Jump to content
IGNORED

HQ Player


Recommended Posts

odd, when I check the signal path in Roon I see DSD256 not DSD256*44.1. I am using HQP version 3.15.1

 

Maybe Roon reports 44.1-base rates as "DSD64", "DSD128" and so on. And 48-base rates as "DSD64x48", "DSD128x48". I have not paid attention to this detail...

 

HQPlayer doesn't affect this part, because Roon gets the output sampling rate just a number from HQPlayer and then translates it to it's preferred presentation.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
Looks like I will be upgrading to that newe Ryzen 1700 CPU!!!! @Miska that CPU should have no issues with DSD512 correct?

 

I'm eying at 1800X for testing, but still waiting for good motherboard options to become available.

 

What I looked at 1700X may be minimum for DSD512, because I recommend at least 3.4 GHz base clock for Intel CPU's.

 

I will know more once I have 1800X system up and running...

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Hi,

 

HQPlayer and Network Audio Daemon can run on Windows, Linux and macOS, but I wonder if there is a preferred OS for running them. If there is any potential advantage on running them on any of these OS.

 

The reason for this question is that some DAC manufacturers claim that their DACs support only Windows and/or macOS, but they don't say anything about Linux, neither if it is supported or unsupported. An example could be the Teac devices. Would it be an issue running HQPlayer (or Network Audio Daemon) over Linux on a device attached to a Teac DAC?

 

I don't know if this question has already been answered in this thread. I have been looking for it but I have not found it. If this is an already answered question, I apologize.

 

Thanks!!

Link to comment
Hi,

 

HQPlayer and Network Audio Daemon can run on Windows, Linux and macOS, but I wonder if there is a preferred OS for running them. If there is any potential advantage on running them on any of these OS.

 

The reason for this question is that some DAC manufacturers claim that their DACs support only Windows and/or macOS, but they don't say anything about Linux, neither if it is supported or unsupported. An example could be the Teac devices. Would it be an issue running HQPlayer (or Network Audio Daemon) over Linux on a device attached to a Teac DAC?

 

I don't know if this question has already been answered in this thread. I have been looking for it but I have not found it. If this is an already answered question, I apologize.

 

Thanks!!

 

Yes, it has been discussed. My understanding (someone please correct me if I am wrong):

 

 

I believe with the Teac devices Windows may allow conversion to higher rate DSD due to Windows ASIO drivers permitting this versus the Linux drivers. Also, people who run their DACs direct to amplifiers without a hardware volume control (preamp, etc.) have reported "thumps" at a level that might be damaging when using Linux. I believe this is being worked on for the Linux driver.

 

 

I am happily running HQPlayer and an NAA on Linux with an iFi micro-iDSD DAC and a pre-amplifier and have none of these problems. They are specific to certain DACs and to systems that don't use analog volume control.

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical Ethernet to Fitlet3 -> Fibbr Alpha Optical USB -> iFi NEO iDSD DAC -> Apollon Audio 1ET400A Mini (Purifi based) -> Vandersteen 3A Signature.

Link to comment

Thank you Miska. I just realized the Ryzen does not come with integrated graphics and I do not have a separate GPU, though I only run HQP and Roonserver on a headless system. Will that be a problem?

 

I'm eying at 1800X for testing, but still waiting for good motherboard options to become available.

 

What I looked at 1700X may be minimum for DSD512, because I recommend at least 3.4 GHz base clock for Intel CPU's.

 

I will know more once I have 1800X system up and running...

12TB NAS >> i7-6700 Server/Control PC >> i3-5015u NAA >> Singxer SU-1 DDC (modded) >> Holo Spring L3 DAC >> Accustic Arts Power 1 int amp >> Sonus Faber Guaneri Evolution speakers + REL T/5i sub (x2)

 

Other components:

UpTone Audio LPS1.2/IsoRegen, Fiber Switch and FMC, Windows Server 2016 OS, Audiophile Optimizer 3.0, Fidelizer Pro 6, HQ Player, Roonserver, PS Audio P3 AC regenerator, HDPlex 400W ATX & 200W Linear PSU, Light Harmonic Lightspeed Split USB cable, Synergistic Research Tungsten AC power cords, Tara Labs The One speaker cables, Tara Labs The Two Extended with HFX Station IC, Oyaide R1 outlets, Stillpoints Ultra Mini footers, Hi-Fi Tuning fuses, Vicoustic/RealTraps/GIK room treatments

Link to comment
Thank you Miska. I just realized the Ryzen does not come with integrated graphics and I do not have a separate GPU, though I only run HQP and Roonserver on a headless system. Will that be a problem?

 

For installation and maintenance it is certainly much simpler when you have a GPU. The cheapest passive cooled graphics cards cost 40€, so it is not such a big cost to have one that is enough for those simple tasks.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
I'm eying at 1800X for testing, but still waiting for good motherboard options to become available.

 

What I looked at 1700X may be minimum for DSD512, because I recommend at least 3.4 GHz base clock for Intel CPU's.

 

I will know more once I have 1800X system up and running...

 

1700x and 1800x are the same chips, just clocked differently (AMD has acknowledged this). Can save money, get the 1700x, and change the clocks in bios. Get exactly the same chip, just costs less. Not clear yet if the 1700 (non-x) is also the same chip. It doesn't support XFR, so might be different. My bet is it isn't other than disabling XFR. Have to wait and see, but I'd bet the 1700 is from the same parts bin also, it just achieves it's lower TDP from lower clocks (which are easily adjusted).

Link to comment

Miska with the release of the AMD Ryzen line are there any generalities than can be made about what qualities are best in a processor for up sampling HQPlayer to DSD512, in a windows machine, with roon, HQP connecting directly to the dac via usb with say a uptone regen in-between dac and PC and NO convolution but a GPU for cuda offload.

 

Is faster base speed more important than more cores? i7-7700k (base 4.2 GHz, 4.5 Turbo), i7-6700k (base 4.0GHz turbo 4.2) vs say a Ryzen 1700x (base 3.4Ghz, turbo 3.8 Ghz). Or does more cores play a factor where the intel's above have 4 cores/8 threads and the ryzen has 8 cores/16 threads. Then or course there is L3 cache which I am not sure means anything but some says it does 8mb for the intel and 20 mb for the Ryzen.

 

In general do you think one should manually overclock their CPU's? The 7700k can easily get 4.8Ghz, the 6700k can easily get 4.6Ghz and the ryzen I'll bet could easily do 4.0Ghz. Or leave bios alone and let the CPU's run with no manual overclock?

Link to comment
Miska with the release of the AMD Ryzen line are there any generalities than can be made about what qualities are best in a processor for up sampling HQPlayer to DSD512, in a windows machine, with roon, HQP connecting directly to the dac via usb with say a uptone regen in-between dac and PC and NO convolution but a GPU for cuda offload.

 

Is faster base speed more important than more cores? i7-7700k (base 4.2 GHz, 4.5 Turbo), i7-6700k (base 4.0GHz turbo 4.2) vs say a Ryzen 1700x (base 3.4Ghz, turbo 3.8 Ghz). Or does more cores play a factor where the intel's above have 4 cores/8 threads and the ryzen has 8 cores/16 threads. Then or course there is L3 cache which I am not sure means anything but some says it does 8mb for the intel and 20 mb for the Ryzen.

 

You need certain amount of base clock speed that no number of cores can offset. What is required for base clock speed also depends on other factors like supported instruction set, instructions per clock cycle (or cycles per instruction) and such. Turbo rates are sort of nice, but that means otherwise under-utilized cores which allow boosting some cores higher.

 

Memory bandwidth and cache size also play a big role.

 

One example for comparison is this load figure of upsampling to DSD512 (with 4-channel CL12 DDR4-2400 RAM):

hqp-load-dsd512.png

 

It is not really even loaded. But that machine is really intended for pushing 8-channel DSD256 with convolution engine to the exaSound e28 DAC...

 

In general do you think one should manually overclock their CPU's? The 7700k can easily get 4.8Ghz, the 6700k can easily get 4.6Ghz and the ryzen I'll bet could easily do 4.0Ghz. Or leave bios alone and let the CPU's run with no manual overclock?

 

I would recommend first just loading the optimal defaults and maybe setting fans to silent mode for starters and see if all the needed things work. Overclocking tends to have impact on reliability and/or lifetime of the hardware.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
1700x and 1800x are the same chips, just clocked differently (AMD has acknowledged this). Can save money, get the 1700x, and change the clocks in bios. Get exactly the same chip, just costs less. Not clear yet if the 1700 (non-x) is also the same chip. It doesn't support XFR, so might be different. My bet is it isn't other than disabling XFR. Have to wait and see, but I'd bet the 1700 is from the same parts bin also, it just achieves it's lower TDP from lower clocks (which are easily adjusted).

 

It is quite typical that processors are speed-sorted. So many models may come out of same manufacturing line, but the ones that pass testing at higher clock speed are stamped as higher models and the ones that don't manage at such high clock speeds are stamped as slower models.

 

Similar thing happens with GPUs, for example all/most may come out of same manufacturing line, but the ones with malfunctioning execution units have some units fused out (disabled) and thus are stamped as lower models with less execution units. In addition the GPUs may be speed sorted too.

 

The overall ratios depend on yields that tend to improve over the manufacturing period. At the beginning yields for the highest class ones may be lower than later after tuning of the manufacturing process.

 

So you really cannot trust that every piece of 1700X would work reliably clocked like 1800X under heavy constant loads. One may be able to run couple of minutes test run, but not necessarily one year of 24/7 100% load test.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

"Memory bandwidth and cache size also play a big role."

 

Miska, that is quite an impressive CPU utilization screen shot. Looking at that I would extrapolate that memory bandwidth is really important, 4 bands vs. 2 bands for the 6700, 7700 and new AMD Ryzens. L3 cache is also very impressive with the 6950. So seeing this I would expect the intel i7-6800K to be a killer DSD512 machine for a very reasonable price when used for 2 channel playback with no convolution and a GPU thrown in for CUDA and of course graphics.

Link to comment
Miska, that is quite an impressive CPU utilization screen shot. Looking at that I would extrapolate that memory bandwidth is really important, 4 bands vs. 2 bands for the 6700, 7700 and new AMD Ryzens. L3 cache is also very impressive with the 6950. So seeing this I would expect the intel i7-6800K to be a killer DSD512 machine for a very reasonable price when used for 2 channel playback with no convolution and a GPU thrown in for CUDA and of course graphics.

 

For the same use case, 6800K is probably better. The load per individual core is lower for 7700K due to higher clock speed though. Because total load is aggregate load of all cores...

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
It is quite typical that processors are speed-sorted. So many models may come out of same manufacturing line, but the ones that pass testing at higher clock speed are stamped as higher models and the ones that don't manage at such high clock speeds are stamped as slower models.

 

Similar thing happens with GPUs, for example all/most may come out of same manufacturing line, but the ones with malfunctioning execution units have some units fused out (disabled) and thus are stamped as lower models with less execution units. In addition the GPUs may be speed sorted too.

 

The overall ratios depend on yields that tend to improve over the manufacturing period. At the beginning yields for the highest class ones may be lower than later after tuning of the manufacturing process.

 

So you really cannot trust that every piece of 1700X would work reliably clocked like 1800X under heavy constant loads. One may be able to run couple of minutes test run, but not necessarily one year of 24/7 100% load test.

 

Yep, all true. A perfect example is the Intel X700 series processors. The K, non-K and T variants are just binned versions of the same processor.

 

But, in this case they really are the same. AMD explicitly acknowledged this. They didn't send out any 1700X review samples, saying if overclocking there is no difference between it and the 1800X. And for what's it's worth, the first 1700 (non x, the 65tdp part) reviews are out and they are reaching the exact same hard cap when overclocking as with the 1800X and 1700X (4-4.1 at same voltages.)

 

All this aside, these chips are testing very strangely. They are doing very well in some math intensive benchmarks, and very poorly in others. Glad you're going to test one =). I'm hoping you'll downclock/undervolt your 1800X to simulate a 1700X/1700 to see how they do...

Link to comment

I've been getting great results with HQPlayer upsampling to DSD256 and playing through my diy DAC. I recently upgraded my DAC to DSD512 but alas my PC was not up the the task as recounted previously (AMD Phenom 6 core).

 

So today i bit the bullet. Bought a new CPU (I7 7700K with Cooler Master Hyper 103 fan), new mobo (Gigabyte z270) and new memory (32 GB of DDR4 2400). I can now upsample anything to DSD512 and CPU useage is only around 28%. And best of all, I can't hear the CPU fan anymore! So I am happy, it sounds great.

 

So if you are waiting for an AMD Ryzen - no need really. The new AMD may or not be better, but the I7 does the job and any more power is not needed for me at least.

Link to comment
I've been getting great results with HQPlayer upsampling to DSD256 and playing through my diy DAC. I recently upgraded my DAC to DSD512 but alas my PC was not up the the task as recounted previously (AMD Phenom 6 core).

 

So today i bit the bullet. Bought a new CPU (I7 7700K with Cooler Master Hyper 103 fan), new mobo (Gigabyte z270) and new memory (32 GB of DDR4 2400). I can now upsample anything to DSD512 and CPU useage is only around 28%. And best of all, I can't hear the CPU fan anymore! So I am happy, it sounds great.

 

So if you are waiting for an AMD Ryzen - no need really. The new AMD may or not be better, but the I7 does the job and any more power is not needed for me at least.

A quick update - with my old PC, CPU temp was sitting at 80 deg when upsampling to DSD256 - which explains the screaming CPU I often experienced. Its a warm night in Melbourne after a 33 deg day. And my bi-amp class A amps generate a lot of heat. CPU temp is now 62 deg, upsampling to DSD512. So all in all - its been a successful day.

Link to comment
I have for a few days this device:

Pure 2 Mastering AD/DA Converter & Clock | Antelope Audio. What best settings are you suggest with HQPlayer and Roon settings?

 

I don't have any specific info about that device, so probably best to run it at 176.4/192k sampling rate and pick a filter based on your preferences, referring to the HQPlayer manual...

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
Which OS? What is the log output when this happens? Any specific steps to reproduce it?
Windows 10.Last time it closed was when it was changing formats.Upsample everything to DSD 256.Closed last timewhen going from a dsd 64 to a 44.1/16 bit file.
Link to comment
Windows 10.Last time it closed was when it was changing formats.Upsample everything to DSD 256.Closed last timewhen going from a dsd 64 to a 44.1/16 bit file.

 

If you are converting everything to DSD256, then nothing should appear/happen at the NAA side when you switch source formats...

 

What is the DAC? It could be some glitch with the ASIO driver...

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
If you are converting everything to DSD256, then nothing should appear/happen at the NAA side when you switch source formats...

Sorry.I upsample everything except my dsd files which are dsd64

What is the DAC? It could be some glitch with the ASIO driver...

Dac is a Simaudio moon dsd280 d with streamer.Uses Asia driver for native dsd playback

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...