HQ Player

[BCRich]

This may be redundant, but can you see the Rendu on the network? If so did you check the status of your DAC and did you also try to restart HQP in the Sonic Orbiter interface?

I know how frustrating this all can be....

Mike

[Ellsworth]

Thanks for the ideas.  It is great that people go out of their way to offer suggestions. I have not played any music since Dec. 18 so that update may be what is causing the issue.  Fortunately,HQP is seeing the Rendu so I feel that I have a good connection.  I have also gone back and forth restarting it in SonicOrbiter.  Tomorrow night I am going to try to play from HQP directly.  That should help narrow the range of possibilities.  Thanks again!

[Bob Stern]

@Ellsworth I forgot to suggest:  The most basic first step when you have problems with a computer is to restart the computer.  Did you do that?

 

Next, launch Activity Monitor and sort by highest CPU usage to see if there is a process using more than 10% CPU.  Wait 30 seconds for Activity Monitor to settle down before reaching conclusions.  The Spotlight-related processes have names beginning "md" (metadata).

[Miska]

On 1/3/2020 at 3:46 AM, ronnie.sengupta said:

The T+A Firmware I installed for Linux flips channels for DSD.

 

There's a separate CPLD firmware for Amanero to deal with that channel swap issue.

 

On 1/3/2020 at 3:46 AM, ronnie.sengupta said:

@miska any thoughts on which way to go - CPU or GPU upgrade for 512 xtr.

 

Hard to say. If you need to replace CPU, motherboard and RAM to upgrade CPU, you can compare the prices. It also depends if you have some other use for the hardware.

 

 

I promised to provide load figures fort poly-sinc-xtr-lp from 44.1k to DSD at 44.1k x512...

 

Here's GTX1080, doesn't work, frequent drop-outs:

 

Here's RTX2080, doesn't work reliably, intermittent drop-outs:

 

Here's RTX2080Ti, works reliably:

 

 

I suspect bigger difference here is due to bigger L2 cache (2/4 MB vs 6 MB) on RTX2080Ti and wider memory bus bandwidth (256-bit vs 352-bit, 320/496 GB/s vs 616 GB/s). Titan RTX would bump the memory bus to 384-bit at 672 GB/s, but same cache size as 2080Ti. Although the RTX2080Ti chip is much bigger compared to RTX2080 too, 13.6 vs 18.6 billion transistors (545 sq mm vs 754 sq mm). GTX1080 is 7.2 billion transistors and 314 sq mm.

 

[ronnie.sengupta]

1 hour ago, Miska said:

 

There's a separate CPLD firmware for Amanero to deal with that channel swap issue.

 

 

Hard to say. If you need to replace CPU, motherboard and RAM to upgrade CPU, you can compare the prices. It also depends if you have some other use for the hardware.

 

 

I promised to provide load figures fort poly-sinc-xtr-lp from 44.1k to DSD at 44.1k x512...

 

Here's GTX1080, doesn't work, frequent drop-outs:

 

Here's RTX2080, doesn't work reliably, intermittent drop-outs:

 

Here's RTX2080Ti, works reliably:

 

 

I suspect bigger difference here is due to bigger L2 cache (2/4 MB vs 6 MB) on RTX2080Ti and wider memory bus bandwidth (256-bit vs 352-bit, 320/496 GB/s vs 616 GB/s). Titan RTX would bump the memory bus to 384-bit at 672 GB/s, but same cache size as 2080Ti. Although the RTX2080Ti chip is much bigger compared to RTX2080 too, 13.6 vs 18.6 billion transistors (545 sq mm vs 754 sq mm). GTX1080 is 7.2 billion transistors and 314 sq mm.

 

 

Thank you Jussi for the results.

 

I got really impatient over the weekend and upgraded CPU to i7 9700k - that didn't help at all. So, this morning I ordered 2080 ti, just installed and ran poly-sinc-xtr-lp from 44.1k to DSD at 44.1k x512

 

Success! Its running CPU at 28% and GPU at 48% - very similar to your results.

 

I could have done without the CPU upgrade the GPU made all the difference.

 

[sbenyo]

On 12/31/2019 at 2:04 PM, Miska said:

 

If you use poly-sinc-long for converting between families, it can take a long time to initialize. Several minutes or hours, depending on machine and settings. Once it has been initialized once, and there are no changes, subsequent starts are faster. GPU is not used for initialization.

 

Did you use the same source rate + output rate combination today and the day before? The conversion ratio is what makes a huge difference (factor of several hundred in initialization times).

 

@Miska, here are my settings. Used to work but something happened. CPU gets high and I have to kill it.

I am using same tracks that used to work. CPU only, gives the same results.

 

Surprisingly, changing to the exterme xtr (CPU only) just works fine.I have AMD 1920x with gtx 2070 super.

 

[Miska]

6 minutes ago, sbenyo said:

@Miska, here are my settings. Used to work but something happened. CPU gets high and I have to kill it.

I am using same tracks that used to work. CPU only, gives the same results.

 

Surprisingly, changing to the exterme xtr (CPU only) just works fine.I have AMD 1920x with gtx 2070 super.

 

With those settings you just need to wait long enough. Most likely it'll finish initialization in less than a minute. During initialization having all the sibling threads of the CPU at 100% is to be expected.

 

Yes, xtr is much faster to initialize. GPU doesn't matter for initialization.

 

[Ellsworth]

20 hours ago, Bob Stern said:

@Ellsworth I forgot to suggest:  The most basic first step when you have problems with a computer is to restart the computer.  Did you do that?

 

Next, launch Activity Monitor and sort by highest CPU usage to see if there is a process using more than 10% CPU.  Wait 30 seconds for Activity Monitor to settle down before reaching conclusions.  The Spotlight-related processes have names beginning "md" (metadata).

Excellent suggestion.  Thanks!  I have tried to play HQP as a stand-alone player and continue to experience the dropouts.  At this point, we can confirm that the interaction with Roon is not causing the problem. 

 

Spotlight doesn't show any activity in the Activity Monitor.  When upsampling to DSD128, I am seeing CPU usage fluctuate from 70 to 100%.  When playing at PCM 44k, CPU usage runs from 2 to 6%.  There are no other activities running that go over 3% CPU usage.  The dropouts happen at both DSD128 and PCM44K.

 

After not changing anything on the computer, i can’t figure out why HQP is doing this.  The only possible solutions I can think of now are to reinstall HQP or upgrade from El Capitan to Catalina.  I am wondering if a setting in HQP might have changed on its own but that doesn’t make any sense.

[Miska]

12 minutes ago, Ellsworth said:

After not changing anything on the computer, i can’t figure out why HQP is doing this.  The only possible solutions I can think of now are to reinstall HQP or upgrade from El Capitan to Catalina.  I am wondering if a setting in HQP might have changed on its own but that doesn’t make any sense.

 

Since you are using Rendu, any changes to the network overall (hardware or software changes), or updates to the Rendu?

 

Are you able to try without Rendu, DAC connected directly to the computer, to simplify the setup even further?

 

[LoryWiv]

6 hours ago, ronnie.sengupta said:

 

Thank you Jussi for the results.

 

I got really impatient over the weekend and upgraded CPU to i7 9700k - that didn't help at all. So, this morning I ordered 2080 ti, just installed and ran poly-sinc-xtr-lp from 44.1k to DSD at 44.1k x512

 

Success! Its running CPU at 28% and GPU at 48% - very similar to your results.

 

I could have done without the CPU upgrade the GPU made all the difference.

 

Thanks for sharing your results. I am a bit confused, though, as I thought the CPU was key in upsampling to high rate DSD and to running modulators whereas GPU is more relevant to running the filters...is that conceptually correct?

[Ellsworth]

1 hour ago, Miska said:

 

Since you are using Rendu, any changes to the network overall (hardware or software changes), or updates to the Rendu?

 

Are you able to try without Rendu, DAC connected directly to the computer, to simplify the setup even further?

 

Ther have been no changes to the network or firmware.  Unfortunately I can’t easily play directly to my DAC as the Mac in on the third floor and the DAC is on the first floor.  This would require some effort.  

 

I originally thought the Rendu was the issue.  However, I can use Roon without any issues.  Unfortunately the sound quality of Roon doesn't come close to HQP.  Thanks for the comment.

[57gold]

Downloaded an album off Nugs of Gov't Mule 12-31-19 show in ALAC-HD and it won't play on HQ Player. 

 

Shows up in iTunes.  So what did I do wrong?  Other Nugs downloads work fine.

[Outlaw]

2 hours ago, 57gold said:

Downloaded an album off Nugs of Gov't Mule 12-31-19 show in ALAC-HD and it won't play on HQ Player. 

 

Shows up in iTunes.  So what did I do wrong?  Other Nugs downloads work fine.

I don't think ALAC is supported in HQPlayer

[blue2]

I believe a lossless conversion from ALAC to FLAC is possible.

[Miska]

8 hours ago, Ellsworth said:

I originally thought the Rendu was the issue.  However, I can use Roon without any issues.  Unfortunately the sound quality of Roon doesn't come close to HQP.  Thanks for the comment.

 

Roon has all kinds of workarounds for problems, so Roon working doesn't yet prove that the problem couldn't exist in that area.

 

[Miska]

58 minutes ago, blue2 said:

I believe a lossless conversion from ALAC to FLAC is possible.

 

Yes, since both formats are lossless, one can convert between the two while data remains identical (just like AIFF/WAV too). Metadata formats used by the two differ quite a lot, so if there's a loss, it would be some lost metadata information in the translation.

 

[Ellsworth]

3 hours ago, Miska said:

 

Roon has all kinds of workarounds for problems, so Roon working doesn't yet prove that the problem couldn't exist in that area.

 

I didn’t realize Roon could work while HQP might not.  The next step will be to bring the computer downstairs when I get home tonight and try to connect directly to the DAC without the Rendu in the chain.  If that configuration works, we can narrow it down to the network connection.   Thanks for your help.   

[57gold]

Thought ALAC HD was some kind oƒ hi def AIFF file. 

 

Tried to convert using XLD but it looks like XLD doesn't do ALAC either.

 

Could have downloaded a FLAC version of same show, but since my entire library is AIFF...

 

Any ideas on how to convert ALAC to AIFF or FLAC?

[Solstice380]

4 minutes ago, 57gold said:

Thought ALAC HD was some kind oƒ hi def AIFF file. 

 

Tried to convert using XLD but it looks like XLD doesn't do ALAC either.

 

Could have downloaded a FLAC version of same show, but since my entire library is AIFF...

 

Any ideas on how to convert ALAC to AIFF or FLAC?

 

iTunes itself will convert AAC to AIFF, maybe it will do ALAC to AIFF?  It's been a few years since I had a Mac.  I've heard of xAct and MAX as other OSX converters.  Is there a dBPoweramp for Mac?  I think dBP Batch Convert will do it.

[perpetualapprentice]

51 minutes ago, Solstice380 said:

...Is there a dBPoweramp for Mac?  I think dBP Batch Convert will do it.

 

+1 for dBPoweramp batch conversion.

 

Yate is my go-to if you need to clean up tags. Steep learning curve but have yet to encounter something it cannot do.

 

--- from a Mac user since 1988

[Bob Stern]

4 hours ago, 57gold said:

Tried to convert using XLD but it looks like XLD doesn't do ALAC either.

 

XLD will convert ALAC with .m4a extension to FLAC.  I've done it a hundred times.  What is the filename extension?  How do you know it's ALAC?  Perhaps the HD refers to some proprietary encoding.

 

In the Terminal program, type afinfo, then a space, then drag the track into the terminal window, then press return.  If it's ALAC, the output should include something like:

File type ID:   m4af

Data format:     2 ch,  96000 Hz, 'alac'

 

While you're in Terminal, you can use ffmpeg to convert it to FLAC.  Type:

ffmpeg -i

type a space

drag the track into the terminal window

type:  -vn

type a space

drag the track into the terminal window again

delete the filename extension and type flac in its place

press return

 

 

 

[Ellsworth]

Thanks again for any comments.  Sorry to keep bothering people on this thread with my problems....I reset the router tonight and had good success playing HQP as a stand-alone player through the Rendu.  This played for 30 minutes with no problem.  I thought all was well.

 

When trying to integrate Roon, all hell broke loose.  I lost connection to the Rendu and could not get Roon or HQP to recognize the Rendu.  After a number of reboot cycles with the DAC, Rendu, and Mac, I had success.  However when controlling HQP with Roon the dropouts started again.   When looking at Roon, I get a rectangle  that flashes across the middle with a blue bar on the left side.  I also get a spinning wheel that prevents me from trying to navigate in Roon.  

 

I am am not sure if that is a HQP issue or if it is a Rendu issue.  At this point, I don’t have the time to move the Mac and try a direct connection to the DAC.  

 

Unless, I can figure out a different fix, that will be my chore for tomorrow night.

 

 

[ronnie.sengupta]

I realized that when I did my tests earlier, I had the multicore setting set to Checked which resulted in stuttering on my i7-9700k with redbook to  ASDM7EC and xtr filters at dsd 256. With Multicore set to "greyed out" I get no stutter and the CPU stays at 55 to 70% with no overclocking, running at clock speeds of around 4600 Mhz.

 

The same settings were running at close to 100% on my i7-7700k overclocked, running at clock speed for 4700Mhz to 5000 Mhz (no stuttering)

 

 

I ran some additional tests on my PC (i7-9700k + RTX 2080 ti) with and without Cuda, all three settings for Multicore,  EC / Non EC modulators and xtr / ext2 filters. 

 

 

@LoryWiv

XTR filters seem to benefit quite a bit from the Cuda offload as can be seen in row 11 and 12 above for ASDMEC and Row 1 and 3 for ASDSM7

 

@Miska

I ran the ASDSM7/xtr/512 tests on my registered HQP 3 version it works without stutter at 85-95% CPU. Same test on HQP 4.2.3 stutters and uses ~74% CPU. Without Cude and Multicore greyed out.

 

HQP 3.25.6

 

 

 

 

HQP 4.2.3

 

 

 

[Carlos269]

Ronnie, does your DAC support 48KHz based clock rates? Are those DSD512 and DSD256 datasets for 48x512 & 48x256 or for 44.1x512 & 44.1x256? I believe that I know the answer but want to confirm. Thanks.

[Miska]

40 minutes ago, ronnie.sengupta said:

 

@Miska

I ran the ASDSM7/xtr/512 tests on my registered HQP 3 version it works without stutter at 85-95% CPU. Same test on HQP 4.2.3 stutters and uses ~74% CPU. Without Cude and Multicore greyed out.

 

HQP 3.25.6

 

 

 

 

HQP 4.2.3

 

 

 

 

This is because core allocation and reservation in HQPlayer 4. On a 8+ core CPU, two cores are reserved for low intensity tasks and for things like Roon. This works generally well until you hit a case where you would need those two cores for running the filters too, this happens only with the heaviest filters though. You can disable core reservation by setting environment variable HQPLAYER_RESERVED_CORES value to 0 (Control Panel -> System -> Advanced -> Environment variables). If the CPU would have HyperThreading, you could also move the reserved cores to the threads by specifying suitable mask, this would free up all primary siblings to intensive tasks.