is probably good option for example with exaSound e28. But I'll probably add more parallelism to the code over time, already did it for CUDA so I can try to scale it also to multi-core CPUs with lot of cores. Especially since there are rumors that AMD is coming up with some 32-core CPU.

 

All this assumes you don't use matrix processing, digital room correction or such, where higher number of cores help more.

 

For those of you in the US, I happened to be in Costco today and noticed that they have a Dell XPS8900 (Service Tag 3GNZZ72) with the i7-6700K chip with 16gb of RAM and a GTX 750 4GB CUDA enabled card for $799 US. That's with a $200 manufacturer's rebate. Seems like a good deal...and its only available in the store. I couldn't tell from the box how big the power supply is...and couldn't get the case open (screwed shut to keep people like me from opening it)....so I'm not sure if you could easily put a GTX970 in it or not.

 

Robert

Software: Roon/HQplayer; System I: Roon Server/HQplayer DSD 512 Upsampling, Custom Windows 10 PC/AO, LPS-1 powered Startech USB card; LPS-1 powered ISO Regen; Holo Cyan DAC; VPI Scout 2 Turntable, Soundsmith Boheme, TTW Clamps and Carbon Matt; Cary SLP-98P Preamp; Van Alstine FET 600 Poweramp;  Aerial Acoustics 6T loudspeakers, SVS SB13 Ultra Subwoofers. System II: Custom PC with Signalyst Linux HQplayer NAA; LPS-1 powered Startech USB card; LPS-1 powered ISO Regen; IFI Micro iDSD Black Label; Primaluna Dialogue 2 with Tung Sol KT-150; Paradigm Studio 20 v3 monitors on Custom Mapleshade stands. Cables: Moon Audio, LUSH, Kimber Kable, Mapleshade, LARRY custom.

Link to comment
Thank you very much Zack, that did the trick!

Ok, I know I'm late at the party, but I have to say: HQPlayer+NAA is a GAME CHANGER!!! All one needs is a good server and then something like the microRendu with the NAA/Roon output and it's game on!!!! This is exciting!!!

 

 

Sent from my iPhone using Computer Audiophile mobile app

 

Oh, good! I was using an old laptop for the NAA, with a lot of tweaks to Windows 8.1. Switching to a Cubox and Miska's SD image made things a lot cleaner and quieter. Is the microRendu available yet? I'd like to try that.

 

It's very exciting and I'm glad it's working for you.

Link to comment
Oh, good! I was using an old laptop for the NAA, with a lot of tweaks to Windows 8.1. Switching to a Cubox and Miska's SD image made things a lot cleaner and quieter. Is the microRendu available yet? I'd like to try that.

 

It's very exciting and I'm glad it's working for you.

 

Not yet available, but soon I'm guessing. Keep an eye on their thread and website.

 

 

Sent from my iPhone using Computer Audiophile mobile app

Link to comment

I've been using HQPlayer with Ubuntu Studio for some time and been very pleased with the performance. But on a whim I decided to try Debian stretch OS this week. I also upgraded HQPlayer from 3.12 to 3.13 at the same time which is a bit silly - its hard to pin down whether changes are due to new OS, or upgraded app. Anyway, after a couple of days listening I am certain that my computer audio system sounds better than ever. The main improvement is in the sound stage. Instruments have a very clear position in space and are totally free of the speaker. But any description of SQ changes are subjective, the objective observation that I have made relates to CPU usage. Previously (Ubuntu Studio, HQP 13.12) when playing hi res file, CPU usage was around 46%. Now (Debian, HQP 13.13) CPU usage is 38% with same settings.

 

I use a JLSounds USBtoI2S card, and HQPlayer upsamples everything to DSD256. Filer is poly sinc mp 2s, modulator DSD7 256+fs.

 

I like the automatic rate selection. I always had to switch the sample rate manually when moving from 48k family (12.3Mhz) to 44.1k family (11.3Mhz), and this is a nice touch.

 

Thanks Jussi, I love HQPlayer.

 

Regards, Hazard

Link to comment

I just purchased an Ifi idsd Dac and will be using it with HQ Player and Mac Mini, i5 2.6 with 8gb or ram. I want to start converting PCM to DSD. Any advice on which filters are usable with this processor? Any to avoid? Also will this processor handle up sampling to 256 or do I have to stay at 128?

 

Thanks,

TC

Link to comment
I just purchased an Ifi idsd Dac and will be using it with HQ Player and Mac Mini, i5 2.6 with 8gb or ram. I want to start converting PCM to DSD. Any advice on which filters are usable with this processor? Any to avoid? Also will this processor handle up sampling to 256 or do I have to stay at 128?

 

Hi TC

 

You didn't say if this is a Nano or a Micro.

 

To get DSD256 working with one of these DACs on a Mac, you'll need to apply a firmware update.

 

Nano firmware here

Micro firmware here

 

Don't worry, iFi has Mac versions of the firmware updaters in addition to Windows.

Link to comment
I've been using HQPlayer with Ubuntu Studio for some time and been very pleased with the performance. But on a whim I decided to try Debian stretch OS this week. I also upgraded HQPlayer from 3.12 to 3.13 at the same time which is a bit silly - its hard to pin down whether changes are due to new OS, or upgraded app. Anyway, after a couple of days listening I am certain that my computer audio system sounds better than ever. The main improvement is in the sound stage. Instruments have a very clear position in space and are totally free of the speaker. But any description of SQ changes are subjective, the objective observation that I have made relates to CPU usage. Previously (Ubuntu Studio, HQP 13.12) when playing hi res file, CPU usage was around 46%. Now (Debian, HQP 13.13) CPU usage is 38% with same settings.

 

I use a JLSounds USBtoI2S card, and HQPlayer upsamples everything to DSD256. Filer is poly sinc mp 2s, modulator DSD7 256+fs.

 

I like the automatic rate selection. I always had to switch the sample rate manually when moving from 48k family (12.3Mhz) to 44.1k family (11.3Mhz), and this is a nice touch.

 

Thanks Jussi, I love HQPlayer.

 

Regards, Hazard

 

Can I just ask concerning the JLSounds card, do I need to do any soldering or anything else, or does it work out of the box? The reason I am asking is that it sounds great, but I am not an engineer if it's meant for DIY builds...

 

 

Sent from my iPhone using Computer Audiophile mobile app

Link to comment

I like poly-sinc with DSD7 or ASDM7. But with you processor you will definitely want to use the poly-sinc 2s to reduce processing demand. I'm guessing you will be right on the edge of being able to do DSD256. I used to use a 2.2gHz i7 Mac Mini and it was hit and miss. From redbook it might work...if not go to DSD128. If you're upsampling redbook also give the closed-form filters a try.

 

On the mac I'm pretty sure you will need to use DoP. Its been a while since I used the IFI connected directly to a mac...I use a linux NAA in between them usually...I'm I'm pretty sure it won't go native...but because it passes PCM 702 you are good up to DSD256.

 

Robert

Software: Roon/HQplayer; System I: Roon Server/HQplayer DSD 512 Upsampling, Custom Windows 10 PC/AO, LPS-1 powered Startech USB card; LPS-1 powered ISO Regen; Holo Cyan DAC; VPI Scout 2 Turntable, Soundsmith Boheme, TTW Clamps and Carbon Matt; Cary SLP-98P Preamp; Van Alstine FET 600 Poweramp;  Aerial Acoustics 6T loudspeakers, SVS SB13 Ultra Subwoofers. System II: Custom PC with Signalyst Linux HQplayer NAA; LPS-1 powered Startech USB card; LPS-1 powered ISO Regen; IFI Micro iDSD Black Label; Primaluna Dialogue 2 with Tung Sol KT-150; Paradigm Studio 20 v3 monitors on Custom Mapleshade stands. Cables: Moon Audio, LUSH, Kimber Kable, Mapleshade, LARRY custom.

Link to comment

Jussi,

 

I'd like to request a small feature in a future version.

 

Specifically it would be a check box in the DSDIFF/DSF Settings menu that requests that the program check the DSD file to be played and if it is the same as the max bitrate level currently selected it plays the file as a pass through, otherwise it upsamples it to the max bitrate. Now that I'm running DSD256 most of the time most the DSD files I player are DSD64 and DSD128 and I'd like to upsample those to DSD256...but there are more and more DSD256 release and obviously I'd like to just pass those through. Unless I'm missing something it shouldn't be too hard to add and makes a lot of sense.

 

While some type of similar PCM feature might be nice it doesn't seem to be as important as DSD is much more processor demanding.

 

Thanks!

 

Robert

Software: Roon/HQplayer; System I: Roon Server/HQplayer DSD 512 Upsampling, Custom Windows 10 PC/AO, LPS-1 powered Startech USB card; LPS-1 powered ISO Regen; Holo Cyan DAC; VPI Scout 2 Turntable, Soundsmith Boheme, TTW Clamps and Carbon Matt; Cary SLP-98P Preamp; Van Alstine FET 600 Poweramp;  Aerial Acoustics 6T loudspeakers, SVS SB13 Ultra Subwoofers. System II: Custom PC with Signalyst Linux HQplayer NAA; LPS-1 powered Startech USB card; LPS-1 powered ISO Regen; IFI Micro iDSD Black Label; Primaluna Dialogue 2 with Tung Sol KT-150; Paradigm Studio 20 v3 monitors on Custom Mapleshade stands. Cables: Moon Audio, LUSH, Kimber Kable, Mapleshade, LARRY custom.

Link to comment
Specifically it would be a check box in the DSDIFF/DSF Settings menu that requests that the program check the DSD file to be played and if it is the same as the max bitrate level currently selected it plays the file as a pass through, otherwise it upsamples it to the max bitrate. Now that I'm running DSD256 most of the time most the DSD files I player are DSD64 and DSD128 and I'd like to upsample those to DSD256...but there are more and more DSD256 release and obviously I'd like to just pass those through. Unless I'm missing something it shouldn't be too hard to add and makes a lot of sense.

 

While some type of similar PCM feature might be nice it doesn't seem to be as important as DSD is much more processor demanding.

 

Problem now is that HQPlayer doesn't know what is the reason for enabling DSP processing. It is not only about rate conversions, but also about volume control and other DSP functionality. For example with my exaSound e28 I use solely HQPlayer's volume control and if one track suddenly comes through at full level without volume control it would be pretty nasty surprise...

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

I'm back to that DSD256 multichannel file problem. I finally got the files downloaded properly and completely, converted from DFF to DSF, and they play fine in JRIver. However, 2 of the three movements (files/tracks) do not even load in HQPLayer, meaning I drag them as a folder and the folder does not load...I drag them as files and only movement 3 loads and plays. Any ideas? Same with the original correctly-sized file of movement 1 as DFF. Doesn't matter, it ain't loadin!

Link to comment
Problem now is that HQPlayer doesn't know what is the reason for enabling DSP processing. It is not only about rate conversions, but also about volume control and other DSP functionality. For example with my exaSound e28 I use solely HQPlayer's volume control and if one track suddenly comes through at full level without volume control it would be pretty nasty surprise...

 

Wow...forgot about that. I typically don't do much with volume (just leave it at -3db) and tend to play full albums...so it didn't make much difference to me. For DSD256 to DSD256 that's one "processing rich" volume control...but then again speakers can be pretty expensive the last time I checked.

 

Robert

Software: Roon/HQplayer; System I: Roon Server/HQplayer DSD 512 Upsampling, Custom Windows 10 PC/AO, LPS-1 powered Startech USB card; LPS-1 powered ISO Regen; Holo Cyan DAC; VPI Scout 2 Turntable, Soundsmith Boheme, TTW Clamps and Carbon Matt; Cary SLP-98P Preamp; Van Alstine FET 600 Poweramp;  Aerial Acoustics 6T loudspeakers, SVS SB13 Ultra Subwoofers. System II: Custom PC with Signalyst Linux HQplayer NAA; LPS-1 powered Startech USB card; LPS-1 powered ISO Regen; IFI Micro iDSD Black Label; Primaluna Dialogue 2 with Tung Sol KT-150; Paradigm Studio 20 v3 monitors on Custom Mapleshade stands. Cables: Moon Audio, LUSH, Kimber Kable, Mapleshade, LARRY custom.

Link to comment
I like poly-sinc with DSD7 or ASDM7. But with you processor you will definitely want to use the poly-sinc 2s to reduce processing demand. I'm guessing you will be right on the edge of being able to do DSD256. I used to use a 2.2gHz i7 Mac Mini and it was hit and miss. From redbook it might work...if not go to DSD128. If you're upsampling redbook also give the closed-form filters a try.

 

On the mac I'm pretty sure you will need to use DoP. Its been a while since I used the IFI connected directly to a mac...I use a linux NAA in between them usually...I'm I'm pretty sure it won't go native...but because it passes PCM 702 you are good up to DSD256.

 

Robert

 

You hit the nail on the head, it is running dsd128 fine so far. Started having dropout about two minutes in on 256. I was running Poly Sinc shrt mp 2s I think. Yes, I had to enable dop to see the dam choice in the drop down menu. Is it normal to get some popping after stopping a track? It pops after a few seconds, right at the same time the drop down menus get to where you can change them again. Any idea on the pop after tracks stop? I also cannot adjust buffer time, is this normal? It is grated out.

 

Are certain filters that are the lowest cpu utilization for dsd 256?

 

Thanks,

Tim

Link to comment
Is it normal to get some popping after stopping a track?

 

Depending on a DAC yes, for example with iFi or Mytek Stereo192-DSD DAC which don't have output mute relays..

 

I also cannot adjust buffer time, is this normal?

 

With CoreAudio it is not possible, because OS X wants to manage buffers on it's own...

 

Are certain filters that are the lowest cpu utilization for dsd 256?

 

Generally the "-2s" ones are the lower CPU utilization ones.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
You hit the nail on the head, it is running dsd128 fine so far. Started having dropout about two minutes in on 256. I was running Poly Sinc shrt mp 2s I think. Yes, I had to enable dop to see the dam choice in the drop down menu. Is it normal to get some popping after stopping a track? It pops after a few seconds, right at the same time the drop down menus get to where you can change them again. Any idea on the pop after tracks stop? I also cannot adjust buffer time, is this normal? It is grated out.

 

Are certain filters that are the lowest cpu utilization for dsd 256?

 

Thanks,

Tim

 

Yes...I frequently get pops on the IFI at the beginning of playback. Stick with the poly-sinc 2s variants...but I'm afraid DSD256 may be beyond your machine. I just mentioned yesterday a "cheap" Windoz machine available at Costco that will get the job done for you easily with DSD256...might also do DSD512...and would give you an ASIO driver for your IFI to play with...would make DSD512 experimentation possible. It packs a MUCH faster i7 processor along with a CUDA enabled card for even more power.

Set it up with HQplayer and maybe Fidelizer and forget about it. You can drive it with roon from your mac.

 

That's more or less what I do. Took me a while to figure this all out...but it is working beautifully now...and DSD256 is really worth the effort. The turntable was just complaining last night that if feels neglected...better spin some vinyl tonight.

 

Robert

Software: Roon/HQplayer; System I: Roon Server/HQplayer DSD 512 Upsampling, Custom Windows 10 PC/AO, LPS-1 powered Startech USB card; LPS-1 powered ISO Regen; Holo Cyan DAC; VPI Scout 2 Turntable, Soundsmith Boheme, TTW Clamps and Carbon Matt; Cary SLP-98P Preamp; Van Alstine FET 600 Poweramp;  Aerial Acoustics 6T loudspeakers, SVS SB13 Ultra Subwoofers. System II: Custom PC with Signalyst Linux HQplayer NAA; LPS-1 powered Startech USB card; LPS-1 powered ISO Regen; IFI Micro iDSD Black Label; Primaluna Dialogue 2 with Tung Sol KT-150; Paradigm Studio 20 v3 monitors on Custom Mapleshade stands. Cables: Moon Audio, LUSH, Kimber Kable, Mapleshade, LARRY custom.

Link to comment
Depending on a DAC yes, for example with iFi or Mytek Stereo192-DSD DAC which don't have output mute relays..

 

 

 

With CoreAudio it is not possible, because OS X wants to manage buffers on it's own...

 

 

 

Generally the "-2s" ones are the lower CPU utilization ones.

 

Thanks for the replay Miska! I have an iMac with a 2.7ghz quad core i5. The mac mini is a dual core. Would this make a difference being able to play did 256?

 

Thanks!!!

Tim

Link to comment
Thanks for the replay Miska! I have an iMac with a 2.7ghz quad core i5. The mac mini is a dual core. Would this make a difference being able to play did 256?

 

Yes... On iMac you can enable "Pipeline SDM" in HQPlayer settings while on dual-core Mac Mini it should be disabled. This makes quite a big difference.

 

My Mac Mini can do DSD256 with the -2s filters, at least on Windows 8.1 which I'm using on it most of the time (BootCamp). I also have a quad-core i5 iMac with only OS X, which has no problems with DSD256.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
I've been using HQPlayer with Ubuntu Studio for some time and been very pleased with the performance. But on a whim I decided to try Debian stretch OS this week. I also upgraded HQPlayer from 3.12 to 3.13 at the same time which is a bit silly - its hard to pin down whether changes are due to new OS, or upgraded app. Anyway, after a couple of days listening I am certain that my computer audio system sounds better than ever. The main improvement is in the sound stage. Instruments have a very clear position in space and are totally free of the speaker. But any description of SQ changes are subjective, the objective observation that I have made relates to CPU usage. Previously (Ubuntu Studio, HQP 13.12) when playing hi res file, CPU usage was around 46%. Now (Debian, HQP 13.13) CPU usage is 38% with same settings.

 

I use a JLSounds USBtoI2S card, and HQPlayer upsamples everything to DSD256. Filer is poly sinc mp 2s, modulator DSD7 256+fs.

 

I like the automatic rate selection. I always had to switch the sample rate manually when moving from 48k family (12.3Mhz) to 44.1k family (11.3Mhz), and this is a nice touch.

 

Thanks Jussi, I love HQPlayer.

 

Regards, Hazard

 

I'm also running Debian Stretch with an Core i3 processor and can upsample to DSD256 with the poly-sinc filters, the processor runs at about 80%, but I don't have any issues. With Ubuntu studio I couldn't even use the poly-sinc 2 filters at DSD64 without studdering. That big of a dfference in performance doesn't make much sense.

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...