Is *signal purity* BS?

[Gerner]

Hi Gerner,

I just spent half an hour penning a long reply only for it to go missing when I came to post it. Suffice to say I won't be wasting another half an hour of my life!

Still I will say that to me all processing leaves a footprint. I prefer to live without as much DSP as I can. Sure it can work wonders for stereo imaging but I'd rather trawl through old vinyl until 3 in the morning than hear the layering these days. I've yet to hear a product that allows me to do both but I live in hope.

No matter what was lost in your writing, I read what you mean in your post here.

And yes....it's the fact actually.

Cheers pal.

Gerner

 

[Gerner]

Fast posters here...

I forgot to let the quote remain.. will remember.

Gerner

 

[ShinOBIWAN]

About a device that is rather popular for EQ, the DEQX:

The latter is about the 16 bit processing the DEQX would apply which really is nothing to hope for results. Even with 24 bits and a digital volume only, the sound gets so much degraded that there's *really* no hope for a better net result.
The least it would need is 32 bit digital input, hence a player that would output that. So we talked about adjusting XX to that, which just can be done.
Note for those who come up with it : E.g. Foobar can do it.
... But what to do if you don't like the sound of Foobar at all (as how we percieved that at that time).

I use 64bit precision filtering with sample rate capability upto 192Khz. Sadly its PC based and fairly convoluted so most music lovers look at it and say "Looks good, makes sense but no thanks". Hopefully in the short term future a more user friendly commercial box will be introduced with comparable specifications but for now its the only way. Units such as the Tact and DEQX are excellent introductions into the arena but aren't as transparent. This field is moving fast but for now it takes some technical know how to get the best results which makes it something of an oddball. The commercial units that are easily accessible are still limited by the antiquated CD playback standards and underpowered DSP with similarly cutback software.

In short there's room for improvement but better quality solutions are to be had with effort and willingness.

 

[Gerner]

http://www.zerogain.com/forum/showthread.php

Hahaha...we know each other more than you maybe think:

The RAAL ribbons...

Aleksander is one of my best friends. He and I are in some very serious developing phase that might show the world something...or maybe not?

Anyhow I think he will regard you as serious using his ribbons for your choice.

Cheers.

Gerner

 

[Gerner]

I use 64bit precision filtering with sample rate capability upto 192Khz. Sadly its PC based and fairly convoluted so most music lovers look at it and say "Looks good, makes sense but no thanks". Hopefully in the short term future a more user friendly commercial box will be introduced with comparable specifications but for now its the only way. Units such as the Tact and DEQX are excellent introductions into the arena but aren't as transparent. This field is moving fast but for now it takes some technical know how to get the best results which makes it something of an oddball. The commercial units that are easily accessible are still limited by the antiquated CD playback standards and underpowered DSP with similarly cutback software.

In short there's room for improvement but better quality solutions are to be had with effort and willingness.

Shin... see you grasp the sucker here... fine

Well...still my doors are open. I never let down ideas unless they were idiotic, and this isn't.

Keep me briefed.

Ciao

Gerner

 

[ShinOBIWAN]

How were the speakers intended to be placed for the sweet spot, zero phase turn true piston behaviour, and how is a mike compared to our ears???? Big question.
The mike is not a human ear or brain, that's for sure. Just look at the flatness performance of the mike and compare it to the two mikes we have been born with..... Jihhaaaa apples and pears.

If this DSP'ed room correction should work out properly, it should be able to copy our brains perception. And still I don't believe it can, neither the proccessing measuring CPU can handle that.

Psychoacoustics is still a relatively under developed branch of science.

But some important discoveries have been made, most of which I have little understanding of. However a few of these psychoacoustic behaviours have been implemented into the filter generation program I use. By no means perfect but it shows that the advanced is spanning interesting concepts and is at least concious that our ears have to be pampered to. One thing I haven't done is compare the filters with and without this processing in place. Might be interesting to see just what audible differences this makes and more importantly whether the music sounds better.

 

[Gerner]

Psychoacoustics is still a relatively under developed branch of science.

But some important discoveries have been made, most of which I have little understanding of. However a few of these psychoacoustic behaviours have been implemented into the filter generation program I use. By no means perfect but it shows that the advanced is spanning interesting concepts and is at least concious that our ears have to be pampered to. One thing I haven't done is compare the filters with and without this processing in place. Might be interesting to see just what audible differences this makes and more importantly whether the music sounds better.

Shin ...you are far ahead of me here...report report report..along the way....

I do look forward to that a *system* can copy our brains/ears. Could even be the case in my remaining lifetime...

I know a person who is working in the digital domains *imaginary' world. His ideas, if true, addresses this.

If not within expected lifetime..well.....I care no more... and if I become 100+ years old my brain and ears will be hopefully be able to appreciate :mana:


Gerner

 

[ShinOBIWAN]

http://www.zerogain.com/forum/showthread.php

Hahaha...we know each other more than you maybe think:

The RAAL ribbons...

Aleksander is one of my best friends. He and I are in some very serious developing phase that might show the world something...or maybe not?

Anyhow I think he will regard you as serious using his ribbons for your choice.

Cheers.

Gerner

Small world. I'm a big fan of his 140-15d ribbon and is preferred over the dome tweeters I've used recently. I made the mistake of asking him about his thoughts on dome tweeters and its probably best I not repeat his reply.

He likes Tool so he's alright in my book.

 

[Tenson]

Not Toole?

 

[Gerner]

Small world. I'm a big fan of his 140-15d ribbon and is preferred over the dome tweeters I've used recently. I made the mistake of asking him about his thoughts on dome tweeters and its probably best I not repeat his reply.

He likes Tool so he's alright in my book.

Hahaha shin...so you got Aleksandars cookbook for controling magnetism and stray flux...

He is a lovely guy and I typically call him Nikola Tesla the 2nd. You know why if you met him.

I shall see him 1st of january and and again together with Bert from BD-Design 3 days in middle of January.

We have quite som big plans togeher. But it's plans in our heads and maybe far too early to start to rave about. But they are crazy those ideas.

I will bring him your greetings, well if it's ok?

He and me a month ago:



I appoligize for not commenting your beautiful loudspeakers. They are damn beautiful but not only that. I can see you have some pedigree in speaker building as well.

Attached is a 10 years old design of mine, totally made by own hands:

500 kg's of a monster 4,0 x 1,3 x 0,5 m's.
D'Apolito mid/high section and 10" driver front loaded W-bass horns.
Time alignment for the bass oooops! Not possible at that moment.
Filter might look complicated but it's not. Battery biased caps 36 Vdc. Only 5 components in total in series with the drivers. The rest is just impedance notches.

A pic of Aleksander (right) and me (left). And a pic of the 500 kg monsters.

Happy New Year

Gerner

 

[Gerner]

I use 64bit precision filtering with sample rate capability upto 192Khz. Sadly its PC based and fairly convoluted so most music lovers look at it and say "Looks good, makes sense but no thanks". Hopefully in the short term future a more user friendly commercial box will be introduced with comparable specifications but for now its the only way. Units such as the Tact and DEQX are excellent introductions into the arena but aren't as transparent. This field is moving fast but for now it takes some technical know how to get the best results which makes it something of an oddball. The commercial units that are easily accessible are still limited by the antiquated CD playback standards and underpowered DSP with similarly cutback software.

In short there's room for improvement but better quality solutions are to be had with effort and willingness.

Hi Shin..

So you have your room correction processing going on on a PC?...64 bit! XP/Vista/? 64 bit MOBO?
So it's dedicated only to support the room correction processors?

But isn't the output just still 16 bit? Or?

Tell me is it something you develope yourself or what?

Many questions, sorry. I have not heard of this approach before.

I can understand that you need at least 64 bit's to master room correction and EQ.

Gerner

 

[Stereo Mic]

Gerner,

you should find that Shin is working in Acourate - a series of digital programs that run from crossover generation to room correction.

I'm no digital engineer but AFAIK it works either in 64 bit floating point or 32 bit fixed point, but either should be sufficient for double precision of CD based media. The key with any of this technology of course is the programming - one slip or oversight (especially when converting from floating point I am led to understand) can ruin the end result.

I haven't tried Acourate so can't vouch for the quality of the programming nor the subjective results but my understanding is that the technology employed isn't that much more advanced than the Tact and DEQX alternatives - especially when used in conjunction with CD quality formats. Maybe Shin can enlighten us there - but Tact certainly works with 32 bit processing - enough to achieve double precision (or an effective doubling of the dynamic range) internally.

From what I have read, it all comes down to the talent of the programmer rather than the chip specifications once we have reached that kind of level of processing. Dr Radomir Bozovic of Tact was previously heavily involved in the Psychoacoustics field - saying that it seemed to have little effect on the programming of the Tact RCS.

 

[Gerner]

Gerner,

you should find that Shin is working in Acourate - a series of digital programs that run from crossover generation to room correction.

I'm no digital engineer but AFAIK it works either in 64 bit floating point or 32 bit fixed point, but either should be sufficient for double precision of CD based media. The key with any of this technology of course is the programming - one slip or oversight (especially when converting from floating point I am led to understand) can ruin the end result.

I haven't tried Acourate so can't vouch for the quality of the programming nor the subjective results but my understanding is that the technology employed isn't that much more advanced than the Tact and DEQX alternatives - especially when used in conjunction with CD quality formats. Maybe Shin can enlighten us there - but Tact certainly works with 32 bit processing - enough to achieve double precision (or an effective doubling of the dynamic range) internally.

From what I have read, it all comes down to the talent of the programmer rather than the chip specifications once we have reached that kind of level of processing. Dr Radomir Bozovic of Tact was previously heavily involved in the Psychoacoustics field - saying that it seemed to have little effect on the programming of the Tact RCS.

Hi Mike

Thanks.

I was not familiar with this tool, but by browsing through the tutorial presentations I see you can do almost anything in there.

My first thought was how convenient it is to linialize the drivers to a wanted low-band-high pass transfiguration. Can my Duelund theories be implemented and would they be superior to steep zero phase turn transfiguratuion?

You know I alowe for a total phase turn of 360 degrees each driver for any frequency in order to achieve a miltiway system act like an all driver piston speaker. A fullranger so to speak.

This is what I do for analogue x-overs to the point where a number of componets and their parasites starts to degrade more than they do good. And there I stop. This is the tweaking part.

With the SW tool it is obvious you can acheive even more ideal curves for the drivers compared to analogue filtering and certainly, if you find the result crazy, tune it in to your taste.

There is this basic rule we cannot/shall not change:

The basic driver must show before we begin to x-over it, a basic sense of call it quality. Not to many anomalities and artifacts should be found on the naked driver, such as crazy peaks, crazy dips nor insane waterfall plots.
Neither a very narrow-banded driver should in use and forced to be broader banded. We will likely hear either distorion from the lows and emphasized breake-ups in the highs.

If, then the digi-tooling will result in forcing the driver to something it doesn't really want to or cannot do at all due to design criterias.

Furthermore great care should be paid to if the we overdo the proccesor part and have too many processes running to correct an imperfect driver, should that be the case.

With such a tool I would likely not go to the limits, but just help a good driver to show sligtly more ideal curves etc. Just to cool down the processing power needed. Less is normally good.

Finally, it's fine it can time allign. Byt why waste procesor power in this field (either Shin do that, I think, looking to his speakers design).
Better to time align in the physical universe, meaning addressing this important point already when you design the speaker cabinet.

However I see the advantage using the SW time aligning tutorial to a whatever branded speaker bought, where this point has not been addressed.

Interesting..everything.

Now I think I'm going to try this spectaclar instrument out and see if the gains, if any, will be superior to my stances on keeping my BS signal quality policy valid.

Thanks for leading me into this interesting world.

I love when there is something to be tested out.

Maybe I should jump to England to have a listen.. if Shin alowes me......would be fun.


Gerner


PS! Dr Radomir Bozovic. So you know him too? I know him just periferically from when Peter Lyngdorf spotted him at Snell speakers. But that was after I left Peters audio-empire as an emplyoeed there.

 

[endust4237]

Just thinking loud.

I have a DEQX unit, I used to use it as a digital crossover (using two Altmann DACs on the digital outputs of the DEQX) between an Avantgarde Trio and a REL Stentor II sub. Worked very well. I really liked it. Practically no signature, nothing against the signal purity and the fun factor.

Few weeks ago I got a chance to use a Danley SH 50 PA speaker and a week ago an SH 100B as well. Very clever minimal phase error construction speakers using the patented tapped horn for the enclosure (more info on the Danley home page). I have tried to use the DEQX as a room corrector with the Danelys as well, but it did not work, the result was always much worse than the pure speaker only config. So Thomas Danley is about something. Now I sold my Avantgarde Trio, and I am extremely happy with the Danleys I really can not imagine any speaker which suits better to my taste, my system and my music collection. Probably for any price.

( A detailed summary of my journey with different Danley speakers in my system will come soon)

 

[Gerner]

Just thinking loud.

I have a DEQX unit, I used to use it as a digital crossover (using two Altmann DACs on the digital outputs of the DEQX) between an Avantgarde Trio and a REL Stentor II sub. Worked very well. I really liked it. Practically no signature, nothing against the signal purity and the fun factor.

Few weeks ago I got a chance to use a Danley SH 50 PA speaker and a week ago an SH 100B as well. Very clever minimal phase error construction speakers using the patented tapped horn for the enclosure (more info on the Danley home page). I have tried to use the DEQX as a room corrector with the Danelys as well, but it did not work, the result was always much worse than the pure speaker only config. So Thomas Danley is about something. Now I sold my Avantgarde Trio, and I am extremely happy with the Danleys I really can not imagine any speaker which suits better to my taste, my system and my music collection. Probably for any price.

( A detailed summary of my journey with different Danley speakers in my system will come soon)

Hi Endust

I have a chance to hear the Danleys in Budapest, just 380 km from here.

Will do that.

Gerner

 

[joel]

Acourate - a series of digital programs that run from crossover generation to room correction.

Ulli Bruggmann is a very smart guy.
Unfortunately, Gerner has failed to convince me that "Signal Purity" as espoused by audiophiles is not BS. I shall continue to live in digital error.

 

[Gerner]

Ulli Bruggmann is a very smart guy.
Unfortunately, Gerner has failed to convince me that "Signal Purity" as espoused by audiophiles is not BS. I shall continue to live in digital error.

Hi Joel

I didn't exactly try to convince you Joel. Why should I?

I just put an opinion here.

Yes the Brugmann-man is clever, but reading his white paper it is to *cheap* to convince people about his theories on passive x-overs when using Butherworth transfigurations as an example.

It should be known to speaker guys that Butherworth is worth nothing. NOTHING!

What do you think would have been his conclusions if he compared his theories to a 1:1 transfiguration?

So this is also BS *theory purity*

Gerner

 

[joel]

It should be known to speaker guys that Butherworth is worth nothing. NOTHING!

Really.

 

[Gerner]

Really.

YEP

 

[ShinOBIWAN]

Hi Shin..

So you have your room correction processing going on on a PC?...64 bit! XP/Vista/? 64 bit MOBO?
So it's dedicated only to support the room correction processors?

Its a little misleading but a 64bit operating system isn't required for 64bit audio processing. You don't even need a 64bit processor although one with 64bit extensions will be roughly twice as efficient should the code take of advantage of that.

64bit data can be represented by two 32bit types and really when you see 64bit marketed on CPU's and OS's they're talking about the memory and data address range which is a physical quantity. You can represent 512bit data and process it provided you have the software coded accordingly. Of course you take a performance hit over processing that same data using a bonafide 512bit instruction set because, say a 32bit CPU, has 16 times less physical address bits so the data needs to be split and sent in 16 chunks compared to just once on genuine 512bit CPU. Very inefficient.

Just for the record I use Vista 64bit with an Intel quad core running at 3Ghz.

But isn't the output just still 16 bit? Or?

If digital then 16-24bits input, this is then processed with 64bits of resolution at either native sample rate or can be upsampled upto 192Khz. If analogue input then its converted to digital using the 24bit ADCs and then processed.
Once any processing is finished its back to the DAC which are 16/24bit so dithering occurs. So in a nutshell 64bit precision is used throughout the processing to maintain sufficient headroom for iterative computational stacking and minimise rounding or truncation. The end result is more transparent processing with less digital footprint.

Tell me is it something you develope yourself or what?

The software is written by Dr. Ulrich Bruggemann. A very bright chap that has worked with room correction for some years including solutions for Tact.

Many questions, sorry. I have not heard of this approach before.

Its not something I was aware of either until somebody mentioned it to me and I gave it a go as I was already enjoying similar but lesser room and driver correction methods.