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

[elcorso]

No, I don't think it is all about theories and graphs, but if you read thru' this thread, you'll see that DSD proponents have offered graphs as evidence that DSD is superior to PCM. Charles Hansen has shown demonstrable evidence the Sony's impulse graphs are both fabrications and deceiving. Even Andreas Koch's figure is deceptive; I disagree that "he didn't hide anything." At best, his figure is apples-to-oranges.

 

I haven't yet heard DSD but plan to try it after I get my QB-9 upgraded next month. But, even if it does sound stupidly good (it's going to have to be AMAZING to sound better than the hi-res I'm already hearing), I'll be listening mostly to the handful of free demos. I've only seen a single sampler that I'm willing to pay for. And that's just the DSD64 stuff. After searching the DSD download sites listed on Audiostream, I found fewer than ten DSD128 titles (and most of those -from 2L - were recorded as DXD files), none of which I would buy.

 

Charles' 1 m LP analogy is silly and, IMO, entirely appropriate.

 

I really don't care about graphs and similar. In this computer music world everyone wants graphics. Then Audacity (and others) are the best 'music players' as today. Somebody understand the graphs, some others don't, but at the end it will be how the music 'sounds' to you.

 

If you find "Andreas Koch's figure is deceptive" I could suggest to listen to some of his DACs.

 

And I'm absolutely sure you will be not disappointed by you upgraded Ayre's QB-9 playing native DSD.

 

Roch

[Charles Hansen]

Regarding noise, I just reproduce here two plots I created in other thread, I don't have time to produce more right now.

 

(since there's now DSD512 DACs too, I could make a plot of that too at some point)

 

Hello Miska -

 

Thank you for the graphs. Unfortunately I am color blind, so they do not help me much. Could you please plot them as a normal X-Y plot for me?

 

Thanks,

Charles Hansen

Ayre Acoustics, Inc.

(Just a Dumb Analog Hardware Engineer)

[Charles Hansen]

Charles,

 

I think you are scolding the wrong person. If you read this thread backwards I'm one the CA member who defended you from somebody that was offending. But also I'm a DSD defender, I own even one of your products, the Ayre C-5 Universal Player (whit the pre-ringinging digital filter et all). BTW, thanks to this unit I bought a lot of SACDs because of his SQ, that now I enjoy by computer playback.

 

You (Ayre) has the right to choose the DSD format you want for your DAC, but we, consumers, have also the right to choose what we want. I never said "don't buy Ayre because it only plays DSD64", but the possible causes of this decision, supporting it.

 

BTW, this is a crazy world, isn't it?

 

Enjoy the music (in the format you like better).

 

Roch

 

Dear Roch,

 

Thanks for your support. I hope you enjoyed your time with the C-5. As we learned more about digital, we implemented our own custom digital filter for the C-5 that had no pre-ringing. It made a very nice improvement in the sound quality. You can read about that at: http://www.ayre.com/pdf/Ayre_MP_White_Paper.pdf

 

Yes, we live in a crazy world...

 

Cheers,

Charlie Hansen

Ayre Acoustics, Inc.

[Charles Hansen]

1) Because it allows to move bulk of the DSD noise one or two octaves up

2) Because it gives extra SNR for the audio band when performing digital volume control and digital room correction for DSD

3) Because it relaxes requirements for the analog reconstruction filter

 

I'm running all this every day. (as well as many other people)

 

Hello Miska,

 

Thank you for the information.

 

But I think you are making a BIG mistake. The only way to upsample the DSD is to filter out the noise. If you do not, the noise will alias INTO the audio band and make the performance much worse. But since the noise starts at 20 kHz, the only way to filter the music from the noise is to use a BRICKWALL FILTER.

 

But the bad sound created by the BRICKWALL FILTER is the only reason that DSD sounds better than Redbook CD! (Let alone high-res PCM!!)

 

So if you want your DSD to sound like Redbook CD, you can either press the button on the front to read the Redbook layer or upsample it to a higher rate!!!!

 

Thanks,

Charles Hansen

Ayre Acoustics, Inc.

(Just a Dumb Analog Hardware Enginee)

[beetlemania]

If you find "Andreas Koch's figure is deceptive" I could suggest to listen to some of his DACs.

 

And I'm absolutely sure you will be not disappointed by you upgraded Ayre's QB-9 playing native DSD.

 

I have no doubt that Andreas Koch's DAC sounds really good. His products are well-reviewed and even Charles Hansen has had kind words in this thread regarding his engineering (but not his comparison of DSD and PCM!). But that doesn't change the accuracy of his graph, does it?

 

And, yes, I'm looking forward to getting my DSD upgrade. More so for the improved overall sonics (with crappy old-fashioned PCM), however, rather than the ability to play a handful of DSD files (albeit I am curious to hear what all the DSD hoopla is about).

[wisnon]

Hello Miska,

 

Thank you for the information.

 

But I think you are making a BIG mistake. The only way to upsample the DSD is to filter out the noise. If you do not, the noise will alias INTO the audio band and make the performance much worse. But since the noise starts at 20 kHz, the only way to filter the music from the noise is to use a BRICKWALL FILTER.

 

But the bad sound created by the BRICKWALL FILTER is the only reason that DSD sounds better than Redbook CD! (Let alone high-res PCM!!)

 

So if you want your DSD to sound like Redbook CD, you can either press the button on the front to read the Redbook layer or upsample it to a higher rate!!!!

 

Thanks,

Charles Hansen

Ayre Acoustics, Inc.

(Just a Dumb Analog Hardware Enginee)

Or filter from say 30khz.

 

I see you fear intermodulation noise...

[Charles Hansen]

SDM stands for "Sigma-Delta Modulation" which is also a digital encoding system. It is a specific subset of PDM or PWM.

 

It doesn't need to have anything to do with A/D or D/A conversion. I have both PCM -> SDM and SDM -> PCM conversion running purely in software, on-the-fly. You can create a PWM or PDM output also without having a sigma-delta modulator at all.

 

Delta-sigma (ΔΣ; or sigma-delta, ΣΔ) modulation is a method for encoding analog signals into digital signals or higher-resolution digital signals into lower-resolution digital signals.

 

 

 

DSD is specifically 1-bit SDM. It can be considered to be either PDM or PWM too. It is SDM specifically because it relies heavily on noise Sigma-Delta modulation process while PDM or PWM don't necessarily imply any noise shaping.

 

So, IMO, calling DSD PDM, PWM or SDM are all correct, while SDM is the most specific.

 

Practically, DSD converters are really discrete time PWM (SAH'ed PWM).

 

Hello Miska --

 

I disagree with you. SDM stands for Sigma-Delta Modulator (many people think it should be more properly referred to as "Delta-Sigma Modulator).

 

Ayre makes an A/D converter box. Inside is a chip from a company called Arda. When you look at the datasheet there is a block diagram of what is inside the chip. The heart of it is a block called "Multibit Sigma-Delta Modulator". I suppose that you could call what comes out of that as "Sigma Delta Modulation" but that would not be very helpful of enlightening. The reason is that nobody ever uses the output of a Delta Sigma Modulator without processing it first.

 

99.99% of the time, the signal from a DSM is sent through a digital low pass filter and the resulting signal is PCM in a standard format (such as 192/24). The more bits in the DSM, the harder it is to make, but the higher the performance will be (assuming equal precision for all the bits used). When CD was first released commercially, Sony could make a 16-bit ladder DAC, but Philips could only make a 14-bit ladder DAC. So they oversampled and noise shaped the output to get 16 bits of resolution.

 

But everybody wants to make everything as cheaply as possible these days. So first they made the 1-bit DSM which had good linearity but poor S/N ratio. So they added more bits. With present day technology, the best tradeoff between price and performance is to use a 4-bit multi-bit DSM. The Arda chip we use is more expensive and higher performance, so it uses a 6-bit DSM.

 

As I said before, one can run the output of the 6-bit 256 Fs DSM through a digital lowpass filter to end up with PCM. OR if you want DSD (with only one bit) you could throw away three bits and just use the MSB. However the performance would be terrible. So instead, we RE-modulate the output of the Delta-Sigma Modulator. We use an FPGA to do this digital math, but it could be done with a microprocessor or with a DSP (Digital Signal Processor) chip, which is simply a microprocessor with additional special instructions built in so that common audio tasks can be programmed more easily.

 

Arda is kind enough to supply a dumb analog hardware engineer like me with instruction on how to re-modulate the 6-bit output of the multi-level DSM chip down to 1-bit for DSD. I will quote here extensively from their app note:

 

The multibit outputs of the AT1201 are remodulated into DSD signals using a 7th-

order single-stage 1-bit sigma-delta modulator, as shown in Figure 1. The

modulator’s noise transfer function has zeros distributed at low frequencies to

maximize dynamic range.

The multibit outputs are first lowpass filtered to attenuate high-frequency

quantization noise before remodulation into a 1-bit stream. To generate the DSD

streams at 64x and 128x, the multibit signal is also decimated by 2 and 4,

respectively.

 

The sigma-delta modulator in the AT1201 supports a high modulation depth and

enters overload at -2 dBFS. The 7th-order, 1-bit Σ∆ modulator in the DSD generator

has lower modulation depth, entering overload at about -6 dBFS. Therefore, signal

scaling is needed between the multibit outputs of the AT1201 and the DSD generator

in order to achieve the best possible performance.

Furthermore, the out of band noise in the multibit outputs of the AT1201 must be

attenuated before remodulation to a DSD stream. Failure to do so results in poor

performance in the high-order modulator.

 

As I noted before virtually every single A/D converter made in the whole world today uses a multi-bit Delta-Sigma Modulator. The ONLY exception that I know of is the Grimm "discrete" converter, which is limited to only DSD-64. Oh, and I was reading this forum that some company (possibly Signalyst) used and older 1-bit TI DSM chip for their A/D converter. But the current top-of-the-line A/D chip from TI uses a 4-level multi-bit DSM. Then to get DSD output aat one bit they must also have an internal remodulator.

 

So ALL modern A/D converters begin with the same basic technology. The only difference is what happens next. Do you digitally re-modulate it to one bit to get DSD, or do you digitally low pass filter it to create a standard PCM format? The 6-bit output IS PCM already, but there is no audio system that uses 6 bits. So by oversampling (256x in the case of the Arda, 64x in the case of the original Sony or the current Grimm) one can create a higher resolution signal.

 

That is why the Arda chip is the best one currently available that I am aware of. It has more bits (6) than the competitors, and a higher oversampling ratio (256x), so the amount of information density is the best. When we convert it to PCM, there is more resolution (higher S/N ratio) out to MUCH higher frequencies, than any other chip.

 

We made one prototype with the TI PCM4222 chip (their best one). The performance was MUCH worse than the Arda. It had the same problems with out of band noise that forces people to use brickwall filters that ring. But with the Arda, the out-of-band noise was SO low that we could use a very gentle filter with NO ringing whatsoever and the noise remained very low.

 

~~~~~~~~~~

 

But the bottom line is this. DSM (or SDM as many call it) is a MODULATOR. It has an output that can be converted to standard PCM rates and bit depths, or it can be converted to single bit DSD.

 

This latter point is obvious. There are several companies that make players where the PCM on a standard CD (or DVD) is converted to 1-bit DSD because they feel that provides the best sound quality.

 

Everybody has their own philosophy, experiences, and opinions. That is why there is not just one company (or many companies all making identical products!). But in my experience, once the signal has been converted by an A/D converter to a certain format, the amount of resolution is then fixed at a given level. One can do all kinds of manipulation of the digital signal, but the lost resolution can never be restored. We use an oversampling filter to simplify the design of the analog circuitry, but we don't pretend that it is going to restore the resolution that was lost when the recording was originally made.

 

So even if one believed that 1-bit systems had some kind of performance advantage, re-modulating the 16-bit data on a CD is not going to make it sound like it was recorded in DSD. We know that the lost resolution can never be restored, so the only way that a CD could be made to sound like an SACD is if that system has a coloration that can be added in to a recording that doesn't already have it.

 

And maybe that is what is going on. When I listen to DSD, I hear much the same thing that Barry Diament describes -- a "sameness" to the sound of the high frequencies that is common to ALL of the recordings, but is not present is real life.

 

There is also the variable of the systems that people use for the rest of their playback system. It may be that there is some coloration to DSD that complements the sound of some systems, so that the owners of those systems prefer DSD. I don't know the answers to these questions.

 

In the past there have been no valid comparisons between the two systems as FAR too many variables were involved. But soon we will add USB output to the DSD of our A/D converter. Then it will be trivial to digitize an analog signal with the Ayre A/D using two different encoding methods (and the same Delta Sigma Modulator) and then play them back through the same Ayre D/A conveerter (using the same DAC chip with a Delta Sigma Modulator) and ALL other variables will be eliminated. Then finally we will have a true comparison of the sonic performance of the two different formats.

 

Best regards,

Charles Hansen

Ayre Acoustics, Inc.

(Dumb Analog Hardware Engineer)

[Charles Hansen]

If you find "Andreas Koch's figure is deceptive" I could suggest to listen to some of his DACs.

 

Hello Corso,

 

That is what I do not understand. Andreas is an outstanding engineer who makes outstanding products. He is not in the position that Sony was in 15 years ago when by telling a few lies they hoped to regain a $1 billion per royalty stream. What does he have to gain by mis-representing the facts? He does not need to mislead people. Everybody knows that the Playback Design is one of the top designs on the market. Everybody knows that he helped build the original hardware that Sony leased to the studios to make SACD's. Everybody knows that he is the person who organized the group for the DoP initiative and fixed the bugs in the system that dCS proposed.

 

It is possible that he does not know the difference between "S/N ratio" and "noise floor of an FFT". But I don't believe that for one second. He is MUCH smarter than that.

 

I suspect that he was asked to make an article for an online magazine and either the magazine supplied the graph or he was very busy and working very quickly and made a simple error. I have met Andreas and spoken with him and I don't believe that he would intentionally deceive people. He is a good man with a good heart.

 

I cannot explain the error, but that does not change the fact that the graph is wrong and will mislead most people who read it. There are very few people with enough technical knowledge to look at that graph and understand that it is wrong. Even when it is explained, it is something of a difficult concept. But it should be corrected just as when an error appears in a magazine, there will be a correction in the next issue.

 

I remember when I had time to read Scientific American or even National Geographic.. In every single issue there was a small sidebar noting all of the errors from the previous issue. Nothing and nobody is perfect....

 

Best regards,

Charles Hansen

Ayre Acoustics, Inc.

[Charles Hansen]

Or filter from say 30khz.

 

I see you fear intermodulation noise...

 

Hello Wisnon,

 

No, that would be a better trade off. That is what we do with our PCM DACs. We have a selectable digital filter. One is a brickwall with almost no aliasing, but it is so sharp that it has a lot of ringing. We call that position "Measure" because it measures better when there is no aliasing and flat frequency response to 20 kHz.

 

The other position is called "Listen" and is designed for people that want to listen to musics instead of measure the DAC. It uses a much more gentle filter with at least 10x less ringing (all post, as it is a minimum phase filter). This cause a slight rolloff at 20 kHz and some aliasing of frequencies above 10 kHz.

 

However I must say that I am surprised that it sounds better to upsample. After all, the noise of DSD should be inaudible so why even bother to upsample it? There is no increase in resolution. It makes no sense to me. Perhaps someone who makes an upsampling DSD DAC would like to explain their reasoning on this forum.

 

Regards,

Charles Hansen

Ayre Acoustics, Inc.

[Charles Hansen]

Quick confirmation Mr. Hansen: will the upgrade be performed by the local agent, or will the agent have to ship the dac to the factory?

 

Hello Freann,

 

The distributor in each country is responsible for upgrades and repairs of the units they sell. They also set their own prices (just as with the products themselves), so the price will vary from the US price.

 

Best regards,

Charles Hansen

Ayre Acoustics, Inc.

[Miska]

The heart of it is a block called "Multibit Sigma-Delta Modulator". I suppose that you could call what comes out of that as "Sigma Delta Modulation" but that would not be very helpful of enlightening. The reason is that nobody ever uses the output of a Delta Sigma Modulator without processing it first.

 

Oh yes they, and I do. Bunch of DSD recording equipment uses PCM4202 or PCM4204 (or the old PCM1804), I use PCM4202 too running at DSD128 (5.6 or 6.1 MHz). It is a pure 1-bit converter, very good one actually with proper dithering.

 

Then there's also PCM4222 which I have also used, it also has DSD output but in addition it can output it's 6-bit modulator data directly. This way it gives it's best performance out.

 

Aside from the 6-bit output, both converter give their best output performance at DSD128.

 

99.99% of the time, the signal from a DSM is sent through a digital low pass filter and the resulting signal is PCM in a standard format (such as 192/24).

 

And this is where the problem starts, because for example for PCM4222 (TI's flagship ADC), SNR suffers 10 dB and THD suffers 2 dB. Plus time domain gets screwed up.

 

You can see more extensive comparison in the other thread.

 

The 6-bit output IS PCM already, but there is no audio system that uses 6 bits.

 

Multi-bit SDM is NOT PCM. Multi-bit SDM most of the time is not two's complement binary representation.

 

But everybody wants to make everything as cheaply as possible these days. So first they made the 1-bit DSM which had good linearity but poor S/N ratio.

 

OK, so TI managed to gain 6 dB SNR by using 6-bits while using single PCM4204 per channel (four converters) would give you same SNR figure. I think PCM4222 is in fact just six PCM4202's in a single package.

 

We made one prototype with the TI PCM4222 chip (their best one). The performance was MUCH worse than the Arda.

 

I get much better performance from it than what I can see here:

Ayre Acoustics QA-9 USB A/D converter Measurements | Stereophile.com

 

But the bottom line is this. DSM (or SDM as many call it) is a MODULATOR. It has an output that can be converted to standard PCM rates and bit depths, or it can be converted to single bit DSD.

 

Well, you can keep your opinion, I don't mind. You can call it "multi-bit PDM" or "multi-bit PWM". Modulator's output is SDM, just like R2R ladder's output is PCM.

 

I could make similar argument that "PCM is a resistor ladder" which of course it not true.

 

This latter point is obvious. There are several companies that make players where the PCM on a standard CD (or DVD) is converted to 1-bit DSD because they feel that provides the best sound quality.

 

One can do all kinds of manipulation of the digital signal, but the lost resolution can never be restored.

 

So far, least loss of delta-sigma converter output I've seen has been has been DSD, aside from the original multi-bit. I will probably at some introduce some 8-bit (9-level SDM) file format, since I'm already using it for DAC interfacing.

 

We use an oversampling filter to simplify the design of the analog circuitry, but we don't pretend that it is going to restore the resolution that was lost when the recording was originally made.

 

Who is? By using oversampling combined with noise shaping, you can however retain original digital SNR within the original frequency band while attenuating the signal in digital domain. That's the added bonus.

 

So even if one believed that 1-bit systems had some kind of performance advantage, re-modulating the 16-bit data on a CD is not going to make it sound like it was recorded in DSD.

 

Of course not, but it helps bypassing some of the poor oversampling and modulation inside DAC chips.

 

In the past there have been no valid comparisons between the two systems as FAR too many variables were involved.

 

That's why I built a converter with both PCM and DSD ADC in parallel with a common analog stage, plus a one single DAC stage. I can switch between the two with flip of a switch.

 

(PCM4222 by the way can give you both PCM and DSD output at the same time)

 

P.S. If I want 192/24 PCM, I can get pretty decent performance from my inexpensive Focusrite Forte or my other ADC based on CS5381.

[Miska]

However I must say that I am surprised that it sounds better to upsample. After all, the noise of DSD should be inaudible so why even bother to upsample it? There is no increase in resolution. It makes no sense to me. Perhaps someone who makes an upsampling DSD DAC would like to explain their reasoning on this forum.

 

Point of oversampling DSD is exactly the same as oversampling PCM. Don't play dumb...

[Miska]

But I think you are making a BIG mistake. The only way to upsample the DSD is to filter out the noise. If you do not, the noise will alias INTO the audio band and make the performance much worse. But since the noise starts at 20 kHz, the only way to filter the music from the noise is to use a BRICKWALL FILTER.

 

Charles,

 

Do you think I do anything without running tons of analysis?

 

I've been doing this DSP stuff now for 15 years, hopefully I've managed to achieve at least something.

[Onkle Je]

Hello Onkle,

 

Probably once a conversion has been made, it is best to leave it alone and play it back in the native format if possible. When we first made our first "universal" player, I said, "Let's convert the DSD to PCM. Then we will get rid of the out-of-band noise and it will sound better." So we made a conversion filter and listened to it.

 

We decided that the native DSD still sounded a *tiny* bit better than our conversion, so we left it out.

 

Today I think I know more. I believe that I could make a conversion filter from DSD to 176.4/24 PCM that would "invisible". But I don't think it would improve the sound either. But if it were truly transparent so that you could not hear if it were in the signal path or not, that would still be a good development. Then anybody with a good PCM system could keep it and still enjoy DSD recordings. I may work on that later this year, but maybe not. I don't like the existing DSD-to-PCM converters I have heard, but that doesn't mean that a much better one couldn't be made. The problem with most digital equipment is that it is designed by digital engineers. They have been inculcated in a system that is a fantasy land. Only the best of them break free from the system that they have been brought up in.

 

Best regards,

Charles Hansen

Ayre Acoustics Inc.

 

Thank you very much for your answer, Charles. Hopefully, you'll find some time to work on your own converter.

Regards

[Charles Hansen]

Then there's also PCM4222 which I have also used, it also has DSD output but in addition it can output it's 6-bit modulator data directly. This way it gives it's best performance out.

 

Thank you for the correction. I said the PCM4222 used a 4-bit Delta Sigma Modulator, which was incorrect.

 

Well, you can keep your opinion, I don't mind. You can call it "multi-bit PDM" or "multi-bit PWM". Modulator's output is SDM, just like R2R ladder's output is PCM.

 

You seem to be like don Quixote, tilting at windmills. Even on the TI data sheet for the PCM4222 they refer to SDM as a "Sigma Delta Modulator":

 

The 6-bit outputs from the delta-sigma modulators are routed to the digital decimation filter, where the output of the filter provides linear PCM data. The linear PCM data are output at the audio serial port interface for connection to external processing and logic circuitry. The multi-bit modulator outputs are also routed to a direct stream digital (DSD) engine, which converts the multi-bit data to one-bit DSD data.

 

(What TI refers to as a "DSD engine" is obviously the re-modulator.)

 

But if you want to make up your own nomenclature that is different from the industry standard -- even the suppliers whose parts you use -- it is a free country and you are free to do so. But in this country we call it "spitting into the wind". It usually doesn't do much good. And when you speak with other engineers, they will have a difficult time understanding you since you use your own terminology....

 

I will probably at some introduce some 8-bit (9-level SDM) file format, since I'm already using it for DAC interfacing.

 

WOW! I was right -- a true don Quixote! It's not enough to make up your own terminology. Now you want to make up your own new format!

 

I think there is already enough problems with two formats - DSD and PCM - and two sampling frequencies - 44.1 kHz and 48 kHz (and their multiples). And now you think that we need a new format because (I suppose) that the existing ones are not good enough for you.

 

Well if Sony (when they still had billions of extra dollars) was unable to convince people to switch formats, I think it will be many, many times more difficult for a small, one-man company to do so. But good luck to you! I will stick with PCM, as I have been able to achieve excellent results with it and everybody already uses it.

 

 

That's why I built a converter with both PCM and DSD ADC in parallel with a common analog stage, plus a one single DAC stage. I can switch between the two with flip of a switch.

 

(PCM4222 by the way can give you both PCM and DSD output at the same time)

 

P.S. If I want 192/24 PCM, I can get pretty decent performance from my inexpensive Focusrite Forte or my other ADC based on CS5381.

 

All I can say is that it is no wonder that you prefer DSD if you are using the standard digital filters supplied by Texas Instruments. If there is one lesson that people have learned from this thread, I hope it would be that the problem with PCM is due to the standard digital filters that everyone uses. This is true for both the record side and the playback side. It is the reason why it is easier to achieve good sound with DSD than with PCM.

 

But by using custom filters (and also being sure to use that wonderful machine between our ears), we can make PCM sound as good or better than DSD, yet have NO problem with out-of-band noise, NO problem to replace all of the recording, mastering, and playback equipment, and NO problem to convert the signal to PCM for processing and then BACK to DSD for playback (adding more noise with each conversion).

 

I would encourage you to use your obvious supply of knowledge of digital systems to improve PCM instead of "tilting at windmills", trying to change industry-standard nomenclature and introducing your own formats.

 

Best regards,

Charles Hansen

Ayre Acoustics, Inc.

(Dumb Analog Hardware Engineer)

[Charles Hansen]

You sent a PM, but your inbox is full and I can't reply. Here is your answer.

 

Love the info you are putting out there, even if I dont agree totally.

 

You clearly know your stuff and as such are entitled to your learned view.

 

Dont allow anyone to get under your skin...just keep doing your thing and keep your customers happy.

 

The way I see it, it like back in the cassette days. We all wanted Dolby, DTS, metal tape compatibility, Chrome compatibility, etc...we just wanted a deck that could play EVERYTHING....no matter what was considered 'best'.

 

A pity you didnt do DSD128...cover all the commercial bases, unless it cost a lot to do so. Then you can still say you prefer PCM but offer all alternatives. Just like inputs/outputs...XLR, SPDIF coax, Toslink, USB are all options commonly offered even if only one of them sound best.

 

Hello Wisnon,

 

Thanks for the message. I didn't realize you had sent it as the default setting are NOT to notify me. Weird...

 

Anyway, thanks for the kind words. If I get cranky, please don't take it personally. I was in an accident and paralyzed and have to take a lot of pain meds and sometimes it affect what I write -- especially when I do so "off the cuff" -- then it comes off the wrong way, like a personal attack, which is almost NEVER what I mean.

 

Your example of a cassette is a good one, and it is an example of the type of complexity that I would like to avoid. I still have a really nice Sony 3 head, 3 motor cassette deck -- just one notch below a Nakamichi. Sometimes I look at that thing and wonder how often the AVERAGE user ever got ANYTHING right.

 

I used to work at a stereo shop and what set us apart was our service department. The owner had been the chief tech at a shop for close to twenty years. So when he started his own shop, his USP (unique selling proposition) was that when you bought a piece of equipment from him, it was GUARANTEED to meet the specs.

 

Nowadays it is not so important but in the old days there were only three sources -- LP, FM, and cassette. Well FM and cassette will ONLY meet the specs if they are "aligned" properly. To do a proper alignment on either one took about an hour for a good tech. When they build them at the factory there is no way in the world that they had time to make all the dozens of adjustments required to get them to meet the spec.

 

But at our store, we would open up the box, take out the unit and fine tune every single setting in there to get it to work at the highest possible level. The factory would just measure the voltage at a test point and set that to some "average" that would get the performance in the ballpark, but there was too much variation in the individual parts to get it to actually meet the spec.

 

The closest thing we have to that today is with a video display. If you ever read a home theater magazine, they will talk about "calibrating" the display, and that is the same thing. The magazines have to do it as otherwise you would NEVER know how good the design really is. BUT if YOU don't have YOUR set calibrated also (typically $300 to $500), you won't get the proper results either.

 

If you are buying a $8000 projector, you would be stupid not to have your display calibrated. But how many people that buy $800 "flatscreens" at Costco are going to calibrate it? ZERO....

 

Back to cassettes. I was looking at my deck the other day and there are so damned many switches and knobs and dials, that I doubt that more than 0.2% of the people using them ever got the most out of them. Even if it had been calibrated, it had to be for a specific brand and model of tape. And if they ever changed (improved) the formulation, you had to have it re-calibrated.

 

People these days don't have the time or the patience for all of that nonsense. Even with something as simple-minded as a USB DAC, how many people do you think really go out and get a good aftermarket player and go through ALL of the settings to get the best sound quality possible.

 

Here's an example -- have YOU oriented the AC plugs of every piece of equipment in your system? If not (and I doubt that you have) you are missing out on a lot of performance....

 

So the LAST thing we need is MORE formats and MORE variations and MORE choices. If we go down that road, then MAYBE 0.2% of the people will know enough and get better sound. But the other 99.8% are going to get WORSE sound because it's too much brain damage for them to understand it and know how to use it properly.

 

End part one (5000 character limit)

 

~~~~~~~~~~

 

The reason that we don't do DSD-128 is because it requires that the USB receiver be capable of doing 8 Fs sample rates. We use the XMOS and they make it in two speed grades. Most of our QB-9's have the faster processor, because it is only about $1 more and it is easier to find. But not all of them that faster processor. Who would have ever thought four years ago that any sane person would WANT 384/24??? So to do DSD-128, we would have to get the unit back, make sure it has the high speed grade part, change the firmware, test it, burn it in, re-test it, box it and ship it. If it DOESN'T have the high speed-grade part, we have to remove the old one, throw it away and replace it with the new one. That is about $10 for the part, but about $50 in labor. With all of the other stuff thrown in, we would have to charge AT LEAST $300 retail (the dealer wants some mark-up too).

 

Now how stupid would someone have to be to pay $300 to play 7 files? It's not like DSD-128 is EVER going to catch on. Even DSD-64 is a SUPER small, niche market product that only appeals to the 1% of computer audio users that think DSD is better than PCM of the 1% of audiophiles that use computer audio of the 0.1% of the general population that are audiophiles.

 

Trust me, the little companies like Channel Classics and Blue Coast and whoever is making that stuff will all be out of business in five years -- if that long. Look at how many companies have even been successful making STANDARD audiophile grade stuff -- Mobile Fidelity went out of business (twice). Classic Records went out of business and they made LP's and CD's and 96/24 DVD-V discs and 192/24 DVD-A discs.

 

So if you think that these new little companies that make formats that are literally 100 times smaller in audience than what MoFi or Classic can appeal to are going to survive, you are smoking crack cocaine. It ain't gonna happen.

 

The best hope for high res succeeding down the road is Pono. All the labels are actually interested in this for a lot of reasons. But what it will be is 192/24 PCM. And 20 years from now, you are going to have a REALLY hard time finding ANY equipment that will play back SACD's. Computer programs are a lot easier to make than hardware (which is why we are having this ridiculous proliferation of formats at all), but when I see guys like Miska seriously propose that he is going to start a new format using 9-bit delta-sigma modulation, I KNOW that guy is so far removed from reality that he will be lucky to be in business at all in two years if he doesn't wake up -- and fast!

 

The ONLY way that we as audiophiles are going to get what we want is to STANDARDIZE on ONE format. Then there is a slim hope of a prayer of a dream that we can create a critical mass so that the REAL record companies will give us some product that they can make money on.

 

They can't make money if we act like a bunch of spoiled little brats and say, "I don't LIKE that format. That format isn't as good as MY format. Nyah, nyah."

 

Right now the ONLY high res music you can buy that has a good selection of titles is HD Tracks. And they ONLY sell PCM. DSD is NOT a better sounding format. It was implemented better than DVD-A was a decade ago when Sony wasn't about to go out of business (like they are now). But properly implemented 192/24 sounds better than properly implemented DSD.

 

And NOBODY is EVER going to go past either 192/24 OR DSD-64. Why? Because it is all turning to downloads and streaming. Hell, we can't even get iTunes to give 44/16 because of the bandwidth required. And you think that they are going to give you DSD-256???? NO FUCKING WAY. Ain't gonna happen.

 

That's is what is so frustrating about these forums. Obviously there are a bunch of really smart people that know a lot about digital technology. But it's the same old question, over and over again, "How can someone who is so smart be SO FUCKING STUPID???"

 

People need to grow up and get a clue. 192/24 implemented right sound un-fucking believable. I think it sounds better than vinyl. But the implementation is what needs the work, not the format. The format is not the problem. If we could quit having these "smart" people pulling the market in umpteen different directions, we MIGHT be able to get 192/24 as a commercial reality. Then when people implement it properly, we will have the best sound of all time.

 

Cheers,

Charlie

[Miska]

You seem to be like don Quixote, tilting at windmills. Even on the TI data sheet for the PCM4222 they refer to SDM as a "Sigma Delta Modulator":

 

I use sigma-delta or delta-sigma, both can be found widely around in the datasheets.

 

(What TI refers to as a "DSD engine" is obviously the re-modulator.)

 

What's the big deal? I have a software implementation(s) of such.

 

It usually doesn't do much good. And when you speak with other engineers, they will have a difficult time understanding you since you use your own terminology....

 

So far I haven't experienced any problems. (I don't count forums where people act like they wouldn't understand or misunderstand on purpose)

 

Now you want to make up your own new format!

 

Of course, if something useful doesn't already exist, it needs to be invented. Good thing is that ASIO on Windows and ALSA on Linux already have a necessary transport vehicle for it.

 

I think it will be many, many times more difficult for a small, one-man company to do so. But good luck to you! I will stick with PCM, as I have been able to achieve excellent results with it and everybody already uses it.

 

Wonderful thing is that I don't need to convince anybody. I do what I need for myself, and sometimes I may offer that for other people to use too.

 

All I can say is that it is no wonder that you prefer DSD if you are using the standard digital filters supplied by Texas Instruments. If there is one lesson that people have learned from this thread, I hope it would be that the problem with PCM is due to the standard digital filters that everyone uses.

 

Good that we almost agree about something.

That's why I prefer to record the modulator output from the ADC straight and perform proper conversion in software. Although I'm introducing the new format because I don't think the conversion should be made at all and then converted back in the DAC. Just keep it at original sampling rate throughout. I can perform any processing I need in the original format.

 

But by using custom filters (and also being sure to use that wonderful machine between our ears), we can make PCM sound as good or better than DSD, yet have NO problem with out-of-band noise

 

That won't happen, because the rate down-conversion and then later conversion back up will have it's penalty. So better skip that part. Your delta-sigma modulator in DAC will re-create the out-of-band noise, and if you use leaky oversampling filters (less than 160 dB attenuation above Nyquist frequency) you will also add nasty ultrasonic image frequencies that are way worse than any out-of-band noise in DSD, because they are heavily correlated and cause all kinds of nasty correlated intermodulation effects at later stages.

 

I would encourage you to use your obvious supply of knowledge of digital systems to improve PCM instead of "tilting at windmills", trying to change industry-standard nomenclature and introducing your own formats.

 

I first did improvements for PCM, then I did it for DSD and now I'm going to do it for multi-bit SDM.

 

So my player can spit out oversampled noise shaped 1536/32 PCM, oversampled 24.576 MHz DSD and soon oversampled 24.576 9-level SDM.

 

Now give us DACs that take this stuff in so that I don't have to spend time building DACs for myself! For the PCM I have not yet seen such on the market. For DSD I have. Maybe we'll see the multi-bit SDM DACs in not too distant future too (so far, I just have to make my own).

 

But I find it pathetic if the PCM DAC actually needs to then go and convert it to some kind of SDM anyway. So if you do PCM, give us ladder DACs, if you do SDM, give us native SDM input.

 

Here's output of typical PCM DAC with 44.1 kHz sweep:

...speaking of out-of-band noise?

[wisnon]

Quote from Charles:

Here's an example -- have YOU oriented the AC plugs of every piece of equipment in your system? If not (and I doubt that you have) you are missing out on a lot of performance....

 

Actually Charles, I am in Switzerland and we have mainly 3 prong plugs here, so not much chance of getting it wrong.

 

I DO have a few 2 prong pieces of equipment, one is a Linear PSU for my Chord Dac, and I use Voltmeter to get it phase aligned! I also have an older Lampizator Silk power conditioner and will soon upgrade it with the Phase flipper where I will get the phase alignment for the 4 outlets by flipping thru the 16 combinations!

 

So yes, I do know about correct plug alignment and to tell you what...PhilipAC has one and while using a lower quality borrowed Dac, he was bummed out until he played around with the flippes switches on his Silk and WOW, the Dac improved immensely, he told me!

 

+++++++

 

Part 2.

 

Here is my take Charles. On my main Gen 4 L4 Lampi Dac, even RBCD sounds like Hirez.

 

However, my take is that a pure DSD recording will ALWAYS sound best via DSD and PCM will sound better via PCM.

There is no real complexity to the consumer in having a Dac that can play both natively. Here is what Lukasz said about his DSD feature:

 

DSD project works ! It is OK without noises and perfect !

It sounds awesome. Orchestra dynamics are explosive !

 

I trust Lukasz implicitly not to BS me, so if he is impressed, you better believe it sounds good and his R2R current output PCM Dac is a killer.

[mav52]

You sent a PM, but your inbox is full and I can't reply. Here is your answer.

 

 

 

Hello Wisnon,

 

Thanks for the message. I didn't realize you had sent it as the default setting are NOT to notify me. Weird...

 

Anyway, thanks for the kind words. If I get cranky, please don't take it personally. I was in an accident and paralyzed and have to take a lot of pain meds and sometimes it affect what I write -- especially when I do so "off the cuff" -- then it comes off the wrong way, like a personal attack, which is almost NEVER what I mean.

 

Your example of a cassette is a good one, and it is an example of the type of complexity that I would like to avoid. I still have a really nice Sony 3 head, 3 motor cassette deck -- just one notch below a Nakamichi. Sometimes I look at that thing and wonder how often the AVERAGE user ever got ANYTHING right.

 

I used to work at a stereo shop and what set us apart was our service department. The owner had been the chief tech at a shop for close to twenty years. So when he started his own shop, his USP (unique selling proposition) was that when you bought a piece of equipment from him, it was GUARANTEED to meet the specs.

 

Nowadays it is not so important but in the old days there were only three sources -- LP, FM, and cassette. Well FM and cassette will ONLY meet the specs if they are "aligned" properly. To do a proper alignment on either one took about an hour for a good tech. When they build them at the factory there is no way in the world that they had time to make all the dozens of adjustments required to get them to meet the spec.

 

But at our store, we would open up the box, take out the unit and fine tune every single setting in there to get it to work at the highest possible level. The factory would just measure the voltage at a test point and set that to some "average" that would get the performance in the ballpark, but there was too much variation in the individual parts to get it to actually meet the spec.

 

The closest thing we have to that today is with a video display. If you ever read a home theater magazine, they will talk about "calibrating" the display, and that is the same thing. The magazines have to do it as otherwise you would NEVER know how good the design really is. BUT if YOU don't have YOUR set calibrated also (typically $300 to $500), you won't get the proper results either.

 

If you are buying a $8000 projector, you would be stupid not to have your display calibrated. But how many people that buy $800 "flatscreens" at Costco are going to calibrate it? ZERO....

 

Back to cassettes. I was looking at my deck the other day and there are so damned many switches and knobs and dials, that I doubt that more than 0.2% of the people using them ever got the most out of them. Even if it had been calibrated, it had to be for a specific brand and model of tape. And if they ever changed (improved) the formulation, you had to have it re-calibrated.

 

People these days don't have the time or the patience for all of that nonsense. Even with something as simple-minded as a USB DAC, how many people do you think really go out and get a good aftermarket player and go through ALL of the settings to get the best sound quality possible.

 

Here's an example -- have YOU oriented the AC plugs of every piece of equipment in your system? If not (and I doubt that you have) you are missing out on a lot of performance....

 

So the LAST thing we need is MORE formats and MORE variations and MORE choices. If we go down that road, then MAYBE 0.2% of the people will know enough and get better sound. But the other 99.8% are going to get WORSE sound because it's too much brain damage for them to understand it and know how to use it properly.

 

End part one (5000 character limit)

 

~~~~~~~~~~

 

The reason that we don't do DSD-128 is because it requires that the USB receiver be capable of doing 8 Fs sample rates. We use the XMOS and they make it in two speed grades. Most of our QB-9's have the faster processor, because it is only about $1 more and it is easier to find. But not all of them that faster processor. Who would have ever thought four years ago that any sane person would WANT 384/24??? So to do DSD-128, we would have to get the unit back, make sure it has the high speed grade part, change the firmware, test it, burn it in, re-test it, box it and ship it. If it DOESN'T have the high speed-grade part, we have to remove the old one, throw it away and replace it with the new one. That is about $10 for the part, but about $50 in labor. With all of the other stuff thrown in, we would have to charge AT LEAST $300 retail (the dealer wants some mark-up too).

 

Now how stupid would someone have to be to pay $300 to play 7 files? It's not like DSD-128 is EVER going to catch on. Even DSD-64 is a SUPER small, niche market product that only appeals to the 1% of computer audio users that think DSD is better than PCM of the 1% of audiophiles that use computer audio of the 0.1% of the general population that are audiophiles.

 

Trust me, the little companies like Channel Classics and Blue Coast and whoever is making that stuff will all be out of business in five years -- if that long. Look at how many companies have even been successful making STANDARD audiophile grade stuff -- Mobile Fidelity went out of business (twice). Classic Records went out of business and they made LP's and CD's and 96/24 DVD-V discs and 192/24 DVD-A discs.

 

So if you think that these new little companies that make formats that are literally 100 times smaller in audience than what MoFi or Classic can appeal to are going to survive, you are smoking crack cocaine. It ain't gonna happen.

 

The best hope for high res succeeding down the road is Pono. All the labels are actually interested in this for a lot of reasons. But what it will be is 192/24 PCM. And 20 years from now, you are going to have a REALLY hard time finding ANY equipment that will play back SACD's. Computer programs are a lot easier to make than hardware (which is why we are having this ridiculous proliferation of formats at all), but when I see guys like Miska seriously propose that he is going to start a new format using 9-bit delta-sigma modulation, I KNOW that guy is so far removed from reality that he will be lucky to be in business at all in two years if he doesn't wake up -- and fast!

 

The ONLY way that we as audiophiles are going to get what we want is to STANDARDIZE on ONE format. Then there is a slim hope of a prayer of a dream that we can create a critical mass so that the REAL record companies will give us some product that they can make money on.

 

They can't make money if we act like a bunch of spoiled little brats and say, "I don't LIKE that format. That format isn't as good as MY format. Nyah, nyah."

 

Right now the ONLY high res music you can buy that has a good selection of titles is HD Tracks. And they ONLY sell PCM. DSD is NOT a better sounding format. It was implemented better than DVD-A was a decade ago when Sony wasn't about to go out of business (like they are now). But properly implemented 192/24 sounds better than properly implemented DSD.

 

And NOBODY is EVER going to go past either 192/24 OR DSD-64. Why? Because it is all turning to downloads and streaming. Hell, we can't even get iTunes to give 44/16 because of the bandwidth required. And you think that they are going to give you DSD-256???? NO FUCKING WAY. Ain't gonna happen.

 

That's is what is so frustrating about these forums. Obviously there are a bunch of really smart people that know a lot about digital technology. But it's the same old question, over and over again, "How can someone who is so smart be SO FUCKING STUPID???"

 

People need to grow up and get a clue. 192/24 implemented right sound un-fucking believable. I think it sounds better than vinyl. But the implementation is what needs the work, not the format. The format is not the problem. If we could quit having these "smart" people pulling the market in umpteen different directions, we MIGHT be able to get 192/24 as a commercial reality. Then when people implement it properly, we will have the best sound of all time.

 

Cheers,

Charlie

 

Dam Charles how do you really feel, after all your a designer.

[digimuse]

Charles,

 

I'd like to thank you for taking the time to enlighten us in this thread.

 

There is a lot of misinformation in public forums about the superiority of DSD. It's fine to posit opinions, likes/dislikes, but to confuse one's opinions with truths is only self-serving. This is especially problematic when (false) technical claims are made. Unfortunately, there is no review by experts to correct these errors, or to have an intelligent debate on the tradeoffs. The problem is compounded by commercial interests which benefit from propagating lies, which are all too easily accepted as truths by non-specialists (like me).

 

The result is that we are left with a sea of non-expert proclamations and rants from which the truth is impossible to discern.

 

However, at least in this thread, you have shed some light. I want to thank you for your educational postings, and debunking:

 

• proof by volume (I make bold claims, and I post often, so I'm right)
  
• proof by blathering (I'm great at putting together words and sentences that appear coherent and intelligent)
  
• proof by repetition (we +1 each others' postings, and keep repeating the same untruths, so it must be true)
  
• proof by I'm an expert (I build systems with DAC components, so I'm right)
  
• proof by I'm a (small) businessman (I'm going to create a brand new format/standard for doing digital audio right)
  
• proof by attacking the messenger (it's so much harder to defend a position that is shown to be incorrect - much easier to deflect and question the person contradicting your cherished beliefs).
  

Computer audio is a hobby for all of us. We strive to move forward and continually make improvements (although perhaps it would be better to spend more time listening to music). We research various alternatives, audition different systems, and go forward with a purchase that we think is the best way to move our hobby forward. A new opportunity, like SACD/DSD, sounds appealing, especially with the marketing BS put forth by Sony. Having committed to a choice, it's very deflating to accept criticisms of this choice (how could I have made a sub-optimal choice?).

 

It would be very helpful if record companies/music retailers were clear about exactly what was done in the recording, editing, and mastering process. Without this, there are too many variables to know why a recording sounds good or bad (DSD/PCM, bit rate, sample rate, filters used, etc.).

 

I myself have fallen prey to the exaggerated claims of SACD/DSD. I have one of those "ringy-ding" DACs that has an upsampler box upstream transmitting encrypted DSD signals (they're driving two Ayre monoblocks made of some billeted metal).

 

However, after reading all your posts in this thread, I'm convinced that the only value DSD provides is when all mixing/editing is done in the analog domain, and this final analog signal is directly fed to a DSD D/A converter (and on playback a DSD D/A converter is used). I don't believe I have any such recording. Converting from DSD to PCM, doing the editing/mastering, and then converting the PCM to DSD degrades audio quality (especially if you go between DSD and PCM repeatedly).

 

Perhaps future high-resolution PCM recordings will abandon brick-wall filters, and we can have more recordings showing what the best of digital audio has to offer. I tried out the John Marks demo of "Jul" in 192 kHz/ 24 bit and thought it had a very natural, warm, realistic, and un-harsh sound (digital recordings often sound grainy - like an unnatural over-sharpened photo).

 

Keep up the great work!

[elcorso]

Charles,

 

I'd like to thank you for taking the time to enlighten us in this thread.

 

There is a lot of misinformation in public forums about the superiority of DSD. It's fine to posit opinions, likes/dislikes, but to confuse one's opinions with truths is only self-serving. This is especially problematic when (false) technical claims are made. Unfortunately, there is no review by experts to correct these errors, or to have an intelligent debate on the tradeoffs. The problem is compounded by commercial interests which benefit from propagating lies, which are all too easily accepted as truths by non-specialists (like me).

 

The result is that we are left with a sea of non-expert proclamations and rants from which the truth is impossible to discern.

 

However, at least in this thread, you have shed some light. I want to thank you for your educational postings, and debunking:

 

• proof by volume (I make bold claims, and I post often, so I'm right)
  
• proof by blathering (I'm great at putting together words and sentences that appear coherent and intelligent)
  
• proof by repetition (we +1 each others' postings, and keep repeating the same untruths, so it must be true)
  
• proof by I'm an expert (I build systems with DAC components, so I'm right)
  
• proof by I'm a (small) businessman (I'm going to create a brand new format/standard for doing digital audio right)
  
• proof by attacking the messenger (it's so much harder to defend a position that is shown to be incorrect - much easier to deflect and question the person contradicting your cherished beliefs).
  

Computer audio is a hobby for all of us. We strive to move forward and continually make improvements (although perhaps it would be better to spend more time listening to music). We research various alternatives, audition different systems, and go forward with a purchase that we think is the best way to move our hobby forward. A new opportunity, like SACD/DSD, sounds appealing, especially with the marketing BS put forth by Sony. Having committed to a choice, it's very deflating to accept criticisms of this choice (how could I have made a sub-optimal choice?).

 

It would be very helpful if record companies/music retailers were clear about exactly what was done in the recording, editing, and mastering process. Without this, there are too many variables to know why a recording sounds good or bad (DSD/PCM, bit rate, sample rate, filters used, etc.).

 

I myself have fallen prey to the exaggerated claims of SACD/DSD. I have one of those "ringy-ding" DACs that has an upsampler box upstream transmitting encrypted DSD signals (they're driving two Ayre monoblocks made of some billeted metal).

 

However, after reading all your posts in this thread, I'm convinced that the only value DSD provides is when all mixing/editing is done in the analog domain, and this final analog signal is directly fed to a DSD D/A converter (and on playback a DSD D/A converter is used). I don't believe I have any such recording. Converting from DSD to PCM, doing the editing/mastering, and then converting the PCM to DSD degrades audio quality (especially if you go between DSD and PCM repeatedly).

 

Perhaps future high-resolution PCM recordings will abandon brick-wall filters, and we can have more recordings showing what the best of digital audio has to offer. I tried out the John Marks demo of "Jul" in 192 kHz/ 24 bit and thought it had a very natural, warm, realistic, and un-harsh sound (digital recordings often sound grainy - like an unnatural over-sharpened photo).

 

Keep up the great work!

 

The same as you I do appreciate Charles posting in this thread and enlighten us. He has a lot of courage posting in an audio forum being an analog designer/owner of one the most valuable audiophile audio brands, Ayre.

 

But also you can't trash away Miska postings, being a successful digital engineer, at least that you have some credentials on this to show us. He also has the courage to post here in C.A. and enlighten us too.

 

I'm talking here about courage, since a lot of 'audiophiles' believes they know better than audio, digital audio and software engineers, then ends even insulting this courageous people.

 

I'm sorry you have a "ringy-ding" DAC that doesn't please you on DSD playback, there are a lot of good ones, and I'm sure the Ayre QB-9 (DSD capable) will also match your spectation feeding the extraordinary Ayre mono blocks.

 

I also DL the John Marks demo of "Jul" in 192 kHz/ 24 bit and it's an extraordinary PCM recording, as all John's recordings, but unfortunately "one swallow does not make a summer". He used the brick-wall free Ayre's ADC, but also, did you see the microphones and mic. placement, etc., from the pictures?

 

I do agree that in order to get an excellent SQ (from PCM or DSD) the first thing is the recording (microphones, mixing and editing procedures, et al). But I have also a lot of extraordinary DSD recordings (as I have in PCM) with all the adjectives you used to qualify "Jul" recording.

 

I'm not an expert, either, but a DSD & PCM lover, and my 'testing' is by listening, and until now learning theory from this thread and others.

 

Happy listening (in the format you like better),

 

Roch

[beetlemania]

So if you think that these new little companies that make formats that are literally 100 times smaller in audience than what MoFi or Classic can appeal to are going to survive, . . . It ain't gonna happen.

 

The best hope for high res succeeding down the road is Pono. . . . it will be is 192/24 PCM. . . .

 

The ONLY way that we as audiophiles are going to get what we want is to STANDARDIZE on ONE format. . . .

 

They can't make money if we act like a bunch of spoiled little brats and say, "I don't LIKE that format. That format isn't as good as MY format. Nyah, nyah."

 

. . . But properly implemented 192/24 sounds better than properly implemented DSD.

 

And NOBODY is EVER going to go past either 192/24 OR DSD-64. Why? Because it is all turning to downloads and streaming. Hell, we can't even get iTunes to give 44/16 because of the bandwidth required. And you think that they are going to give you DSD-256???? NO FUCKING WAY. Ain't gonna happen.

 

 

People need to grow up and get a clue. 192/24 implemented right sound un-fucking believable. I think it sounds better than vinyl. But the implementation is what needs the work, not the format. The format is not the problem. If we could quit having these "smart" people pulling the market in umpteen different directions, we MIGHT be able to get 192/24 as a commercial reality. Then when people implement it properly, we will have the best sound of all time.

 

Excellent post, Charles. You've made several highly educational posts in this thread and that last one really brings it all together with a cogent argument for why audiophiles should get behind 24/192.

 

I've been lovin' the QB-9. Computer audio has some quirks but I'm getting the best sound out of my system with hi-rez. I used to have a CX-7. After the MP upgrade I thought that it was pretty close to vinyl but with the best vinyl still sounding better than the best RBCD. With the QB-9, the tables are turned. My best examples of hi-rez sound a bit better than my best vinyl.

 

Interesting your exchanges with Miska re: standard filters. After I got the QB-9, I tried pretty much every PC player I could find (except XXHighend which has the most byzantine installation imaginable - truly for the computer geek crowd). I found them them all to sound pretty much the same (tho' at least one other QB-9 owner reports differently) and I have to wonder whether your filters (and other design choices) have lead to this result whereas owners of other DACs report substantial differences among music players. If I had to choose the best, I'd pick JPlay, but the difference is so subtle I'd rather put the money to more music (or the DSD upgrade). If I had to pick the worst, I'd pick Miska's HQPlayer. It has many filter settings. On most of them, it sounded pretty much the same as Foobar or JRMC but on some it sounded a bit "dull". I have to wonder if some of the HQPlayer filters are redundant with the QB-9's, leading to some negative interactions. At any rate, you and Miska have different equipment and different experiences. IME, the QB-9 kicks ass and doesn't need help from ancillary software - I'm really looking forward to the DSD upgrade (and not because I'll be able to play DSD files)!

 

One potential boon for DSD capability is if SACD titles ever become available as DSD downloads. That would be pretty cool (unless they charge $30-40 per title). In the meantime, your choice not to include DSD128 seems entirely reasonable. As I wrote in a previous post, there's barely any DSD64 I'm interested in, much less the pittance of DSD 128 titles.

[Charles Hansen]

Quote from Charles:

Actually Charles, I am in Switzerland and we have mainly 3 prong plugs here, so not much chance of getting it wrong.

 

Hello Wisnon,

 

Three-prong plug does not help unless the transformer manufacturer always winds it correctly and the equipment manufacturer always corrects it correctly. Probably less than 20% of all equipment has both done correctly so that you can count on proper polarity.

 

Then many times DIY homeowner replace the outlets, or DIY audiophiles put in "super" outlets and they don't know and put it in backwards....

 

Only way to know for sure is to measure each component separately.

 

Cheers,

[Charles Hansen]

Charles,

 

I'd like to thank you for taking the time to enlighten us in this thread.

 

Thank you for the kind words.

 

Both high-res PCM (especially quad-rate) and DSD are capable of excellent sound. Most DSD releases in the beginning of the format war 12 years ago sounded better because Sony paid Ed Meitner and Andreas Koch to build hardware with EXCELLENT implementation.

 

Probably the best high-res PCM ADC for many years was the Pacific Microsonics Model Two, but that was not released until 2003 (I think) and was very expensive. Plus Pacific Microsonics played the LYING game a lot when it came to the HDCD process. But the box itself was designed by Keith Johnson and was EXCELLENT.

 

During that time it was easy to make a good sounding SACD because the ONLY converters available were owned by Sony and lent to the studios. But in the end, when they are both done PROPERLY there is not much difference between them. But if you look at how many studios are fully equipped with PCM and how many studios are fully equipped with DSD, it is at least 1,000 to 1 ratio. So if you like music (and not just pretty sounds), the reality is that we live in a PCM world.

 

DSD failed once and it will fail again. Do you think iTunes will offer DSD downloads? What about Amazon? They used to sell SACD's -- do you think the will start selling DSD downloads? Do you think Spotify and Mog will start streaming DSD? Or do you think they will start streaming DSD-256??? Or Miska's 9-bit raw Delta-Sigma output??????

 

No, of course none of these things will ever happen.

 

The only thing we can hope for is high-resolution PCM. All Universal Music companies (Deutsche Gramophon, EMI, et cetera) all record in 96/24 at a minimum now. Warner is said by Neil Young to have transferred 9,000 (albums or tracks - it is not clear) from analog to 192/24. Every week there are three or four new Warner high-res titles on HD Tracks. All of the record companies are looking at Pono VERY seriously. Most Pono will be 192/24.

 

When 192/24 is done PROPERLY, it sounds better than DSD.

 

So the obvious thing to do is keep improving the implementation of PCM. It is very simple. It is simple to see that is the future. It is simple to improve it. It is simple to understand the advantages of DSD and incorporate them into high-res PCM. It is simple for the studios to record PCM. It is simple for them to keep using the equipment they are familiar with.

 

Music is a big business. The companies that control it want to make as much money as possible. That is not done by changing to a new format. The musicians that make music want to sell as many albums as possible. If they can sign with Warner or they can sign with Blue Coast, which company do you think they will choose? This is not rocket science...

 

The only reason that DSD even exists is because Sony and Philips patents on CD were expiring. They couldn't patent high-res PCM. So they took something that didn't make very much sense and turned it into a format so they could license it. The big prize for the record companies was that nobody could copy it with a computer. That seemed like a good idea to some people 15 years ago when Napster was giving away MP3 for free. But now the only reason the DSD has a second life is because they made a version that will play on the computer.

 

But just because you can DO something, doesn't make it a good idea...

 

Best regards,

[Charles Hansen]

I'm not an expert, either, but a DSD & PCM lover, and my 'testing' is by listening, and until now learning theory from this thread and others.

 

Happy listening (in the format you like better)

 

Hello Corso,

 

Yes, both are capable of excellent sound. But only one will survive. It is only a matter of time. Nobody wants two formats. Not the record companies, not the retailers, not the customers. If you look at history, every time there is a format war, one will always lose.

 

So VHS beat Beta (even though Beta was better). Blu-ray beat HD-DVD (even though HD-DVD made much more sense, Sony knew the right people to bribe -- they owned a movie studio!). Sometimes two formats can co-exist because they serve different needs. So cassette and LP coexisted. But cassette beat 8-track.

 

If DSD were inherently a superior format that was capable of much better sound than high-res PCM, it MIGHT survive. But even then probably not, because it is too expensive to convert all the studios and too difficult to process in modern recording methods. But since it is not hard to make high-res PCM that has all of the advantages of DSD, it is only a matter of time before DSD dies off. That doesn't mean it is bad or it sounds bad. It just doesn't offer any real advantage and it has many, many practical disadvantages.

 

Best regards,