"Chord Electronics FPGA DAC Technology Explained" - What went wrong?

[amirm]

What I said was not nonsense. You started your review with a bunch of complaints about the DAVE that had nothing to do with sound. Because of that, I anticipated you would trash it because you don't like Chord products in general. That's an honest description of what I thought. But, from everything you have said since, it's clear that what you hated about the DAVE was the oversampling issue. You hate it not because it's the king of Chord but because it's the king of taps.

Of course it is nonsense. The core of my reviews are objective measurements that don't change because I like or dislike the look of something. The instrument makes those measurements and I can't impact that with my feelings. My reviews always include a few comments about the look and operation of the device. These observations are orthogonal to the technical performance.

In sharp contrast, you or random subjectivist reviewer relying on faulty listening tests and tons of other bias, most definitely are influenced by non-sound factors. A reviewer getting a $10K DAC would die before he would say it sounds the same as an $80 DAC. Me on the other hand, show measurements that can very well show that as a fact.

I also don't "hate" oversampling whatsoever. It is an overkill feature but I have not stated that the technique is a bad thing let alone hating it. Do I think it is wasted of money, sure. But that doesn't make me hate it.

As I said you are just wasting our time with these silly arguments. Facts are these:

1. Chord products in general are competent and have noise and distortions that are not audible to audiophiles.

2. They over-emphasize filtering when there is no evidence that they do anything useful audibly, yet add considerable cost to products.

3. They have odd and in some cases, terrible user interfaces. Some may like this, others like me don't.

4. They have the mother of all marketing people preaching non-scientific things about them in a sponsored forum. People there buy into it while closing their eyes to realities of the technology and what it does for them.

5. Sighted observations from the marker or reviewers/owners is useless. I can take two identical devices and put them in different boxes and get people to say one is better. And that the difference is night and day. This is what is wrong with your only data point you are using to praise these technologies. Learn how your body works for heaven's sake.

6. There are very high performance DACs that cost as low as $90 that are objectively better, and subjectively the same for just about anyone.

There is really nothing else here.

 

[Geert]

You most certainly do not know that these taps don't provide audible benefits. Show me the graph that displays audible benefits. Since you're the one making the assertion, you have the burden of proof.

Taps are a well known concept in the realm of digital signal processing (which is much wider than audio). They are part of the implementation of digital filters. How taps work is pure math; how much taps you need to meet the quality requirements (frequency resultion) of a filter is a calculation. And the results of that calculation can be measured objectively.

As such the burden of proof is with the people who claim that for audio we need more taps than science tells us.

Here's some more testing of that claim by an expert member https://www.audiosciencereview.com/...nd-better-than-a-16k-one-lets-find-out.36793/

 

[Keith_W]

Taps are a well known concept in the realm of digital signal processing (which is much wider than audio). They are part of the implementation of digital filters. How taps work is pure math; how much taps you need to meet the quality requirements (frequency resultion) of a filter is a calculation. And the results of that calculation can be measured objectively.

As such the burden of proof is with the people who claim that for audio we need more taps than science tells us.

Here's some more testing of that claim by an expert member https://www.audiosciencereview.com/...nd-better-than-a-16k-one-lets-find-out.36793/

Can you help me understand how taps are relevant to DAC's? I understand how they are relevant for DSP, but not DAC's. I get from this thread that more taps = higher oversampling, but what is the exact calculation, and how much oversampling is considered "adequate", and how much is "overkill"?

(Again, not challenging or arguing, I only want to know the answer!)

 

[BigEarsBill]

Taps are a well known concept in the realm of digital signal processing (which is much wider than audio). They are part of the implementation of digital filters. How taps work is pure math; how much taps you need to meet the quality requirements (frequency resultion) of a filter is a calculation. And the results of that calculation can be measured objectively.

As such the burden of proof is with the people who claim that for audio we need more taps than science tells us.

Here's some more testing of that claim by an expert member https://www.audiosciencereview.com/...nd-better-than-a-16k-one-lets-find-out.36793/

Thanks for the civil reply. I'll check out the link.

 

[IAtaman]

Thanks for the civil reply. I'll check out the link.

Frankly, you should have done that before you started looking for comparable taps to Dave, but better late than never I guess.

Can you help me understand how taps are relevant to DAC's? I understand how they are relevant for DSP, but not DAC's. I get from this thread that more taps = higher oversampling, but what is the exact calculation, and how much oversampling is considered "adequate", and how much is "overkill"?

(Again, not challenging or arguing, I only want to know the answer!)

Simply put, If you have an oversampling DAC, you need the filter out the frequencies above Nyquist otherwise say a 32KHz signal might be creating the same samples as 12KHz signal would, distorting the sound. You use FIR filters to do that.

 

[voodooless]

Can you help me understand how taps are relevant to DAC's? I understand how they are relevant for DSP, but not DAC's.

The concept is exactly the same, the DAC contains some DSP bits.

I get from this thread that more taps = higher oversampling, but what is the exact calculation, and how much oversampling is considered "adequate", and how much is "overkill"?

More taps != more oversampling. It just gives you to ability to make steeper filters with lower stop bands and less band pass ripple. The oversampling (in case of synchronous oversampling) is the process of adding zero samples between the real ones, and the low passing this using convolution. The convolution bit is where the taps come in. “Normal” DACs over sample to a much higher rate, often 64 to 128x, and then use only a filter with a few hundred taps at best. And it worlds perfectly fine.

 

[BDWoody]

I will reply when I think you have misrepresented what I have said or are saying something with which I disagree.

Not anymore. Not enough of a good faith effort to actually understand.

Just...too much.

Time out.

 

[DSJR]

Bottom line in some ways is this I think. There are many newly retired men (usually) with bulging pension pots who want to indulge their audio and music interest. The kind-of steam punk looks and quirkiness and 'story' of the Chord digital gear can seriously interest them, it's generally well made and good after sales service is offered here in the home market should it be needed. Price in this instance isn't at all an issue for these people. What they're buying are luxury eye-candy digital playback 'goods' with generally good performance and a tale to tell, but in my case, I'm so glad the sub $/£500 level has seen such huge strides in generally excellent but rather simpler dacs

I've had a few dems of Chord gear and knew hands-on the now old 1200E amp. Sure, the M-Scaler made an audible difference when their rep demo'd it, but I felt that the sound was slightly quieter when it was switched in via a button on it, thereby 'enhancing' the sense of depth-of-image. It turns out the levels are reduced with it working, thereby making a comparison such as this totally useless frankly. I'd hardly say the company's amps are especially transparent either, being very 'dry' in perceived tone (I say perceived here as it's a subjective thing until they're thoroughly tested objectively)...

Mr Watts does claim it's the *effects* of the -300dB issues he claims to hear, rather than actual levels and in the talk he gave (I had to button my lip during said talk on a few occasions, but the audience were nodding away in acceptance regardless), he made a huge thing also about 'noise floor modulation' which the Dave here seemed to exhibit (is this also a non-event in the inexpensive utility dacs tested and loved here?)

 

[antcollinet]

Show me the graph that shows that.

They show that as a suite of tests together.

There is only frequency response, noise, and distortion, and we can measure those to an accuracy far greater than human hearing. We also have a good understanding of the levels of those things the human ear is cabpable of detecting.

So when a device measures below those levels we know it doesn't have a "sound"


Beyond that - if we have it wrong, it is easy for you (or anyone) to prove that by a properly executed blind test showing you can hear a difference when the measurements say you can't. But despite all the motivation from 1000's of outraged audiophiles, and invested manufacturers - so far no-one has been able to do so.

 

[Mart68]

They show that as a suite of tests together.

There is only frequency response, noise, and distortion, and we can measure those to an accuracy far greater than human hearing. We also have a good understanding of the levels of those things the human ear is cabpable of detecting.

So when a device measures below those levels we know it doesn't have a "sound"


Beyond that - if we have it wrong, it is easy for you (or anyone) to prove that by a properly executed blind test showing you can hear a difference when the measurements say you can't. But despite all the motivation from 1000's of outraged audiophiles, and invested manufacturers - so far no-one has been able to do so.

It's not easy for the ordinary punter to construct a blind test that has results that would be accepted.

But there's no excuse when it comes to the manufacturers, or the magazines, or the larger subjective forums, who do have the resources and could commission a neutral third party to organise and document the test. If the results showed that at least some people could reliably hear the claimed improvements the expense would easily be recouped in increased sales.

That they never do this is really the prime indicator that their claims are bogus - and that they know it.

 

[IAtaman]

It's not easy for the ordinary punter to construct a blind test that has results that would be accepted.

But there's no excuse when it comes to the manufacturers, or the magazines, or the larger subjective forums, who do have the resources and could commission a neutral third party to organise and document the test. If the results showed that at least some people could reliably hear the claimed improvements the expense would easily be recouped in increased sales.

That they never do this is really the prime indicator that their claims are bogus - and that they know it.

pkane's test that was linked above is pretty conclusive as far as I can tell. Any difference between a 16K tap filter vs a 4M tap filter is less than 250db below 20KHz. I don't think any listening validation is required.

Does a 4 million taps FIR filter sound better than a 16K one? Let's find out.

Claim A claim is frequently made that DAC digital filters with a lot of taps can make a significant difference in sound quality. Some software packages and even hardware manufacturers are touting the benefits of large FIR filters, often with millions of taps, supposedly because these produce a...

audiosciencereview.com

 

[Geert]

It's not easy for the ordinary punter to construct a blind test that has results that would be accepted. But there's no excuse when it comes to the manufacturers

Shouldn't be to difficult for Chord, since this is what Watt's has to say about the Hugo M Scaler:

"That is, you will hear better depth and detail resolution, unless you are deaf, or not blessed with good hearing, or not interested in the SQ and musical performance, in which case why on earth would you be interested in an expensive up-sampler".

 

[Mart68]

pkane's test that was linked above is pretty conclusive as far as I can tell. Any difference between a 16K tap filter vs a 4M tap filter is less than 250db below 20KHz. I don't think any listening validation is required.

Does a 4 million taps FIR filter sound better than a 16K one? Let's find out.

Claim A claim is frequently made that DAC digital filters with a lot of taps can make a significant difference in sound quality. Some software packages and even hardware manufacturers are touting the benefits of large FIR filters, often with millions of taps, supposedly because these produce a...

audiosciencereview.com

I quite agree but then we get wave after wave of people who have seen the test bench results but still ask 'But have you listened to it?'

Telling them to do a blind test and come back with the results showing they can hear the difference is pointless because if the result was positive we quite rightly wouldn't believe them. Their test methodology would be called into question, they would have no independent scrutiny or documentation to verify the test, and so forth.

 

[Mart68]

Shouldn't be to difficult for Chord, since this is what Watt's has to say about the Hugo M Scaler:

"That is, you will hear better depth and detail resolution, unless you are deaf, or not blessed with good hearing, or not interested in the SQ and musical performance, in which case why on earth would you be interested in an expensive up-sampler".

Exactly. He could do it tomorrow. But he won't and we all know why.

Quoting the usual rubbish from the Audiophile Rainbow is cheap and easy and good enough to convince the uneducated punter, so there's no need for him to prove anything.

 

[Geert]

Telling them to do a blind test and come back with the results showing they can hear the difference is pointless because if the result was positive we quite rightly wouldn't believe them.

Sorry but that's now true. If they can demonstrate that they've performed a reliable test, a lot of people here would by very interested. I have no dog in this fight, I'm an engineering that wants to learn.

 

[Mart68]

Sorry but that's now true. If they can demonstrate that they've performed a reliable test, a lot of people here would by very interested. I have no dog in this fight, I'm an engineering that wants to learn.

That element is massively non-trivial.

 

[Geert]

That element is massively non-trivial.

But not that big a challenge that it was totally impossible for hundred thousand of manufacturers and audio enthousiast in more than 20 years of digital audio. Amir measured more than 400 DAC's in a few years on his own. I expect us to have fusion energy before we see some evidence of audibility of taps and what not.

 

[fpitas]

Not anymore. Not enough of a good faith effort to actually understand.

Just...too much.

Time out.

This must be the new trendiness: make an ignorant PITA of yourself on ASR until you get booted. I guess we should be honored. Or something.

 

[Mart68]

But not that big a challenge that it was totally impossible for hundred thousand of manufacturers and audio enthousiast in more than 20 years of digital audio. Amir measured more than 400 DAC's in a few years on his own. I expect us to have fusion energy before we see some evidence of audibility of taps and what not.

measured yes, but setting up a blind test requires a lot more effort. And if it comes up with surprising results you are going to have to show independent verification for those results.

A while back it was proposed that a blind test was set up to demonstrate the unmeasurable differences in DACs that 'Goldensound' was saying he could hear.

That proposal fell down for a number of reasons, one of them was finding someone who had the knowledge and experience to organise and oversee the test and who was acceptable to all parties as independent and credible.

We've had people here before claiming that they have successfully identified differences in power cables and other highly unlikely differences in blind testing, they are not believed and their claims are dismissed since they offer no reliable verification that such a test ever even happened, let alone was conducted correctly.

Like you I would love to see it happen regardless of the outcome but it's not a simple task.

 

[IAtaman]

Sorry but that's now true. If they can demonstrate that they've performed a reliable test, a lot of people here would by very interested. I have no dog in this fight, I'm an engineering that wants to learn.

This part I don’t get. If a null test is saying 250db difference in audible range, what kind of engineering curiosity would drive you to look for a listening test? Do you think the test is missing something?