Does DSP belong in State of the Art Systems?

[the sound of Tao]

Precisely!
You can fall of the cliff much quicker using dsp
It takes many years to reach the bottom using a coventional approach
Phil

All of the worst systems I’ve heard have either been digital or analogue.

The best I’ve heard have been in the analogue spectrum but great digital is still possible if it has been implemented in a kind of digital friendly or digital ideal way. Analogue and digital l see as having very different implementation pathways if you want the best out of either.

Still yet to hear a dsp system (or class d amp) at the best end (or for me even an ok end but that’s just for my needs) but that doesn’t mean it can’t happen.

Everyone’s experiences, expectations and perceptions are clearly variable in this area. There have been a few guys making positive noise about class d lately. That’d be good if it is better now but the only digital I’ve really liked has involved not using too many digital layers or components or at the more intense end of digital processing.

In my experience much fragmentation and multi layering with very heavy digital processes can tend with it then ending up seeming just more like something of a digital soup.

I think some of this discussion is unfair to the pro-DSP camp as it is an as yet (I believe) undisclosed fact on this thread that digital anything -- including DSP -- simply is anathema to a subset of the WBF membership.

 

[Bjorn]

All good questions. How can an acoustician measure the difference in transparency I hear from a vocal recording reproduced by electrostatic loudspeakers and the same vocal recording reproduced by dynamic driver loudspeakers?

Transparency is an audible difference that I suspect is not cognizable in the frequency domain.

By measuring the drivers, the speakers and considering how they integrate with the room. Transparency of speakers is mostly related to distortion and the directivity, not so much the frequency response.

 

[pjwd]

I

All of the worst systems I’ve heard have either been digital or analogue.

The best I’ve heard have been in the analogue spectrum but great digital is still possible if it has been implemented in a kind of digital friendly or digital ideal way. Analogue and digital l see as having very different implementation pathways if you want the best out of either.

Still yet to hear a dsp system (or class d amp) at the best end (or for me even an ok end but that’s just for my needs) but that doesn’t mean it can’t happen.

Everyone’s experiences, expectations and perceptions are clearly variable in this area. There have been a few guys making positive noise about class d lately. That’d be good if it is better now but the only digital I’ve really liked has involved not using too many digital layers or components or at the more intense end of digital processing.

In my experience much fragmentation and multi layering with very heavy digital processes can tend with it then ending up seeming just more like something of a digital soup.

Digital filters do no more or less than passive filters to the signal excepting they have none of the phase issues etc. Not talking about any a/d d/a coversions here
Its all about implementation .. if you are interested in a trip up to byron in 3 - 4 months I should have a system up and running that may (or may not ) give some idea of what is possible. Beats talking about it

Cheers
Phil

 

[the sound of Tao]

I
Digital filters do no more or less than passive filters to the signal excepting they have none of the phase issues etc. Not talking about any a/d d/a coversions here
Its all about implementation .. if you are interested in a trip up to byron in 3 - 4 months I should have a system up and running that may (or may not ) give some idea of what is possible. Beats talking about it

Cheers
Phil

Agree that implementation is everything but also I find everything different does something different and mostly this is why things sound different… but that’s just my experience in this.

 

[tima]

In my experience much fragmentation and multi layering with very heavy digital processes can tend with it then ending up seeming just more like something of a digital soup.

 

[Holmz]

I
Digital filters do no more or less than passive filters to the signal excepting they have none of the phase issues etc. Not talking about any a/d d/a coversions here
Its all about implementation .. if you are interested in a trip up to byron in 3 - 4 months I should have a system up and running that may (or may not ) give some idea of what is possible. Beats talking about it

Cheers
Phil

Byron is a long ways from WA - but not that long of a way.
Interested in what you’re doing…

I got an Octo DAC8 Pro, and have been intending on FIR based active XOs.

 

[pjwd]

Byron is a long ways from WA - but not that long of a way.
Interested in what you’re doing…

I got an Octo DAC8 Pro, and have been intending on FIR based active XOs.

Cool .. I have the merging happi .. there are a few decent options these days .. I am doing a dipole system .. will post something when It gets assembled
Phil

 

[morricab]

Transparency is a subjective perception and not necessarily a property of the signal.

It is of course a property of the signal...to suggest otherwise is silly. Whether you can perceive subtle differences in the signal that contribute to the perception of transparency is another issue.

 

[morricab]

I
Digital filters do no more or less than passive filters to the signal excepting they have none of the phase issues etc. Not talking about any a/d d/a coversions here
Its all about implementation .. if you are interested in a trip up to byron in 3 - 4 months I should have a system up and running that may (or may not ) give some idea of what is possible. Beats talking about it

Cheers
Phil

They do DIFFERENT than passive filters. If you don't have jitter and other digital artifacts under control, and for sure all but the most dedicated designs do not, then you will hear the effect of a digital filters, especially if from a hardware implementation and also if it has its own DA conversion and analog stages. These are often cookbook recipe circuits as the guys are mostly digital engineers. This means these will not sound nearly as good as properly implemented DA solutions. If it is software, well, it still has go into some kind of DA system to get to analog and a computer card will NOT cut it sonically.

 

[Bjorn]

They do DIFFERENT than passive filters. If you don't have jitter and other digital artifacts under control, and for sure all but the most dedicated designs do not, then you will hear the effect of a digital filters, especially if from a hardware implementation and also if it has its own DA conversion and analog stages. These are often cookbook recipe circuits as the guys are mostly digital engineers. This means these will not sound nearly as good as properly implemented DA solutions. If it is software, well, it still has go into some kind of DA system to get to analog and a computer card will NOT cut it sonically.

No need to us extra DA conversion when a DSP has volume control. And the quality of the volume control and DA conversion in the best DSPs can rival the best DACs.

Using software and a quality DAC with computer can work as good as anything else. Noise and distortion can be measured.

 

[Kingrex]

What does State Of The Art mean? Has this been fleshed out?

I have heard a decent handfull of fine stereo. At a level, I really don't know its the equipment that sets them apart. A lot of what makes a system great that I hear is the room.

And I do gravitate to the sound of larger speakers. Moving air seems to matter. Scale helps.

If you have a great room, I think you can go any route you want with electronics. I don't see a right or wrong. I think any path takes some time and attention to detail. It doesn't just happen. If you work with or without DSP, your going to create a sound that hopefully is what you are looking for.

In saying that, DSP might give someone a little advantage in getting to a sound they like with the equipment they have as they can probably affect more variations in tone from what they are using. If you don't have DSP, your speakers are what they are.

 

[Ron Resnick]

Transparency of speakers is mostly related to distortion and the directivity, not so much the frequency response.

Do you mean whichever topology of loudspeaker driver surface distorts less? For example, electrostatic membrane distorting less than dynamic driver?

 

[Ron Resnick]

In saying that, DSP might give someone a little advantage in getting to a sound they like with the equipment they have as they can probably affect more variations in tone from what they are using. If you don't have DSP, your speakers are what they are.

If we are talking about DSP merely as a system tone control, why not just use an inexpensive pro-audio equalizer — and not bother diving into the digital domain?

(I realize that a DSP system can “hear” the room and solve the frequency response equation and compensate automatically at every frequency — effectively a self-calculating, automatic equalizer.)

 

[schlager]

What does State Of The Art mean? Has this been fleshed out?

I have heard a decent handfull of fine stereo. At a level, I really don't know its the equipment that sets them apart. A lot of what makes a system great that I hear is the room.

And I do gravitate to the sound of larger speakers. Moving air seems to matter. Scale helps.

If you have a great room, I think you can go any route you want with electronics. I don't see a right or wrong. I think any path takes some time and attention to detail. It doesn't just happen. If you work with or without DSP, your going to create a sound that hopefully is what you are looking for.

In saying that, DSP might give someone a little advantage in getting to a sound they like with the equipment they have as they can probably affect more variations in tone from what they are using. If you don't have DSP, your speakers are what they are.

That is what I would call the "modern" approach and most I know within audio is true to that. Even some "old timers" analog guys has swayed over to active DSP speakers and they have never looked back.

In 20 years time everything will be based on DSP and cable talk and distortion profiles are long gone discussions.

 

[tima]

If you don't have DSP, your speakers are what they are.

Your speakers are what they are, dsp or not.

You do have control over speaker position. Some speakers allow you to alter time arrival from different drivers. Some let you fiddle with the cross-over. Field-coil speakers allow alteration of tsp characteristics.

Some box listens to music in your room and adjusts how it sounds. What is the basis of preference of that box?

 

[microstrip]

I think this is tautological and, therefore, not terribly helpful. The only reason you would consider it "proved to be heard" is because it can be measured.

Well, fortunately people who work professionally in perceptual areas disagree with you. There are accepted methods to consider something perceptual as proved.

 

[Rt66indierock]

I can’t imagine building a system today without Audiolense XO or something like it.

 

[Atmasphere]

Transparency is a subjective perception and not necessarily a property of the signal.

This statement is easily demonstrated to be false unless there is a difficulty distinguishing between the perceiver and the perceived. Sure- all of our perception occurs in the brain but there are perceptual rules that are used by all individuals. On the other hand, take away the signal (regardless of where in the musical chain it occurs) and nothing will be perceived. So it is easy to see that the signal can vary in how much of the property of 'transparent' it manifests.

Distortion obscures detail. Harmonics and intermodulations get added that change the tone of instruments. The more distortion you can remove the greater the transparency; the most transparent a system can be will also be the system with the lowest distortion. One problem that is poorly understood is how the ear/brain system uses higher ordered harmonics to sense sound pressure; this means that some 'transparent' systems might also sound bright and harsh, since the ear assigns that tonality to higher ordered harmonics, just as those harmonics sculpt the sound of any musical instrument.

 

[pjwd]

This statement is easily demonstrated to be false unless there is a difficulty distinguishing between the perceiver and the perceived. Sure- all of our perception occurs in the brain but there are perceptual rules that are used by all individuals. On the other hand, take away the signal (regardless of where in the musical chain it occurs) and nothing will be perceived. So it is easy to see that the signal can vary in how much of the property of 'transparent' it manifests.

Distortion obscures detail. Harmonics and intermodulations get added that change the tone of instruments. The more distortion you can remove the greater the transparency; the most transparent a system can be will also be the system with the lowest distortion. One problem that is poorly understood is how the ear/brain system uses higher ordered harmonics to sense sound pressure; this means that some 'transparent' systems might also sound bright and harsh, since the ear assigns that tonality to higher ordered harmonics, just as those harmonics sculpt the sound of any musical instrument.

I would agree transparency is on the signal and is revealed by increasing levels of precision in playback
Would this percieved harshness be one of the reasons for a slightly rolled off top end that is present in all the target curves. Linkwitz had a quite an interesting theory on the reason as I recall
Of course the biggest distorters are speakers and room
Phil

 

[Kal Rubinson]

But then why doesn’t the DSP camp acknowledge that perceptions that can be heard may not be able to be measured?

Because the perceptions of a single individual are subjective and unique and, therefore, are not objectively significant. Can you offer an example of a audible perception that has been statistically authenticated across a large sample and yet is not measurable? It is logically possible but I do not know of one.

It is of course a property of the signal...to suggest otherwise is silly.

Nope. It is a subjective observation that does not exist without human perception. An analogous term for the signal would be "accurate."