Sound Galleries SGM 2015

[asiufy]

I've been playing with the new 3.16 poly-sync-xtr filter for the last 2 days on the Aqua Formula, and so far, I still prefer closed-form.
I'll do a bit more back/forth, but there *is* a significant difference.

cheers,
alex

 

[Steve Williams]

Alexandre
What do you mean by closed form

 

[Steve Williams]

Alexandre
What do you mean by closed form

 

[microstrip]

I've been playing with the new 3.16 poly-sync-xtr filter for the last 2 days on the Aqua Formula, and so far, I still prefer closed-form.(...)

Can you translate it, please? What is closed-form?

 

[asiufy]

hahaha, sorry for the techy-speak

That is a filter setting in HQplayer, and that is one of the parameters that influence the overall presentation given by HQplayer, and consequently, the SGM.

Ed traditionally has been using a filter in the poly-sync-* family, and I guess I was curious enough to go through all the filters (20-odd I think?), and picked one of the the closed-form-* filters as sounding considerably better than the poly-sync-* that Ed has recommended. That turned out to be valid in more than my setup, as Ed has verified closed-form sounds better in other setups as well. At least with the Aqua Formula DAC.

Version 3.16 of HQplayer introduced a new filter in the poly-sync-* family, called poly-sync-ext. Don't ask me what's new with it, or what it does differently, I just listen to stuff As I said, I tried it on for a few, but I still thought my previous fave is still my fave!

Hope this clears things up...

 

[audio.bill]

Can you translate it, please? What is closed-form?

Mike Moffat who designed the original Theta DACs and more recently the Schiit Yggdrasil DAC is a strong proponent of closed form digital filtering, and you can read his further thoughts on the Schiit website. Here's an excerpt with a basic explanation of a closed form digital filter:

Closed-Form Digital Filter Preserves Original Samples

Most DACs simply use the stock digital filters embedded in their D/A converters. But even the most sophisticated ones, using their own digital filter algorithms, don’t have what Yggdrasil has—a time- and frequency-domain optimized digital filter with a true closed-form solution. This means it retains all the original samples, performing a true interpolation. This digital filter gives you the best of both NOS (all original samples retained) and upsampling (easier filtering of out-of-band noise) designs.

Hope this helps!

 

[asiufy]

Cool, thanks for the explanation!

 

[Mike Lavigne]

hahaha, sorry for the techy-speak

That is a filter setting in HQplayer, and that is one of the parameters that influence the overall presentation given by HQplayer, and consequently, the SGM.

Ed traditionally has been using a filter in the poly-sync-* family, and I guess I was curious enough to go through all the filters (20-odd I think?), and picked one of the the closed-form-* filters as sounding considerably better than the poly-sync-* that Ed has recommended. That turned out to be valid in more than my setup, as Ed has verified closed-form sounds better in other setups as well. At least with the Aqua Formula DAC.

Version 3.16 of HQplayer introduced a new filter in the poly-sync-* family, called poly-sync-ext. Don't ask me what's new with it, or what it does differently, I just listen to stuff As I said, I tried it on for a few, but I still thought my previous fave is still my fave!

Hope this clears things up...

and evidently some filters use more processing resources than others, so it can limit higher resolution formatted files. whether this is a hardware, or HQ Player software issue, I don't know.

 

[Taiko Audio]

and evidently some filters use more processing resources than others, so it can limit higher resolution formatted files. whether this is a hardware, or HQ Player software issue, I don't know.

From the designer: The new Poly-sinc-xtr is a "long" filter which emphasizes the frequency domain where the previous mostly preferred Poly-sinc-shrt is a "short" filter emphasizing the time domain. Keep in mind that time and frequency domains are inversely related. The new Poly-sinc-xtr filter has a stop-band attenuation spec of an insane -240 dB which would be equivalent to 40-bit PCM. There are no dac's supporting 40 bit PCM samples, so any benefits of this aspect would mainly apply to higher rate DSD. For comparison, the filters of the Chord Dave, ESS Sabre and other "high tech dacs" are at about -120 dB sba. The penalty for all this is that it requires an insane amount of processing power. Poly-sinc-xtr is about 5 times as heavy as Poly-sinc-shrt. It was named -xtr (for extreme) in stead of a more obvious -lng or -long because its not just a "long" filter but also includes the insane stop-band attenuation. At this point in time it is more a technology demonstration of insane numbers, as no commercial processor exists that can run this at the highest (being DSD) sampling rates for prolonged times. There is however a lighter version available capable running all rates and there is room to develop filters in between.

As having been an IT professional for the past 20 years or so it always makes me smile we can develop filters for reconstructing audio waveforms from a mere 16/44.1 source that requires as much processing power as for example our mail servers serving 30.000 e-mail boxes.

 

[Elberoth]

IBM Blue Gene ... coming soon to audiophile listening rooms

 

[Taiko Audio]

Does that come with an usb output?

 

[hifial]

Thanks. Do you know which room at Axpona will have the server? And with which DAC(s) it will be demoed?

Oops, sorry I meant LA, not Axpona. I have two audio industry buddies that will be at Axpona. Merrill Audio and Distinctive Stereo.

 

[Sablon Audio]

From the designer: The new Poly-sinc-xtr is a "long" filter which emphasizes the frequency domain where the previous mostly preferred Poly-sinc-shrt is a "short" filter emphasizing the time domain. Keep in mind that time and frequency domains are inversely related. The new Poly-sinc-xtr filter has a stop-band attenuation spec of an insane -240 dB which would be equivalent to 40-bit PCM. There are no dac's supporting 40 bit PCM samples, so any benefits of this aspect would mainly apply to higher rate DSD. For comparison, the filters of the Chord Dave, ESS Sabre and other "high tech dacs" are at about -120 dB sba. The penalty for all this is that it requires an insane amount of processing power. Poly-sinc-xtr is about 5 times as heavy as Poly-sinc-shrt. It was named -xtr (for extreme) in stead of a more obvious -lng or -long because its not just a "long" filter but also includes the insane stop-band attenuation. At this point in time it is more a technology demonstration of insane numbers, as no commercial processor exists that can run this at the highest (being DSD) sampling rates for prolonged times. There is however a lighter version available capable running all rates and there is room to develop filters in between.

As having been an IT professional for the past 20 years or so it always makes me smile we can develop filters for reconstructing audio waveforms from a mere 16/44.1 source that requires as much processing power as for example our mail servers serving 30.000 e-mail boxes.

I spent a couple of hours yesterday playing with the new xtr filters and was able to eek out a better sound at 128DSD using the most processor hungry iteration than I was able to achieve from the less demanding 2S version filter could achieve with 512DSD. This listening was done using a Tidal stream to the SGM and then into the T+A dac, which many feel only comes to alive with 512 frequency. As Carlos (CMarin) has observed earlier in this thread, software seems to be the gift that keeps giving just now.

 

[wisnon]

Mike Moffat who designed the original Theta DACs and more recently the Schiit Yggdrasil DAC is a strong proponent of closed form digital filtering, and you can read his further thoughts on the Schiit website. Here's an excerpt with a basic explanation of a closed form digital filter:

Closed-Form Digital Filter Preserves Original Samples

Most DACs simply use the stock digital filters embedded in their D/A converters. But even the most sophisticated ones, using their own digital filter algorithms, don’t have what Yggdrasil has—a time- and frequency-domain optimized digital filter with a true closed-form solution. This means it retains all the original samples, performing a true interpolation. This digital filter gives you the best of both NOS (all original samples retained) and upsampling (easier filtering of out-of-band noise) designs.

Hope this helps!

Closed form may be better for PCM up sampling but likely not for high rate DSD, where the poly-sinc family rules.

 

[wisnon]

IBM Blue Gene ... coming soon to audiophile listening rooms

LoL Already outdated. Cant run DSD1million upsampling.

 

[Steve Williams]

what is the function of these filters? I hate to sound dumb but as I read everything written am I to infer that they serve as tone controls

 

[asiufy]

I can only talk about the practical, subjective part of it, but they are so much more than tone controls. The options within HQplayer, and the filters being one of them, describe how the digital signal is going to handled. You can do less or more with it, and use different methods of "massaging" the bits. Some will add leading edge to transients, some will blur them a bit, some will increase soundstage width, some will add depth, etc... Of course, there are always trade offs, and there's why picking one setting isn't so clear cut, and tends to be very system-dependant (and DAC dependant, of course).

Here's hoping Roon integrates HQplayer fully into its interface, so we're able to tweak HQplayer settings from the comfort of Roon's UI!


cheers,
alex

 

[Steve Williams]

I can only talk about the practical, subjective part of it, but they are so much more than tone controls. The options within HQplayer, and the filters being one of them, describe how the digital signal is going to handled. You can do less or more with it, and use different methods of "massaging" the bits. Some will add leading edge to transients, some will blur them a bit, some will increase soundstage width, some will add depth, etc... Of course, there are always trade offs, and there's why picking one setting isn't so clear cut, and tends to be very system-dependant (and DAC dependant, of course).

Here's hoping Roon integrates HQplayer fully into its interface, so we're able to tweak HQplayer settings from the comfort of Roon's UI!


cheers,
alex

Thanks Alex

In the great scheme of things aren't they just tone controls of sort

 

[Steve Williams]

I can only talk about the practical, subjective part of it, but they are so much more than tone controls. The options within HQplayer, and the filters being one of them, describe how the digital signal is going to handled. You can do less or more with it, and use different methods of "massaging" the bits. Some will add leading edge to transients, some will blur them a bit, some will increase soundstage width, some will add depth, etc... Of course, there are always trade offs, and there's why picking one setting isn't so clear cut, and tends to be very system-dependant (and DAC dependant, of course).

Here's hoping Roon integrates HQplayer fully into its interface, so we're able to tweak HQplayer settings from the comfort of Roon's UI!


cheers,
alex

Thanks Alex

In the great scheme of things aren't they just tone controls of sort

 

[Taiko Audio]

Thanks Alex

In the great scheme of things aren't they just tone controls of sort

Looking at it this way they would be more akin to distortion and noise levels.