Ayre wants $1.5K for DSD'ed QB-9

[labjr]

I'd like to know why he chose the ES9016 instead of the ES9018 reference part?

[pawel8]

Charles

Thank you for explanation of Ayre view.

 

I do not have any agenda,like some other members do.

I am also not interested in measurements,which seems to be obsession of some engineers posting on this great forum.

I am participating to learn and exchange experiences.

By the way your original QB 9 sounded better than 13 k Burmester 099 and much better than Berkeley Alpha Dac mark II(I had both).

Vx5 is outstanding sounding piece.

I heard it both in 300000$ system and in my little "cheap" home system:great!

Best regards

Pawel

[The Computer Audiophile]

The more I read this thread the more I find it hard to believe anything from people who haven't built a component for sale to the public. Text books and mathematics are all well and good but the real world is vastly different.

 

The US Space Shuttles travel tens of thousands of miles per hour under certain conditions. That has nothing to do with how I drive my car so I don't really care to discuss it in a car discussion about my car. Thus, the topic of the QB-9 doesn't deserve some of this nonsense it's receiving.

[labjr]

I find it hard to believe some of the nonsense that comes out of someone who builds components for sale to the public. However, I find it rather entertaining.

[The Computer Audiophile]

I find it hard to believe some of the nonsense that comes out of someone who builds components for sale to the public. However, I find it rather entertaining.

Hi labjr - Until one wears the shoes of a manufacturer and designer its somewhat Monday morning quarterbacking to talk theoretical numbers and criticize designs.

[The Computer Audiophile]

One additional thought. Whatever format my favorite music is available in is the format I want. If its available in PCM and DSD I may want both because ill get two different sounds.

 

People arguing and getting so caught up in the theoretical merits of a format have really lost the point of this wonderful hobby. It doesn't matter if the format goes to 11 or 11 GHz.

[Charles Hansen]

Charles, regarding your ESS Sabre DAC upgrade. At what bit rate and sample rate does the ESS DAC convert PCM/DSD to analog? Also, which low pass filter (50kHz, 60kHz or 70kHz) are you recommending for DSD playback? No agenda here...I'm just curious.

 

Jesus R

 

Hello Jesus,

 

When we changed from the Burr-Brown DAC chip to the ESS, we retained our proprietary digital filter implemented in a Xilinx FPGA. In both cases we bypassed the internal digital filter, which I know is 8x in the B-B and believe is also 8x in the ESS although they are not so forthcoming with the internal workings of their parts. Our digital filter operates at 16x for single sample rates, 8x for dual sample rates, and 4x for quad sample rates. Therefore the output to the DAC is 705.6/768 kHz regardless of the original data rate.

 

There are ten sets of filter coefficients -- "Listen" and "Measure" for the 4 data rates gives 8, plus we also have two more for pre-emphasized 44 and 48 kHz data for a total of ten. All filters are minimum-phase (no pre-ringing) and all are accomplished in a single pass (no compounding of rounding errors). We use 32-bit filter coefficients and a 64-bit data accumulator so there can never be any clipping due to data overflow.

 

The default mode for the ESS chip uses an ASRC (asynchronous sample rate converter), but I prefer to use the original data from the disc instead of replacing it with other data, so we bypass the ASRC. In our rests we found that these parts give higher performance with a higher clock speed, so we changed from 512 Fs master clocks to 1024 Fs master clocks. These run synchronously with the incoming data. I believe (but don't know for sure) that the modulator inside the ESS chips runs at the same rate as the master clock. This means that there is an additional 32x oversampling filter inside the ESS. Most companies use what is called a "zero order hold" for he interpolator, which means that they feed it the same sample 32 times in a row. I have heard a rumor that ESS is the only company that does something more sophisticated than this, but I don't know if that is true, or if so, what it is that they do.

 

Since we are not licensees of Sony, we do not have to follow the recommendations of the Scarlet Book. The ESS chip has lower levels of OOB noise than most other chips, so we use a much gentler filter than the normal 3rd order at 50 kHz thatis required by the Scarlet Book. We have a split pole design with one pole at 80 kHz the other at around 200 kHz. Also our equipment is completely undisturbed by the OOB noise as everything is built with zero-feedback, fully-balanced circuitry using JFET inputs. We have been using a similar design in the C-5xe for nearly a decade and have only run into compatibility problems in one case. One customer had a vacuum tube preamp that was transformer coupled. The transformers were poorly designed and the combination of the leakage inductance and the stray capacitance led to a +10 dB peak at around 130 kHz. He could hear some high frequency "birdies". Further investigation showed that this occurred with ALL SACD players, but just at a somewhat lower level with units that complied with the spec.

 

Best regards,

[Charles Hansen]

Charles

Thank you for explanation of Ayre view.

 

I do not have any agenda,like some other members do.

I am also not interested in measurements,which seems to be obsession of some engineers posting on this great forum.

I am participating to learn and exchange experiences.

By the way your original QB 9 sounded better than 13 k Burmester 099 and much better than Berkeley Alpha Dac mark II(I had both).

Vx5 is outstanding sounding piece.

I heard it both in 300000$ system and in my little "cheap" home system:great!

Best regards

Pawel

 

Hello Pawel,

 

Thank you for the kind words. I am glad that you found the QB-9 to be an enjoyable component for listening to music. We spend a lot of time listening as we find it much more helpful than all of the theory. Every once in a while one can find a measurement to correlates with the actual sound quality, but this is quite rare.

 

DSD is a good example of this as the specifications are would indicate much less good sound than it actually has. As I have noted in the past, I have never been particularly "blown away" by the sound, but it did surprise me by being better than one would expect. This led us to do some investigations, some of which helped us to improve the sound of PCM products.

 

We also learn something new with each product we build, so some of the lessons from the QA-9 A/D converter and the AX-5 integrated amplifier were incorporated into the upgrade for the QB-9. I was very happy with the performance of the QB-9 before and was pleasantly surprised at how much we were able to improve it. I am confident that you will also be impressed when you get a chance to hear it also.

 

Best regards,

[elcorso]

Hi folks! So, I suggest reading this:

 

http://www.grimmaudio.com/whitepapers/dsd%20faq.pdf

 

Grimm has an interesting perspective on DSD 128 and higher… This seems sensible enough to me, suggesting that there is no analog circuitry which could even take advantage of the higher resolution of DSD128 and higher...

 

BTW, Miska, it seems from your last post that you claim to have measured S/N ratios of ~200 dB at the analog outputs of a DAC? Sorry, but that is impossible for me to believe.

 

Hi Barrows,

 

This paper from 2005 is a little 'old' for modern times...

 

I was reading about Grimm releasing a new (I don't exactly know if ADC or DAC) with DSD128 capability in some Germany audio show, but I could be wrong since my German language is so bad...

 

I'm listening with an exaSound e20 DAC to DSD128 and until now I don't any problems with the analog circuitry, but sonic advantages (to my ears).

 

Sources: Only 3 albums I got from DL, but also by up sampling from DSD64. The chosen recordings were recorded in analog or DSD (never from DXD)

 

Best,

 

Roch

[barrows]

hahaha:

 

"This paper from 2005 is a little 'old' for modern times…"

 

Nonsense. The analog limitations have not changed significantly in ten years.

[tubesound]

This paper is worth reading: AES E-Library » Audibility of a CD-Standard A/DA/A Loop Inserted into High-Resolution Audio Playback

[Ned Kelly]

The more I read this thread the more I find it hard to believe anything from people who haven't built a component for sale to the public. Text books and mathematics are all well and good but the real world is vastly different.

 

The US Space Shuttles travel tens of thousands of miles per hour under certain conditions. That has nothing to do with how I drive my car so I don't really care to discuss it in a car discussion about my car. Thus, the topic of the QB-9 doesn't deserve some of this nonsense it's receiving.

 

Agree, but I've learnt from past experience that its never a good idea to question DiYers on their criticism of a given design - it only makes things worse. Of course, in the case of several in this thread I have no idea whether their knowledge has ever extended to the hot end of a soldering iron. Those familiar with the nwavguy saga will know that he was the only one who didn't make any money from his designs - anyone who feels they can compete in the DSD space should play their cards very close to their chests, IMO : unless your name is Kardashian, chances are than fans alone wont pay your bills.

[One and a half]

Hi labjr - Until one wears the shoes of a manufacturer and designer its somewhat Monday morning quarterbacking to talk theoretical numbers and criticize designs.

 

Taken in the context of Ralph Nader, who didn't build a car in his life and sell it to the public, but was aware of significant shortcomings of designs that he could look at from a different perspective. His observations are well documented to their outcome, questioning designs, is like all questions. People are enthusiastic about this hobby, some indeed have taken to a soldering iron, built prototypes and with their peers in different countries on different levels have achieved remarkable results without huge budgets. DSD has a large following in Japan for example, occasionally their inputs reach these pages.

We should encourage to listen to all points of view, even those not in the "know".

[Miska]

BTW, Miska, it seems from your last post that you claim to have measured S/N ratios of ~200 dB at the analog outputs of a DAC? Sorry, but that is impossible for me to believe.

 

No, as I said I talk about what comes out of my player software and goes to DAC. Of course analog side is going to limit it. The point is that digital volume control doesn't limit dynamic range compared to original data.

 

For DSD256 the noise dips lower than 32-bit PCM with flat noise floor. 384 kHz noise shaped 32-bit PCM has even lower noise at audio frequencies. Flat dithers like TPDF are pointless for 4x and higher rates and limit the digital domain performance unnecessarily.

[Miska]

Charles,

 

That 8x interpolation and ZOH/SAH are exactly what makes these "PCM" DACs perform poorly. Combined with low order analog reconstruction filter it generates those nasty images every 352.8/384 kHz as I posted earlier. That's what I mean by resource constrained implementation, ugly.

 

And no, your quote of ScarletBook is incorrect btw.

[Hiro]

hahaha:

 

"This paper from 2005 is a little 'old' for modern times…"

 

Nonsense. The analog limitations have not changed significantly in ten years.

 

So you're saying (or Bruno) that there's no benefit in going beyond 64fs and recording to 128fs or 256fs via the QA-9, because the analog limitations of the converter limit its resolution?

[Miska]

That 8x interpolation and ZOH/SAH are exactly what makes these "PCM" DACs perform poorly. Combined with low order analog reconstruction filter it generates those nasty images every 352.8/384 kHz as I posted earlier. That's what I mean by resource constrained implementation, ugly.

 

Here's output of 44.1k PCM sweep of such DAC. You can also compare size of the "SDM noise problem" peaking at 500 kHz to the PCM noise problem peaking every 352.8 kHz throughout the spectrum.

 

Here's output of 352.8k PCM sweep upsampled to DSD128 in my software and played back with Sabre DAC:

 

You can see roll-off of Sabre's DSD filter.

 

I'll make a new plot later with something like 0 - 1.4 MHz sweep for comparison. Now the sweep output abruptly stops at 176.4k because input is just DXD.

[Miska]

Speaking of 1024x rates. Just for fun of it, I quickly optimized one of my 1-bit modulators to higher rates and converted 1 kHz test tone to DSD1024 and then back to 325.8/32 PCM. This is how it looks like. Below 70 kHz or so, noise floor becomes capped by the 32-bit PCM conversion (32-bit quantization noise is not shaped here).

 

[The Computer Audiophile]

Agree, but I've learnt from past experience that its never a good idea to question DiYers on their criticism of a given design - it only makes things worse. Of course, in the case of several in this thread I have no idea whether their knowledge has ever extended to the hot end of a soldering iron. Those familiar with the nwavguy saga will know that he was the only one who didn't make any money from his designs - anyone who feels they can compete in the DSD space should play their cards very close to their chests, IMO : unless your name is Kardashian, chances are than fans alone wont pay your bills.

Hi Nad - I agree. Without names I have no idea if people commenting are even DIYers or just people with book smarts or neither.

[The Computer Audiophile]

Taken in the context of Ralph Nader, who didn't build a car in his life and sell it to the public, but was aware of significant shortcomings of designs that he could look at from a different perspective. His observations are well documented to their outcome, questioning designs, is like all questions. People are enthusiastic about this hobby, some indeed have taken to a soldering iron, built prototypes and with their peers in different countries on different levels have achieved remarkable results without huge budgets. DSD has a large following in Japan for example, occasionally their inputs reach these pages.

We should encourage to listen to all points of view, even those not in the "know".

Hi one and a half - I understand your example completely and think good work has been done by people like Ralph Nader. However, Ralph consulted many in the industry and even more important he didn't hide his identity, agenda, or references behind an internet moniker. Ralph used the knowledge of many people who actually built cars to back his position. Here I don't know why anyone is saying anything unless they provide information in their signatures.

[Audio_ELF]

Chris... With respect to your point of view; Charles appears to have a large chip on his shoulder about Sony; SACD and DSD. His aggressive postings are as much aimed at Miska (who I believe has well known experience in this area) as at the "armchair engineers".

 

As you commented earlier: this thread is so far off its original topic and so argumentative I really don't know why you haven't locked it...

 

Just the opinion of an armchair reader and general trouble maker ("pot", "kettle" and "black" come to mind considering I've posted in this thread)...

Eloise

[wisnon]

But there are marvelous info gems in this thread at the same time. The stress of the point-counterpoint challenges, has really brought out some interesting perspectives. Best thread on the site at the moment!

[labjr]

Hi one and a half - I understand your example completely and think good work has been done by people like Ralph Nader. However, Ralph consulted many in the industry and even more important he didn't hide his identity, agenda, or references behind an internet moniker. Ralph used the knowledge of many people who actually built cars to back his position. Here I don't know why anyone is saying anything unless they provide information in their signatures.

 

This kind of thing happens in most forums I've seen on the internet. Probably why most manufacturers stay out of the fray. Charles Hansen chooses to participate and does so at his own risk. He could go away and say nothing but he doesn't. He actually, seems to enjoy dumping gas on the fire.

[mav52]

Hi Nad - I agree. Without names I have no idea if people commenting are even DIYers or just people with book smarts or neither.

 

The way I read this back and forth stuff. If these people who are digging into the product Charles is responsible for proclaiming the rights and wrongs of the technology are so right in their beliefs, why are they not designing, engineering, building and selling their OWN product. Talk is cheap but putting a product out there to support "the talk" carries a lot more weight than internet chatter. The stuff gets old, fast...

[Audio_ELF]

The way I read this back and forth stuff. If these people who are digging into the product Charles is responsible for proclaiming the rights and wrongs of the technology are so right in their beliefs, why are they not designing, engineering, building and selling their OWN product. Talk is cheap but putting a product out there to support "the talk" carries a lot more weight than internet chatter. The stuff gets old, fast...

Miska DOES design this stuff...