Jump to content
IGNORED

HQ Player


Recommended Posts

2 minutes ago, Miska said:

 

I wonder how much there are unit variations. Also zoom-in plots would be useful to better see where the deviation begins.

 

 

I have no seen anybody measuring Gungnir, but the DAC chip itself is 18-bit part. How linear it is, is another question, possibly somewhat less, but hard to say how much. You can try different settings from 18-bit down to 14-bit with NS9 dither/noise-shaping. With 18/17-bit and likely even with 16-bit you are also fine with TPDF dither.

 

Really quiet portions of music are good for testing this. As long as you don't start hearing background hiss you are good.

 

You can also emphasize the differences with normal loud material by turning down digital volume from HQPlayer to -40 dB for example and then turning up volume from amp. Just be careful when doing such test not to accidentally blast at loud volume! So music playback is better started first, then carefully increase volume from amp and turn it back down before touching the playback or settings again.

 

Do you have any recommendation when to use NS9, or NS5 or other noise shaping?

 

 

Link to comment
Just now, Yviena said:

Do you have any recommendation when to use NS9, or NS5 or other noise shaping?

 

There are some in the manual...

 

But overall, when digital noise level with flat dither like TPDF/Gauss1 is below DAC's analog SNR (background noise) it is just fine to use those flat dithers. 20-bit is about -120 dB and so on.

 

When the digital noise floor would dominate and you have extra bandwidth to spare (sampling rate >=176.4k), you can use noise shaping to push down the noise floor in lower frequencies. This allows you to reduce number of bits in the output and stick to the converter's linear range. For the lower sampling rates (below 352.8k), NS9 gives needed steeper step on the noise distribution. At higher rates you also have an option to choose more gentle ones like NS5.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
30 minutes ago, Miska said:

 

There are some in the manual...

 

But overall, when digital noise level with flat dither like TPDF/Gauss1 is below DAC's analog SNR (background noise) it is just fine to use those flat dithers. 20-bit is about -120 dB and so on.

 

When the digital noise floor would dominate and you have extra bandwidth to spare (sampling rate >=176.4k), you can use noise shaping to push down the noise floor in lower frequencies. This allows you to reduce number of bits in the output and stick to the converter's linear range. For the lower sampling rates (below 352.8k), NS9 gives needed steeper step on the noise distribution. At higher rates you also have an option to choose more gentle ones like NS5.

I see... So for 8x rates NS5 is probably ideal.

Link to comment

Jussi,

 

My DAC is the Brooklyn, not +. It can accept up to DSD256. In the settings, if I make it 44.1x128, the upsampling is to 5.6m, equivalent to DSD128. In any higher setting, the upsampling is to 6.1m and the DAC says no lock. Why don't I get to choose 11.2m in the settings as seen in the guides?

 

Link to comment
1 hour ago, clang said:

My DAC is the Brooklyn, not +. It can accept up to DSD256. In the settings, if I make it 44.1x128, the upsampling is to 5.6m, equivalent to DSD128. In any higher setting, the upsampling is to 6.1m and the DAC says no lock. Why don't I get to choose 11.2m in the settings as seen in the guides?

 

Do you have DoP enabled? The "SDM Pack" setting should be set to "None". If your Ubuntu install is recent, you should have "hwe" kernel version and thus 4.18 which probably has the generic DSD support for Mytek. You can check with "uname -r" on terminal.

 

If you are still on Ubuntu kernel older than 4.18, you have two alternatives, either installing my custom kernel and booting it up, or switching to Hardware Enablement (hwe) kernel.

 

I'm also assuming you have latest firmware installed in Mytek.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
On 3/27/2019 at 7:23 PM, Miska said:

[...]

I have no seen anybody measuring Gungnir, but the DAC chip itself is 18-bit part. How linear it is, is another question, possibly somewhat less, but hard to say how much. You can try different settings from 18-bit down to 14-bit with NS9 dither/noise-shaping. With 18/17-bit and likely even with 16-bit you are also fine with TPDF or Gauss1 dither.

[...]

 

 

Miska, I am trying various bit settings with Gungnir according to your instructions and it seems to me I prefer 18 or 16 for now. I also use ext2 filter which is my favorite with Gungnir and with very different Chord Mojo as well. Closed-form-M also sounds nice with Gungnir, very analogue and relaxed. Would you suggest these filters for Gungnir, or some others? Thank you.

Link to comment
7 hours ago, AnotherSpin said:

Miska, I am trying various bit settings with Gungnir according to your instructions and it seems to me I prefer 18 or 16 for now. I also use ext2 filter which is my favorite with Gungnir and with very different Chord Mojo as well. Closed-form-M also sounds nice with Gungnir, very analogue and relaxed. Would you suggest these filters for Gungnir, or some others? Thank you.

 

Filters are more up to personal preferences and type of music than a DAC. Both options are technically very good. Closed-form being strictly non-apodizing likely fits best with good quality source material. While ext2 being stricly apodizing is likely best all-rounder giving more consistent results regardless of source. This is also a good way to compare apodizing vs non-apodizing.

 

Since the DAC is R2R PCM ladder, modulator doesn't have a play here like with SDM DACs, so this is quite a bit simpler case. You can then choose between different dithers/noise-shapers, in this case it pretty much boils down to selection between TPDF, Gauss1 and NS9.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

1588403520_ScreenShot2019-03-30at5_47_19PM.thumb.png.2a9d8e0ba31380e0a9aa392d5b316d79.pngI have an STi7 with Roon and HQPe connected to an ultrarendu. This morning, I updated to versions 4.9 and 3.5.6, respectively, and now nothing works. Roon gives me a message that playback failed because it can't connect to HQP. I have tried rebooting and powering down everything multiple times. Attached is the HQP log. @Miska, any ideas? Thanks.

Link to comment
On ‎3‎/‎20‎/‎2017 at 9:55 AM, Miska said:

I prefer to buy DACs that allow me to use them as plain DACs, without forced DSP or at least as little as possible

 

I am sure the answer to my question is buried somewhere in the hundreds of pages of this thread. I know the T+A DAC8 DSD has been mentioned as the tool of choice many times. As have some ifi models. RME ADI2 DAC has a DSD direct mode that bypasses DSP (but will do the oversampling stuff prior to D/A conversion).

@Miska and others: which DACs would you recommend for use with HQ player and do you have an order of preference?

 

Link to comment
9 hours ago, mtcs said:

I have an STi7 with Roon and HQPe connected to an ultrarendu. This morning, I updated to versions 4.9 and 3.5.6, respectively, and now nothing works. Roon gives me a message that playback failed because it can't connect to HQP. I have tried rebooting and powering down everything multiple times. Attached is the HQP log. @Miska, any ideas? Thanks.

 

Looks like Rendu is not seeing the DAC...

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Hi @Miska, I hope you can help me for this problem: I have a gentoo player streamer based on raspberry (some boards for i2s connections) system with minimal version of gentoo and the daemon of HQPlayer. This streamer is connected by lan with a Intel i5 6500 with 16gb di ram, Windows 10, Roon and HQPlayer for audio render. I can play musix for 1 minute or two, then the daemon on raspberry closes (the gentoo and the raspberry works correctly) when I use roon Bridge on raspberry everything works perfectly.

This is the log of the networkaudiod. 

https://bpaste.net/show/839edff9c2dd

 

Any ideas? 

Thanks 

Paolo

Link to comment
On 3/29/2019 at 11:37 PM, Miska said:

 

Filters are more up to personal preferences and type of music than a DAC. Both options are technically very good. Closed-form being strictly non-apodizing likely fits best with good quality source material. While ext2 being stricly apodizing is likely best all-rounder giving more consistent results regardless of source. This is also a good way to compare apodizing vs non-apodizing.

 

 

Miska, I am sure this question was answered already, but what is the the difference between apodizing and non-apodizing filters in a plain, non-technical words for those, who do not understand special terminology?) In my set closed-form gives more natural sound, while ext2 gives very sweet, but slightly artificial (in comparison with closed-form) sound. And, which other filters are non-apodizing besides closed-form?

Link to comment
9 hours ago, paboccardi said:

Hi @Miska, I hope you can help me for this problem: I have a gentoo player streamer based on raspberry (some boards for i2s connections) system with minimal version of gentoo and the daemon of HQPlayer. This streamer is connected by lan with a Intel i5 6500 with 16gb di ram, Windows 10, Roon and HQPlayer for audio render. I can play musix for 1 minute or two, then the daemon on raspberry closes (the gentoo and the raspberry works correctly) when I use roon Bridge on raspberry everything works perfectly.

This is the log of the networkaudiod. 

https://bpaste.net/show/839edff9c2dd

 

I found one strange error from the log (about pthread_cond_timedwait()), but nothing else. And that only led to playback being stopped.

 

Maybe the binary is just not very well compatible with gentoo.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
49 minutes ago, Miska said:

 

I found one strange error from the log (about pthread_cond_timedwait()), but nothing else. And that only led to playback being stopped.

 

Maybe the binary is just not very well compatible with gentoo.

 

Thank you @Miska. I forgot to write that my dac is Denafrips Pontus and that I use in HQPlayer convolution filters (In addition to upsampling filters).

A couple of my friends use Gentooplayer with HQPlayer without problems.

 Do you know any particular settings for Dac Denafrips? Thanks 

Link to comment
45 minutes ago, Miska said:

 

Since all ADCs that produce something like 44.1k output are oversampling, there's a decimation filter to convert the higher rate down to lower one. In addition, these days is quite common that recording is made for example using 96k sampling rate and then converted down to 44.1k using another decimation filter.

 

Point of apodizing filter is to replace original decimation filter's impulse response with another one. This allows altering time- and frequency domain behavior of the original filter. You can get generally shorter "ringing" one, with something like poly-sinc-short or longer one with something like poly-sinc-ext2. Or you can change to a minimum-phase one.

 

Another maybe more important point is to clean up aliasing band at the highest frequencies that happen due to "modern" half-band ADC and DAW decimation filters that have pretty much no suppression at the fs/2 (Nyquist) frequency, and thus content exceeding that in the higher rate source data folds down into lower frequencies. For example original CD release of Pink Floyd DSOTM didn't have this dirt band at top, but the latest remaster does. This applies primarily for source content at 44.1/48k rate, and to much lesser extent to hires content.

 

closed-form is not really a filter, but instead interpolator, so it is non-apodizing due to that. poly-sinc-hb is a non-apodizing half-band filter a bit like the ones in modern ADC and DAC chips, but just better (higher precision and stop-band attenuation). minringFIR is another non-apodizing half-band filter with fairly slow roll-off, so it will let more images through too. You could compare poly-sinc-hb vs poly-sinc and minringFIR vs poly-sinc-short. Also poly-sinc-ext2 vs poly-sinc-xtr (xtr is quite a bit less apodizing than ext2).

 

When you use non-apodizing filter the results largely depend on what kind of decimation filter was used for the source content, because all it's faults come through as-is. While apodizing filters give more consistent performance across the board by correcting faults of the original decimation filter and giving same impulse/frequency response across the board regardless of source content.

 

Quote

Since all ADCs that produce something like 44.1k output are oversampling, there's a decimation filter to convert the higher rate down to lower one. In addition, these days is quite common that recording is made for example using 96k sampling rate and then converted down to 44.1k using another decimation filter.

Thanks for the great explanation Miska. Does this suggest that for 96khz (or higher) sample rate source material we would be better off using a non-appodizing filter, whereas for 44.1 (which has been generally down-sampled using a decimation filter) we'd be better off with one of the apodizing filters?

 

 

Owner of: Sound Galleries, High-End Audio Dealer, Monaco

Link to comment
2 hours ago, Miska said:

 

Since all ADCs that produce something like 44.1k output are oversampling, there's a decimation filter to convert the higher rate down to lower one. In addition, these days is quite common that recording is made for example using 96k sampling rate and then converted down to 44.1k using another decimation filter.

 

Point of apodizing filter is to replace original decimation filter's impulse response with another one. This allows altering time- and frequency domain behavior of the original filter. You can get generally shorter "ringing" one, with something like poly-sinc-short or longer one with something like poly-sinc-ext2. Or you can change to a minimum-phase one.

 

Another maybe more important point is to clean up aliasing band at the highest frequencies that happen due to "modern" half-band ADC and DAW decimation filters that have pretty much no suppression at the fs/2 (Nyquist) frequency, and thus content exceeding that in the higher rate source data folds down into lower frequencies. For example original CD release of Pink Floyd DSOTM didn't have this dirt band at top, but the latest remaster does. This applies primarily for source content at 44.1/48k rate, and to much lesser extent to hires content.

 

closed-form is not really a filter, but instead interpolator, so it is non-apodizing due to that. poly-sinc-hb is a non-apodizing half-band filter a bit like the ones in modern ADC and DAC chips, but just better (higher precision and stop-band attenuation). minringFIR is another non-apodizing half-band filter with fairly slow roll-off, so it will let more images through too. You could compare poly-sinc-hb vs poly-sinc and minringFIR vs poly-sinc-short. Also poly-sinc-ext2 vs poly-sinc-xtr (xtr is quite a bit less apodizing than ext2).

 

When you use non-apodizing filter the results largely depend on what kind of decimation filter was used for the source content, because all it's faults come through as-is. While apodizing filters give more consistent performance across the board by correcting faults of the original decimation filter and giving same impulse/frequency response across the board regardless of source content.

 

 

While very well stated, it nonetheless is a very technical explanation that does not use  common audio description language.  It would be great to hear the HQPlayer filters etc explained in Stereophile or The Absolute Sound type audio language.  

Link to comment
1 hour ago, DancingSea said:

While very well stated, it nonetheless is a very technical explanation that does not use  common audio description language.  It would be great to hear the HQPlayer filters etc explained in Stereophile or The Absolute Sound type audio language.  

 

I'm probably not mentally capable of writing such... :D And since everybody hears things a bit differently I don't see such so valuable either. I think the best is just for everyone to listen themselves and establish their own view how the different filters sound.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
3 hours ago, Geoffrey Armstrong said:

Thanks for the great explanation Miska. Does this suggest that for 96khz (or higher) sample rate source material we would be better off using a non-appodizing filter, whereas for 44.1 (which has been generally down-sampled using a decimation filter) we'd be better off with one of the apodizing filters?

 

No, there's no particular advantage of using non-apodizing filters, unless you are looking the hear faults of the original filters. The effect of apodizing filters is just smaller with hires. As you go up in the rate, the ADC decimation filters get gradually worse, but on the other hand the content level at those frequencies is also lower. At 176.4k and higher rate material there's likely no aliasing left anymore, so cleaning up for that is not important anymore. You still have the potential other effects too, but to lesser extent.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
4 hours ago, paboccardi said:

Thank you @Miska. I forgot to write that my dac is Denafrips Pontus and that I use in HQPlayer convolution filters (In addition to upsampling filters).

A couple of my friends use Gentooplayer with HQPlayer without problems.

 Do you know any particular settings for Dac Denafrips? Thanks 

 

Hard to say, you could also check HQPlayer side logs for errors.

 

I don't have a Denafrips to test with, but based on what I know about it's construction, you should be fine at DSD256 from ASDM7 modulator and filter of your choice.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
1 hour ago, Miska said:

 

No, there's no particular advantage of using non-apodizing filters, unless you are looking the hear faults of the original filters. The effect of apodizing filters is just smaller with hires. As you go up in the rate, the ADC decimation filters get gradually worse, but on the other hand the content level at those frequencies is also lower. At 176.4k and higher rate material there's likely no aliasing left anymore, so cleaning up for that is not important anymore. You still have the potential other effects too, but to lesser extent.

 

1 hour ago, Miska said:

 

Thanks. Do you see any reason at all for using different filters for higher-rez material?

Owner of: Sound Galleries, High-End Audio Dealer, Monaco

Link to comment
3 hours ago, DancingSea said:

 

While very well stated, it nonetheless is a very technical explanation that does not use  common audio description language.  It would be great to hear the HQPlayer filters etc explained in Stereophile or The Absolute Sound type audio language.  

 

Maybe an easier way to do what you want than speaking generally about Stereophile or TAS language is to ask specific questions, if you can come up with them.

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
7 hours ago, Miska said:

 

Since all ADCs that produce something like 44.1k output are oversampling, there's a decimation filter to convert the higher rate down to lower one. In addition, these days is quite common that recording is made for example using 96k sampling rate and then converted down to 44.1k using another decimation filter.

 

Point of apodizing filter is to replace original decimation filter's impulse response with another one. This allows altering time- and frequency domain behavior of the original filter. You can get generally shorter "ringing" one, with something like poly-sinc-short or longer one with something like poly-sinc-ext2. Or you can change to a minimum-phase one.

 

Another maybe more important point is to clean up aliasing band at the highest frequencies that happen due to "modern" half-band ADC and DAW decimation filters that have pretty much no suppression at the fs/2 (Nyquist) frequency, and thus content exceeding that in the higher rate source data folds down into lower frequencies. For example original CD release of Pink Floyd DSOTM didn't have this dirt band at top, but the latest remaster does. This applies primarily for source content at 44.1/48k rate, and to much lesser extent to hires content.

 

closed-form is not really a filter, but instead interpolator, so it is non-apodizing due to that. poly-sinc-hb is a non-apodizing half-band filter a bit like the ones in modern ADC and DAC chips, but just better (higher precision and stop-band attenuation). minringFIR is another non-apodizing half-band filter with fairly slow roll-off, so it will let more images through too. You could compare poly-sinc-hb vs poly-sinc and minringFIR vs poly-sinc-short. Also poly-sinc-ext2 vs poly-sinc-xtr (xtr is quite a bit less apodizing than ext2).

 

When you use non-apodizing filter the results largely depend on what kind of decimation filter was used for the source content, because all it's faults come through as-is. While apodizing filters give more consistent performance across the board by correcting faults of the original decimation filter and giving same impulse/frequency response across the board regardless of source content.

 

 

I have a feeling I understand something from this, and will dig deeper into it with more knowledge of technical things somewhere later. I did comparisons of closed-form-M with other non-apodizing filters, such as poly-sinc-hb, minringFIR with classical, chamber and jazz material through Schiit Gungnir. Closed-form-M gives more well-rounded, full, natural and comfortable sound. Poly-sinc-hb, minringFIR (and ext2) give more polished and articulated but thinner sound, which seem to be slightly fatiguing in a long run.

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...