MarcinGD Posted May 24, 2018 Share Posted May 24, 2018 3 hours ago, tboooe said: I personally feel using a Windows based NAA with drivers sound better based on my limited testing. I've only compared the Allo Sparky to my i3 Server 2012 R2 NAA. I believe some other folks have reported the same results. I need to next put the image on the same i3 machine and compare. Hey tboooe, I also started to test NAA solutions. A day ago I assembled RPi 3 with digione. Why digione and not usbridge ? Well because drivers for any kind of usb under linux based systems make it sound without life. With digione It's different. As stand alone unit it's just good for the money. But as NAA I have problem now what I like more 48x512 or 192k from digione. What is Your opinion about LPS on HQPlayer control machine that feed signal to NAA ? Link to comment
Miska Posted May 24, 2018 Share Posted May 24, 2018 5 hours ago, westside said: @Miska Why do you not answer me ? I pay the software like eveyone no ? Because I don't have a good answer... Just answering "I don't know" is not so great. Did you try on the same HQPlayer machine first locally without NAA with the same settings? Usually if you have drop-outs/stuttering, the CPU load is too high. But without knowing your settings, details of your HQPlayer machine and it's load situation it is hard to say. It could also be network problem between HQPlayer and NAA, depending on what kind of network infra you have. Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
Miska Posted May 24, 2018 Share Posted May 24, 2018 4 hours ago, jimdukey said: What's new with HQ 3.21? See https://www.signalyst.com Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
Miska Posted May 24, 2018 Share Posted May 24, 2018 8 hours ago, Sloop John B said: I'm unsure where to post this as it includes many aspects, HQP, SU-1 and the Holo Audio Spring DAC. I seem to have developed issues with DSD 512 though ultra Rendu with HQP to SU-1 so am now using an older NUC I had lying around as an NAA to try and isolate if the issue is with the Rendu (as I suspect). I have the network NAA daemon running on a Win 10 machine using the Singxer ASIO driver with the following settings Are there advantages (or disadvantages for that matter) to using the bootable image version of the NAA? .sjb You are using very tiny buffers which is likely to cause playback problems like pops/clicks and increases stress on the machine. When I use a Thesycon's driver on Windows (like the own shown in screenshot), I just set buffer size to as big as possible. It still tends to be on a small side for DSD512 though. Advantage of the bootable image is that the entire OS is built precisely for the NAA use and nothing else. A lot smaller size (around 100 MB for entire OS with everything). Another advantage is that there's no such shiny graphics output, only text console, so the GPU is not actively used for anything. It is also very simple, "fire and forget", you boot it up and it's ready for use, that's it. Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
jimdukey Posted May 24, 2018 Share Posted May 24, 2018 Thanks Miska, I didn't know about that! Link to comment
Miska Posted May 24, 2018 Share Posted May 24, 2018 7 hours ago, Le Concombre Masqué said: I never boost more than 6 dB (less indeed) and to fill nulls I rather lower the target level ; however I always introduce a Manual LS in REW. I just designed new files with less ambitious LS gain (110/3) and the IR gain is #4.5 (vs 8+ with the previous set) : I'll give it a try Let's say you have 4.5 dB max boost, and you have gain comp in HQPlayer set to 0, then you shouldn't turn volume control in HQPlayer higher than -7.5 dB (-3 for headroom and -4.5 for IR gain -> -7.5 total). 7 hours ago, Le Concombre Masqué said: I started without success a thread with a pinch of laugh at MQA, stressing than rather than addressing ADC idiosyncrasies they should rather focus on target curves used to calibrate the monitors. I feel that different convolution files sets might be an alternative to changing filters (ie linear vs mp) depending on source. However I think I'll end my crisis soon with a BK target and a whateveritbringswesing mindset.... Yes! BK target is pretty good. I use similar, with -6 dB slope down from 20 Hz to 20 kHz. Has been working good for me... Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
AnotherSpin Posted May 25, 2018 Share Posted May 25, 2018 Miska, Volume dial becomes red from time to time (on 1 album from 10 or 15), even though DirectSDM mode is on (checked) and volume controlled outside the HQP. Vol max is set to -6.0dB. How you can explain it and what would be done? Thank you. Link to comment
Miska Posted May 25, 2018 Share Posted May 25, 2018 When DirectSDM is enabled (checked), SDM output is at fixed -3 dBFS volume. You need to disable DirectSDM in order for volume settings to have effect. Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
Le Concombre Masqué Posted May 25, 2018 Share Posted May 25, 2018 16 hours ago, Miska said: Let's say you have 4.5 dB max boost, and you have gain comp in HQPlayer set to 0, then you shouldn't turn volume control in HQPlayer higher than -7.5 dB (-3 for headroom and -4.5 for IR gain -> -7.5 total). Yes! BK target is pretty good. I use similar, with -6 dB slope down from 20 Hz to 20 kHz. Has been working good for me... Thank you I might keep on hand a set targeted on JBL synthesis curve, for those tracks where the bass line is so attractive (and overwhelming other criteria) with that curve that you may think it was mastered with such a curve. Too bad disclaiming its target curve is not mandatory for mastering engineers ! Miska 1 Link to comment
rickca Posted May 25, 2018 Share Posted May 25, 2018 @Miska, which commercial DAC's do you feel perform best with HQPlayer? What are a couple of top contenders for DSD? For PCM? Pareto Audio AMD 7700 Server --> Berkeley Alpha USB --> Jeff Rowland Aeris --> Jeff Rowland 625 S2 --> Focal Utopia 3 Diablos with 2 x Focal Electra SW 1000 BE subs i7-6700K/Windows 10 --> EVGA Nu Audio Card --> Focal CMS50's Link to comment
luisma Posted May 26, 2018 Share Posted May 26, 2018 On 5/24/2018 at 10:19 AM, tboooe said: I personally feel using a Windows based NAA with drivers sound better based on my limited testing. I've only compared the Allo Sparky to my i3 Server 2012 R2 NAA. I believe some other folks have reported the same results. I need to next put the image on the same i3 machine and compare. Thanks @tboooe Please keep us posted on this. Can someone confirm this? I get the SQ improvement of Windows over Linux and Direct Connection to DAC vs network renderer only (or I should say mostly) applies to owners of Holo Spring (and/or SU-1) combinations for which Windows Drivers (as @tboooe has mentioned before) may be better. Possibly other DAC's may not shown such difference in SQ when using Win vs Linux or DirectUSB vs NAA. Also we cannot rule out the synergy between SU-1 -- I2S -- Holo Dac My trial version of Roon is expiring and need to decide which way to go, I believe HQP is one of them being fed by MinimServer - BubbleUPNP+HQP and Rygel, all running over Ubuntu (I believe that's what Jussi recommends) I'm just afraid of losing SQ with my Pro-ject S2 or my future (yet to purchase DAC) which I believe may be a Brooklyn+ or Denafrip Ares or else. Link to comment
macuniverse Posted May 27, 2018 Share Posted May 27, 2018 OK, I'm using the HQP in try (30 min limit) mode to see whether I can improve over the JRiver player. For some reason, when I play a Flac track from the NAS, I don't get the drop-down options in the player (such as resampling filter, dithering, sample rate, etc). They are all dimmed. My setup is Win10 (latest), Wasapi driver, default endpoint selected, Win sound device is on-board toslink which is selected as the default device. I've set the toslink device directly as well but didn't help. Any assistance much appreciated Link to comment
jimdukey Posted May 27, 2018 Share Posted May 27, 2018 Max Volume is at -60, not -6. Miska told me -3 on both is a good setting. Link to comment
westside Posted May 27, 2018 Share Posted May 27, 2018 On 5/24/2018 at 11:12 PM, Miska said: Because I don't have a good answer... Just answering "I don't know" is not so great. Did you try on the same HQPlayer machine first locally without NAA with the same settings? Usually if you have drop-outs/stuttering, the CPU load is too high. But without knowing your settings, details of your HQPlayer machine and it's load situation it is hard to say. It could also be network problem between HQPlayer and NAA, depending on what kind of network infra you have. Hi @Miska @allmember With NAA and without NAA is there a big difference in term of sound quality ? Cheers. Link to comment
jimdukey Posted May 27, 2018 Share Posted May 27, 2018 My USB Dacs have responded well to the AQ Jitterbug, and IFI USB gadget whose correct name I can't think of. The Ifi cost about the same as the MicroRendu. I'n very pleased with them! The NAA thing is too complex for me... I also like my HAP Z1 Digital Player. Link to comment
Miska Posted May 27, 2018 Share Posted May 27, 2018 1 hour ago, westside said: With NAA and without NAA is there a big difference in term of sound quality ? It completely depends on your system. It can vary from no difference to a big difference. And sometimes reason for using NAA is not because of direct difference in sound quality but because of other reasons. For example a powerful server running HQPlayer can be acoustically noisy due to lot of fans used for cooling. Using a NAA allows one to have the server outside of listening room, in the basement or something. Or if you have one server and various DACs in different rooms and you want to be able to play to these DACs from the same server (which is true for example in my case). Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
Miska Posted May 27, 2018 Share Posted May 27, 2018 On 5/26/2018 at 7:49 PM, luisma said: Thanks @tboooe Please keep us posted on this. Can someone confirm this? I get the SQ improvement of Windows over Linux and Direct Connection to DAC vs network renderer only (or I should say mostly) applies to owners of Holo Spring (and/or SU-1) combinations for which Windows Drivers (as @tboooe has mentioned before) may be better. Possibly other DAC's may not shown such difference in SQ when using Win vs Linux or DirectUSB vs NAA. Also we cannot rule out the synergy between SU-1 -- I2S -- Holo Dac My trial version of Roon is expiring and need to decide which way to go, I believe HQP is one of them being fed by MinimServer - BubbleUPNP+HQP and Rygel, all running over Ubuntu (I believe that's what Jussi recommends) I'm just afraid of losing SQ with my Pro-ject S2 or my future (yet to purchase DAC) which I believe may be a Brooklyn+ or Denafrip Ares or else. I have only one Windows machine left in regular use which has T+A DAC8 DSD and exaSound e28 connected to it. Everything else is Linux. I use my Holo Spring with it's built-in USB straight from fanless HQPlayer Embedded machine and it performs perfectly fine. Pro-Ject S2 and Mytek Brooklyn DAC+ both work perfectly fine on Linux too (I've implemented native DSD support for both). Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
Miska Posted May 27, 2018 Share Posted May 27, 2018 On 5/26/2018 at 12:34 AM, rickca said: @Miska, which commercial DAC's do you feel perform best with HQPlayer? What are a couple of top contenders for DSD? For PCM? There are some devices listed as "Recommended hardware" on my web page. In most cases you'd want to play PCM sources upsampled to high rate DSD and forget sending PCM to the DAC. Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
asdf1000 Posted May 30, 2018 Share Posted May 30, 2018 On 10/27/2017 at 7:07 PM, Miska said: Here's a starting point for iFi micro iDSD (BL): Set the DAC filter to "Standard" Hi Jussi, I've always been confused by the iDSD's filter settings, for DSD, even when iFi have posted about it. Attached below is from the manual. Does the 'standard' setting filter the most ultrasonic noise, i.e. lowest cutoff/corner frequency for the low pass analogue filter? And the bit-perfect setting ('extreme' for DSD) has the highest cutoff frequency? Do you have the actual cutoff frequencies (approx.) for each setting, i.e. 1 standard (standard for DSD?) 2 min-phase (extended for DSD?) 3 bit-perfect (extreme for DSD?) I know those words not in brackets apply to the PCM filters but that's how the physical buttons are labelled on the DAC itself. Cheers in advance Link to comment
lmitche Posted May 30, 2018 Share Posted May 30, 2018 3 hours ago, Em2016 said: Hi Jussi, I've always been confused by the iDSD's filter settings, for DSD, even when iFi have posted about it. Attached below is from the manual. Does the 'standard' setting filter the most ultrasonic noise, i.e. lowest cutoff/corner frequency for the low pass analogue filter? And the bit-perfect setting ('extreme' for DSD) has the highest cutoff frequency? Do you have the actual cutoff frequencies (approx.) for each setting, i.e. 1 standard (standard for DSD?) 2 min-phase (extended for DSD?) 3 bit-perfect (extreme for DSD?) I know those words not in brackets apply to the PCM filters but that's how the physical buttons are labelled on the DAC itself. Cheers in advance HI Em2016, I share your confusion. We asked IFI to explain the differences and they refused with their usual disdain for customers. FWIW, I use #1 standard for playback here as it just sounds best at DSD512. Jud and Alex C agree. It will be curious to learn what Jussi has to say. asdf1000 1 Pareto Audio aka nuckleheadaudio Link to comment
asdf1000 Posted May 30, 2018 Share Posted May 30, 2018 1 minute ago, lmitche said: FWIW, I use #1 standard for playback here as it just sounds best at DSD512. Thanks lmitche Standard is also the loudest of the 3 (easily heard with using the headphone output, less obvious if using the iDSD as a preamp with RCA output). So I'm also curious how & why the different analogue filters affect overall volume also. Link to comment
Popular Post Miska Posted May 30, 2018 Popular Post Share Posted May 30, 2018 You can go directly to the source, TI's datasheet here: http://www.ti.com/lit/ds/symlink/dsd1793.pdf On pages 40 and 41 you can find frequency responses of four analog filters the DAC has - practically different data feed arrangements to the conversion elements. Differing gains of the filters are also stated there. Those gains are just direct result of the configurations. I have also my wide band measurement plots of the different filters, I'll post those later. I'm not sure how those exactly map to the iFi settings, but probably something like: Bit-perfect: AFIR 1 Standard: AFIR 3 or 4 asdf1000 and Jud 1 1 Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
asdf1000 Posted May 30, 2018 Share Posted May 30, 2018 16 minutes ago, Miska said: I have also my wide band measurement plots of the different filters, I'll post those later. I'm not sure how those exactly map to the iFi settings, but probably something like: Bit-perfect: AFIR 1 Standard: AFIR 3 or 4 Thanks Jussi, that looks right looking at the loudness/gain plots. The 'standard' toggle switch is loudest and the 'bit-perfect' toggle switch is lowest volume, so that matches with the gain plots you mention. Yes your own measurement plots would be the icing on the cake. We can have this clarification solved once and for all, without iFi's help ? Link to comment
jimdukey Posted May 30, 2018 Share Posted May 30, 2018 And there's the Pro IDSD Dac, coming out tomorrow I read! That has a LOT of settings, and how it would work with HQ/Mac is something I'll be interested to find out. When someone actually Gets one... Link to comment
Miska Posted May 30, 2018 Share Posted May 30, 2018 4 hours ago, Miska said: I have also my wide band measurement plots of the different filters, I'll post those later. Here are the plots, I've made these quite a while ago at DSD256 rate because it leaves more noise visible while DSD512 the noise floor is practically flat and as result is not useful for this test. Standard: Minimum-Phase: Bit-Perfect: (note - these are all "peak-hold" spectrums of 0 - 22.05 kHz sweep, so max peak level ones) asdf1000 1 Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
Recommended Posts
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 accountSign in
Already have an account? Sign in here.
Sign In Now