Watts Up...?

[Rob Watts]

It was designed for the B&W 803 D3, which has the lowest impedance of 3 ohms. So it will be fine so long as it doesn't fall below 4 ohms nominal and 3 ohms minimum.

 

[musickid]

edit

 

[GuiltyRocker]

Thanks Rob

If a speakers printed specs simply say 4 ohm impedance how can i be sure and know that the speakers impedance doesn't fall below 4 ohms nominal and 3 ohms minimum in real use? Assuming i dont have the means or knowledge to take actual measurements.

Try to find a technical measurement review, if the speaker was reviewed by Stereophile they usually have the impedance graph with explanation. If not, call the manufacturer and ask them what is the lowest impedance of the speaker and at what frequency that is. Then report to Rob what you find.

 

[musickid]

Ty

i found the following description regarding the audio physic classic 25 floor stander im thinking of purchasing (as a treat for my 50th next month) . driven by tt2 single ended.

"The critical impedance minimum is to be found at 3 ohms @ 80Hz.The impedance shows an uncritical phase response"

the last point if Rob could clarify is how i can make sure that the nominal impedance of this speaker doesn't fall below 4 ohms in real use. its printed specs state 4 ohms impedance. without having the means to physically measure this.

 

[Daorom]

..

 

[Rob Watts]

Ty

i found the following description regarding the audio physic classic 25 floor stander im thinking of purchasing (as a treat for my 50th next month) . driven by tt2 single ended.

"The critical impedance minimum is to be found at 3 ohms @ 80Hz.The impedance shows an uncritical phase response"

the last point if Rob could clarify is how i can make sure that the nominal impedance of this speaker doesn't fall below 4 ohms in real use. its printed specs state 4 ohms impedance. without having the means to physically measure this.

That sounds fine, as it's lowest impedance is the same as my 803D3.

 

[musickid]

many thanks to everyone for the support. the german audio physic towers should be with me within 2 weeks. tt2 single ended to spkrs and xlr to powered sub as im doing now.

 

[Jawed]

@Rob Watts it occurred to me that a key constraint of any correctly implemented digital to analogue convertor is that if time were ran backwards, the resulting output would be a time-reversed version of the output produced by running the DAC normally. You'd take the input file, reverse the order of the samples and then run it through the DAC, for comparison with the original file.

Is that practicably possible? Is it useful as a way of thinking about DACs?

 

[Rob Watts]

@Rob Watts it occurred to me that a key constraint of any correctly implemented digital to analogue convertor is that if time were ran backwards, the resulting output would be a time-reversed version of the output produced by running the DAC normally. You'd take the input file, reverse the order of the samples and then run it through the DAC, for comparison with the original file.

Is that practicably possible? Is it useful as a way of thinking about DACs?

Hmm - so a symmetric FIR filter (like the WTA filter) is time invariant, that is the impulse response is identical for when time is running forwards or backwards. But there are some filters, such as all the analogue ones and some digital ones (EQ and HF filters) which give different outputs depending upon time direction.

But I have huge problems with the idea of comparing a DAC analogue output with the original sampled data. The ideal DAC would be completely different to the sampled data, as the ideal DAC (and the intent of my WTA filters) is NOT to reproduce the digital data, but to reconstruct the original analogue signal before it was sampled. So by necessity, the perfect DAC would and should dramatically fail on this test. Another issue is bandwidth limiting; for the best sound quality, you need to very aggressively remove HF content from the source (particularly with HD recordings), and this results in much better sound quality; but of course, it would be a massive failure against the original data. And finally, we have the issue of the ADC creating it's own problems... So in my opinion, the comparison test is a waste of time, and would give the contradictory conclusion that the best performing with this test would be the worst sounding, assuming both DACs are technically capable.



Other news - my son Daniel is working with me full time for the next 6 months. So I have thrown him into the deep end, by giving him the pro ADC project to do. He has gotten to grips with the PCB software very quickly, and I am having difficulty keeping up with him on giving more work to do. This will speed things up dramatically, as this project is a big one - it won't be a simple 2 channel ADC, but will include full EQ and processing, mic pre-amps, and DAC outputs.

 

[Reactcore]

Genes are passed, the next Watts generation is a fact lucky us music lovers. Keep up the marvelous work!! But dont forget to enjoy the fruits

 

[Jawed]

Hmm - so a symmetric FIR filter (like the WTA filter) is time invariant, that is the impulse response is identical for when time is running forwards or backwards. But there are some filters, such as all the analogue ones and some digital ones (EQ and HF filters) which give different outputs depending upon time direction.

But I have huge problems with the idea of comparing a DAC analogue output with the original sampled data. The ideal DAC would be completely different to the sampled data, as the ideal DAC (and the intent of my WTA filters) is NOT to reproduce the digital data, but to reconstruct the original analogue signal before it was sampled. So by necessity, the perfect DAC would and should dramatically fail on this test. Another issue is bandwidth limiting; for the best sound quality, you need to very aggressively remove HF content from the source (particularly with HD recordings), and this results in much better sound quality; but of course, it would be a massive failure against the original data. And finally, we have the issue of the ADC creating it's own problems... So in my opinion, the comparison test is a waste of time, and would give the contradictory conclusion that the best performing with this test would be the worst sounding, assuming both DACs are technically capable.

I didn't mean comparison with the original samples, what I meant was if you put the original data through the DAC and compared the analogue waveforms of the original data and the time-reversed data (but reversed in the analogue domain). So my thought experiment relates to the analogue waveforms entirely.

Anyway, it seems you're saying that this experiment would show a difference because of the analogue filtering and some of the digital filters.

On the other hand it seems that in the digital domain this reversibility could be achieved. A key feature of WTA (or any linear FIR filter) is reversibility, so it's the non-reversible digital filters that create a difference..

I'm wondering whether reversibility is, ultimately, relevant to sound quality. If WTA retains reversibility, but the rest of the DAC (even if only digital domain) ignores reversibility, what is the sound quality penalty?

 

[miketlse]

Other news - my son Daniel is working with me full time for the next 6 months. So I have thrown him into the deep end, by giving him the pro ADC project to do. He has gotten to grips with the PCB software very quickly, and I am having difficulty keeping up with him on giving more work to do. This will speed things up dramatically, as this project is a big one - it won't be a simple 2 channel ADC, but will include full EQ and processing, mic pre-amps, and DAC outputs.

Does this mean he is a full time employee, or possibly alternatively a full time intern working on a real world challenging project (ie not the infamous photocopy 500 long documents), to support his studies?

 

[AnyDave]

Hmm - so a symmetric FIR filter (like the WTA filter) is time invariant, that is the impulse response is identical for when time is running forwards or backwards. But there are some filters, such as all the analogue ones and some digital ones (EQ and HF filters) which give different outputs depending upon time direction.

But I have huge problems with the idea of comparing a DAC analogue output with the original sampled data. The ideal DAC would be completely different to the sampled data, as the ideal DAC (and the intent of my WTA filters) is NOT to reproduce the digital data, but to reconstruct the original analogue signal before it was sampled. So by necessity, the perfect DAC would and should dramatically fail on this test. Another issue is bandwidth limiting; for the best sound quality, you need to very aggressively remove HF content from the source (particularly with HD recordings), and this results in much better sound quality; but of course, it would be a massive failure against the original data. And finally, we have the issue of the ADC creating it's own problems... So in my opinion, the comparison test is a waste of time, and would give the contradictory conclusion that the best performing with this test would be the worst sounding, assuming both DACs are technically capable.



Other news - my son Daniel is working with me full time for the next 6 months. So I have thrown him into the deep end, by giving him the pro ADC project to do. He has gotten to grips with the PCB software very quickly, and I am having difficulty keeping up with him on giving more work to do. This will speed things up dramatically, as this project is a big one - it won't be a simple 2 channel ADC, but will include full EQ and processing, mic pre-amps, and DAC outputs.

Watts^2

@Rob Watts , are you going to be at Axpona next month? It'd be nice to see you again. It's been a minute since I've worked a trade show.

 

[GuiltyRocker]

The so called "Ladder DACs," and R2R attempt to retain the digital data as possible and are not concerned with "reconstructing" the original analog waveform, to me they sound overly soft, boring, flat as a pancake, and unexciting, I think Rob is on the right track to focus on reconstructing the analog waveform as much as possible.

 

[Rob Watts]

I'm wondering whether reversibility is, ultimately, relevant to sound quality. If WTA retains reversibility, but the rest of the DAC (even if only digital domain) ignores reversibility, what is the sound quality penalty?

Just to re-cap there are broadly two forms of filters - causal filters (where the output depends solely on the current and past values) and symmetric acausal filters (where the output depends upon future and past values, and when the coefficients for samples offset into the past or the future are identical). Symmetric acausal filters have exactly the same output from an impulse whether the time is forward, or run in reverse or backwards in time. An example of a symmetric acausal filter is an symmetric FIR filter, like the WTA filter. Examples of causal filters are pretty much all analogue filters, and digital IIR filters (like the EQ in Mojo's UHD DSP), in short causal filters are real time filters.

FIR means finite impulse response - the output from an impulse stops once the impulse has past through all of the coefficients. An IIR filter is an infinite impulse response - as the filter feeds back to the input, it can create an output for an infinite amount of time, so long as the filter has infinite resolution.

The issue that's important here is that symmetric acausal filters pre-ring and post-ring, that is you get a response before the impulse. Causal filters on the other hand only create an output when an impulse occurs, or after an impulse has happened.

For a perfect interpolation reconstruction filters, it MUST be an infinite sinc function filter, and this is an ideal symmetric acausal filter. In reality, we need to reconstruct the timing of transients as accurately as possible, and that's what the WTA filter uniquely does, so that it reconstructs the timing of transients as accurately as possible given the limitations of real processing. But it will pre-ring; but that isn't a problem, as an appropriately bandwidth limited impulse will show minimal pre and post ringing with a WTA filter, and as the filter gets longer, the less ringing it will show. And if there was ringing, it's at an inaudible 22kHz (for 44.1k sample rate), so pre-ringing for an interpolation filter isn't an issue at all, in spite of what the rest of the audio industry says.

But if we are filtering in the audio domain, then pre-ringing is an issue, because pre-ringing artifacts would be in the audio bandwidth and hence audible, and it's clearly unnatural to have a response before the impulse. So it's very important to use causal filtering or IIR for EQ.

Does this mean he is a full time employee, or possibly alternatively a full time intern working on a real world challenging project (ie not the infamous photocopy 500 long documents), to support his studies?

Yes he is an employee, and my company has sponsored him during his Tonmeister degree. He is currently full time, and in September he will go back to Surrey, for his final year.

Watts^2

@Rob Watts , are you going to be at Axpona next month? It'd be nice to see you again. It's been a minute since I've worked a trade show.

Unfortunately not. I will be doing Singapore CanJam in early April though, with London CanJam in July. My next USA show is Socal CanJam, followed by Dallas CanJam.