HQ Player

[rickca]

6 hours ago, Miska said:

This is done using the input feature of HQPlayer Embedded.

Where can I read more about this input feature?

[giordy60]

I also tried the closed form 16m on sdm (I stop at 128)
but I prefer the poly-sinc-xtr-mp -2s filter, with DSD7 256 + fs has a bit more detail. 

 

[jimdukey]

I've been trying some new things I read about here.

How do you differentiate the various Modulators?

All from DSD 5 up to to the 256/512 give me DSD 6.14 on Mytek and Lampi Dacs.

DSD7256+fs gives DSD 5.6, no 6.14.

I thought that DSD 256/512 Modulators were only for those Rates.

Miska said he updated DSD5 Modulator recently.

What do they all do?

I've been using ASDM7.

 

Closed Form 16M gives DSD 5.6 on DSD Files, But DSD 64 upsampling PCM Files.

All Poy-sinc Filters give DSD 6.14 on everything.

There's so much to compare, mix and match.

[shadowlight]

10 hours ago, Miska said:

 

As time goes on, Microsoft keeps improving WSL in each major Win10 upgrade. So over time the UPnP functionality may start working. But I doubt that for example ALSA things would start working any time soon. So things that use "generic unix" are more likely to become available and things that use hardware interfaces (sound, graphics) are less likely to become available soon.

 

I just updated the Windows instance with WSL and the problem of Rygel bombing out is still there.

[Miska]

...maybe in a year or two...

 

[Jamesroy]

23 hours ago, Le Concombre Masqué said:

?Great news!

How about the PS Audio Stellar Gain Cell DAC

[Le Concombre Masqué]

On 8/28/2018 at 9:48 PM, Miska said:

Note, I also did some re-tuning of DSD5 modulator somewhat recently, worth checking out again if you like DSD5v2.

Wonder why you re-tuned it while keeping the name but anyway I like it. "Raw power" comes to my mind to qualify it ; it's a keeper and the new selection in my settings.

Right or wrong.

 

Poly sinc short/DSD5 to 128 are my new settings for my TEAC 501

Replacing xtr...mely long...

 

Thank you for sharing the results of the intellectual challenges you take exploring the various algorithms options that account for this or that DAC's signature : the definite cure to audiophile craziness ; better to change by a click rather than collecting DACs !

 

There's the set of tracks chosen, the aesthetic that suits my mood at a period, expectation biases, including my vision of some rational too ; anyway, to whom it may concerns, I report here:

 

I had not fiddled with filters since I designed convolution filters via REW/RePhase. Funny thing is that now that my system is as minimum phase as I could make it, I prefer linear phase filters. Yes, compared to the mp version the linear was ear scorching on this or a that track but I'm going to accept it as more revealing. Until I would really like to tame : mqa mp will certainly prove useful once I'll be able to use it on 24/192 popular music ; but since I heard it mess, even in the linear version, and thicken Serkin playing Beethoven, I don't think I'll use it unless I want to beef fatten etc in order to make the recording/mastering suit my taste.

Closed form 16 M has never been the winner with hi res (88.2 only for me at the moment) but I think I will launch it manually systematically with classical RBCD, at least in the near future !

incumbent xtr 2s might be worth checking with RBCD when CF is not overwhelming but the difference with poly sinc short might not justify to bother 

 

MinringFIR could have taken the place of Poly sinc short in my settings if your set of rules allowed it with n48 hires  yet. Wish it will be too !

I have expectation bias that, being extremely short, it's going to roll off with RBCD and i actually thought here and there that it was slow and fat, but without the disbelief brought my mqa, but I marvelled at how good acoustic guitars sounded with that filter, even with RBCD : definitely wish your set of rules for that filter will allow me to use it with 96 and 192 hires !

 

 

 

[copy_of_a]

4 hours ago, Le Concombre Masqué said:

MinringFIR could have taken the place of Poly sinc short in my settings if your set of rules allowed it with n48 hires  yet. Wish it will be too !

not sure what you are referring to...? Are you talking about 48x->DSD?

Generally minringFIR can be used for any source rate, but due to its design is only suited to resample by power of 2.

So 48kHz sources can be upsampled to 96, 192 etc. PCM or to 3.072 (etc.) DSD ... but not to 88.2kHz, 176.4kHz (etc.) PCM or 2.822 (etc.) DSD.

minringFIR is non-apodizing; with 44.1kHz soruces the transition starts at around 20kHz. However, due to its special design the ringing is very, very low (short). Personally I use minringFIR-mp all the time for everything (I used to switch to polysinc for classical music but meanwhile I just stay with minringFIR-mp).

 

[Le Concombre Masqué]

19 minutes ago, copy_of_a said:

not sure what you are referring to...? Are you talking about 48x->DSD?

Generally minringFIR can be used for any source rate, but due to its design is only suited to resample by power of 2.

So 48kHz sources can be upsampled to 96, 192 etc. PCM or to 3.072 (etc.) DSD ... but not to 88.2kHz, 176.4kHz (etc.) PCM or 2.822 (etc.) DSD.

minringFIR is non-apodizing; with 44.1kHz soruces the transition starts at around 20kHz. However, due to its special design the ringing is very, very low (short). Personally I use minringFIR-mp all the time for everything (I used to switch to polysinc for classical music but meanwhile I just stay with minringFIR-mp).

 

yes ; can't do 48x to DSD with my DAC that does not do 3.072 (etc.) DSD ;  minringFIR simply does not start with 48x source (but if I choose to output PCM and though it's appealing once in a while I rule out this option with my DAC that globally works better outputting DSD 128.

 

I should probably stick to Miska's terminology :

 

My DAC doesn't accept DSD data at 48k-base DSD rates

Thus, I wish Miska allows rates conversions to a different DSD rate family for non apodizing filters such as minring and mqa

[Le Concombre Masqué]

there are comments about filters but none about integrators....

 

I take the chance that Miska declares it can't be but placebo effect but I rate :

 

CIC>FIR>IIR ; I'm used to hearing via FIR (chosen for my DAC has FIR filters for DSD...)

https://www.stereophile.com/content/ayre-acoustics-qa-9-usb-ad-converter might set ground for CIC (moving average)

I'm back to DSD5V2, yield the most credible presence of a RBCD harpsichord with CF 16M

 

 

[Miska]

6 hours ago, Le Concombre Masqué said:

Thus, I wish Miska allows rates conversions to a different DSD rate family for non apodizing filters such as minring and mqa

 

Just a small note, poly-sinc-mqa is apodizing. There are few non-apodizing filters though, like poly-sinc-hb, minringFIR and closed-form.

 

One benefit of apodizing filters is that you get more consistent sonic performance regardless of what kind of filter was used to produce the recording. While non-apodizing will expose this directly.

 

As typical these days, newer recordings are commonly worse and benefit more from apodizing filters.

[LoryWiv]

12 hours ago, Le Concombre Masqué said:

 

Closed form 16 M has never been the winner with hi res (88.2 only for me at the moment) but I think I will launch it manually systematically with classical RBCD, at least in the near future !

 

I am also enjoying Closed form -16M w/RBCD and even DSD64. I am outputting to DSD256. Placebo or not, I hear more presence,  micro-detail and instrument separation / stage versus poly-sinc-xtr-mp -2s and it keeps CPU / GPU utilization in the 12-15% range.

 

I do wish closed form-16M could be compatible with 24/96 and above. If so I would probably use it as my default setting.

[Le Concombre Masqué]

1 hour ago, LoryWiv said:

 

I am also enjoying Closed form -16M w/RBCD and even DSD64. I am outputting to DSD256. Placebo or not, I hear more presence,  micro-detail and instrument separation / stage versus poly-sinc-xtr-mp -2s and it keeps CPU / GPU utilization in the 12-15% range.

 

I do wish closed form-16M could be compatible with 24/96 and above. If so I would probably use it as my default setting.

with DSD 64 only integrator and moduler matter

[Sloop John B]

5 hours ago, Miska said:

 

Just a small note, poly-sinc-mqa is apodizing. There are few non-apodizing filters though, like poly-sinc-hb, minringFIR and closed-form.

 

One benefit of apodizing filters is that you get more consistent sonic performance regardless of what kind of filter was used to produce the recording. While non-apodizing will expose this directly.

 

As typical these days, newer recordings are commonly worse and benefit more from apodizing filters.

 

I’ve always wondered what is the best filter for brickwalled CDs which in reality (well my reality anyway) are irretrievable, but perhaps a certain filter un- nastys them more than others, do you have an opinion on this ( subjective or objective, I don’t mind). 

 

.sjb 

[jimdukey]

"As typical these days, newer recordings are commonly worse."

What recordings?

SF and Seattle Syms, among others make beautiful sounding recordings.

The ECM albums I have are also wonderfully recorded.

They aren't the only ones either.

[AnotherSpin]

19 minutes ago, jimdukey said:

"As typical these days, newer recordings are commonly worse."

What recordings?

SF and Seattle Syms, among others make beautiful sounding recordings.

The ECM albums I have are also wonderfully recorded.

They aren't the only ones either.

 

Probably, Miska had in mind popular music recordings.

[juanitox]

Hi Miska do you plan to build an Auto-filter fonction ( with preset analyser) to adapt every kind of music/recording with your variants of filters , it will be easy for us ??

[Miska]

5 hours ago, Sloop John B said:

I’ve always wondered what is the best filter for brickwalled CDs which in reality (well my reality anyway) are irretrievable, but perhaps a certain filter un- nastys them more than others, do you have an opinion on this ( subjective or objective, I don’t mind). 

 

Any of the apodizing filters (most of poly-sinc family for example) are good for the purpose, depending on your personal preferences.

 

[Miska]

2 hours ago, jimdukey said:

"As typical these days, newer recordings are commonly worse."

What recordings?

SF and Seattle Syms, among others make beautiful sounding recordings.

The ECM albums I have are also wonderfully recorded.

They aren't the only ones either.

 

I don't know about those, but as an example first Pink Floyd CD releases were better in certain technical ways compared to the most recent remasters.

 

Many of the latest ADCs and mastering tools have some problematic characteristics. Those can be mostly fixed during playback time though. Hires PCM (>= 88.2 kHz) or DSD material doesn't have these problems though.

 

[jimdukey]

 Hires PCM (>= 88.2 kHz) or DSD material doesn't have these problems though.

 

That's good to know!

[Le Concombre Masqué]

I recall why I did not let CIC as integrator : stutters with DSD files. However, I insist : when usable it's even better than FIR : ie deeper firmer bass and when you would wonder if it's not too smooth to be honestly good, remarkable transients show off. At least it happened this lunch time and a musician friend heard the same.

anybody else has fiddled with integrators ?

[louisxiawei]

16 hours ago, Le Concombre Masqué said:

I recall why I did not let CIC as integrator : stutters with DSD files. However, I insist : when usable it's even better than FIR : ie deeper firmer bass and when you would wonder if it's not too smooth to be honestly good, remarkable transients show off. At least it happened this lunch time and a musician friend heard the same.

anybody else has fiddled with integrators ?

CIC is a bit unstable and can cause stutter in most cases when I upsample high rate DSD files (128,256) to dsd 512. 

 

I remember Miska told me that dual/quad core is stable for CIC while 6+ core may easily cause problem.

 

[Le Concombre Masqué]

1 hour ago, louisxiawei said:

CIC is a bit unstable and can cause stutter in most cases when I upsample high rate DSD files (128,256) to dsd 512. 

 

I remember Miska told me that dual/quad core is stable for CIC while 6+ core may easily cause problem.

 

I found out (was too late for my better's half ears to test loud) that I should be able to do 64 to 128 if convolution engine is add not save.

 

I'll then test SQ wise CIC/overlap-add vs FIR/overlap-save

 

Miska wrote once (quoted from memory) that add and save achieve about the same results in different ways and that add should be preferred for its lesser load on cpu ; however I preferred save when I explored that option.

 

My Mac (running Windows server) is quad core.

 

I guess that if you take the chance to have to switch integrators with high rate DSD files it is because you too appreciate CIC : good to know ! Many among us have reported and debated the merits of various filters but googling or searching CA brought back no opinion on integrators...

 

 

[Miska]

5 hours ago, louisxiawei said:

CIC is a bit unstable and can cause stutter in most cases when I upsample high rate DSD files (128,256) to dsd 512. 

 

I remember Miska told me that dual/quad core is stable for CIC while 6+ core may easily cause problem.

 

 

I think I need to look into optimizing it a bit more. Overall it shouldn't be much different from the FIR/IIR.

 

[Le Concombre Masqué]

2 hours ago, Miska said:

 

I think I need to look into optimizing it a bit more. Overall it shouldn't be much different from the FIR/IIR.

 

Do you expect it to sound a bit less good ?

 

Still... As of now, I have decided to let CIC as integrator and overlap-add as convolution engine (for convenience ; a vinyl fan for long, hires (and sacd rips) are overrepresented in my collection ; if I was rbcd only I would keep overlap-save)

 

This morning, with CIC and DSD sources, I marvelled at firmer tighter everything, including bass, meaningful details and spatial cues 

 

CF16M/DSDV2 + CIC is a magic combo for classical/best of breed RBCD

When I'm not overwhelmed by this combo for RBCD I first switch to xtr 2s ; then minringFIR and mqa mp if I have harshness to tame.

 

poly sinc short being my base settings (once again hires is overrepresented in my digital collection) but I'm looking forward being able to use mqa and minringFIR with 96 and 192 with my DAC that does not allow x48 DSD rates