Does DSP belong in State of the Art Systems?

[morricab]

For some reason many members here believe that DSP = Room Correction. I was speaking of DSP for implementing digital crossovers for instance or format conversions. Room Correction is a separate discussion on its own and should be addressed as “Digital Room Correction” and not as “DSP”.

I use digital crossovers in my home brew system. Works well but digitized analog doesn’t sound as good as analog direct when I use an analog active crossover.

 

[Carlos269]

I use digital crossovers in my home brew system. Works well but digitized analog doesn’t sound as good as analog direct when I use an analog active crossover.

Could be. It d pends which digital crossover you are using. I’m using an Electro-Voice EV-One, with the latest software, and getting great results.

 

[morricab]

Could be. It d pends which digital crossover you are using. I’m using an Electro-Voice EV-One, with the latest software, and getting great results.

I use a mini-DSP nano-digi, which only works in digital domain…so I need two outboard DACS. I have to digitize with a TASCAM digital recorder.

 

[Steve Rowan]

That's absolutely right. Few high end systems include "room correction DSP" because they are likely to be feeding similarly high end speakers. The better the system (provided it's set up properly in a sympathetically furnished room), the less the need for DSP.

Also as you say, it's a boon to DIY speaker builders who don't have the R&D and testing facilities for multiple prototypes to get an ideal response from their home-builds. Here DSP will be of enormous help. It also helps with poor / mediocre branded speakers where corners are cut, although it still won't match a well constructed and designed branded one of much the same size.

The other place where DSP may be helpful is with Home Theatre multi-speaker systems where the owner is unlikely to have the skills to set up his system for best performance. It's not easy, so resorting to DSP is almost forgivable, particularly because in HT system the audio is only part of the performance and our brains are less critical of less-than-perfect sound if they are also processing the video sensation.

However, DSP can be beneficial without spoiling the delicate signal if it's used in an active system where only the bass frequencies are sent to a dedicated bass-only amplifier that serves the bass drivers. My own speakers have this DSP facility though the software is not intuitive so I am not even using this - let alone Dirac Live.

In your later post referring to analogue, surely no one who is a true analogue fan would ever use DSP in their system as it involves analogue to digital conversion, then DSP, then digital to analogue conversion? Better perhaps to ditch the analogue source and start off with digital and thus cut out the ADC. The more a signal is processed, the more it will suffer damage.

Others (users of DSP) will disagree of course!

I agree to a point. I’m not a fan of Dirac at all. It seems artificial. At least the Dirac built into my Lexicon HT receiver. I believe it’s much to technical for most users. I prefer Audessy for what it’s worth.

But I know of two examples of great DSP. JL Audio provides DSP which is included in the Gotham and Fathom subwoofers. It’s very easy to use. It’s all over in about 10 chirps. One can hear the algorithm as it improves the bass.

The other is the Legacy Audio Wavelet DSP system. It is a preamp, DSP, crossover, and DAC.
It comes with its own microphone which is set up about 4 foot from each speaker. Three chirps for the right side then move the microphone and three chirps for the left side. The DSP is complete.

Yes I’ve been troubled with the conversion analog to digital and then digital to analog conversions.

However the conversions are so transparent I can switch every component before the Wavelet including footers, weights on components, moving the listening chair and instantly hear what I’ve accomplished or maybe unaccomplished.

Maybe its because ofmy inferior system that I’m hearing

 

[Steve Rowan]

Sorry I had to take a short break and couldn’t finish.

Maybe it’s because of my sorry midfi system that I can hear these differences and I need DSP.
I wouldn’t consider a Lampi Horizon, J Sikora turntable, CODA preamp and amps, midfi though. If I just had better components and speakers I wouldn’t need DSP?

But wait if I’m using DSP there can’t be any differences in components, tubes, interconnects, power cords, footers, VPI bricks etc, right? All should sound the same after analog to digital and digital to analog conversions.

Not.

 

[Mike Lavigne]

there is more than one way to approach optimizing a room. and more than one way to manage a signal path or crossover. and not everyone believes in (or wants to commit to) a dedicated space with serious uncompromised acoustic considerations. and we all have our sources we like and our own choice as to how we think they are most optimized.

for my 2 channel, i'm all the way into the purist all analog signal path, twin tower speakers with analog adjustments for the bass tower, passive crossover main tower, and a dedicated room with serious acoustical treatments and lots of acoustic room tuning efforts. and.....i'm happy with where i'm at. OTOH if dsp does it for you in any or all of these areas then congrats.

i have a dedicated home theater with a Trinnov Altitude 16 dsp processor. 9.3.6 Dolby Surround with three subs. the room is a nice size but it's compromised with double doors and patio doors and a bit of a bulge on one side where the double doors are. so the placement of some of the surround speakers is less than ideal. Trinnov has a set-up wizard that can 'locate' the speaker position anywhere. and a dsp engine second to none. even with minimal room treatments i get amazing Dolby Atmos object based surround performance supporting my movie soundtracks. concert videos can be fun, and involving, but it's not going to compete on two channel sources. the Trinnov is all dsp all the time.

so i am not anti dsp. and i could see using dsp for 2 channel in many situations. but it's not my first choice. YMMV.

 

[Carlos269]

Sorry I had to take a short break and couldn’t finish.

Maybe it’s because of my sorry midfi system that I can hear these differences and I need DSP.
I wouldn’t consider a Lampi Horizon, J Sikora turntable, CODA preamp and amps, midfi though. If I just had better components and speakers I wouldn’t need DSP?

But wait if I’m using DSP there can’t be any differences in components, tubes, interconnects, power cords, footers, VPI bricks etc, right? All should sound the same after analog to digital and digital to analog conversions.

Not.

Ideally, DSP should be neutral and allow you to hear the differences between components. The opposite of what you have stated as a “Not”. If you are referring to Digital Room Correction then the concept is that by correcting the resultant sound at the calibrated microphone the differences between components used is minimized. You sound like you are really proud of your equipment. Maybe you could give us a taste of how it sounds.

 

[Bjorn]

As we have clearly seen in this thread, it's impossible to discuss DSP when people continue to misunderstand what it is and primarily think it equals "room correction". Sad but true.

 

[Tim Link]

Indeed, room correction is just one aspect of DSP, and it's one that I'm having less confidence in all the time, after experimenting with it quite a bit.
A study was done on how much DSP could improve room clarity, and the answer was basically that it can't, at least not in a 2 channel system. A good room doesn't get any better with DSP, and a bad room usually doesn't either, unless you are sitting close enough to the speakers, where room problems tend to be simpler and lower in magnitude. My experience has been the same. Room correction can't do much about reverberation.

https://salford-repository.worktribe.com/preview/1501574/start.pdf

One thing this same study showed is that a good speaker does help! It makes sense that the room can't be any clearer than the speaker, so a bad speaker starts you off bad, and the room can only make it worse from there. And this is where DSP is interesting because using DSP to create crossover networks can allow excellent frequency and phase response with minimal comb filtering from driver overlap that is simply impossible without it. And, my experience shows that when I get my three way speakers dialed in with good crossover settings that give me a nearly perfect step response, the room clarity at the listening position measures significantly higher. It's true that you can get a good minimum phase step response with passive filters, but it requires 1st order crossovers which bring a lot of headaches and sonic issues of their own. With DSP you can get excellent minimum phase or even linear phase response with minimum overlap crossovers. I'm not sure it matters taking it all the way to linear phase, but I can do it with the flip of a switch, comparing uncorrected phase to corrected. It's a subtle difference, but after a lot of listening I tend to leave the phase correction on.

I know a lot of people feel that digital does something to the sound that's bad, and I'm not going to doubt them. But it also does things to the sound that are very, very good if used right, and those things can't be done analog. So the issue comes down to what matters to you.

As for DIY, I would never consider spending a lot of money on anything other than DIY. If you really think about it, there are only levels of DIY. It's not clearly defined. Very few people make their own electronic components or speaker drivers. The furthest you could get from DIY would be to have someone else pick everything for you, including the listening space and all furnishings and treatments.

 

[stehno]

Regardless of the situation, all electronic signal processing potentially induces distortions. If straight-wire is the ideal or most unadulterated, then it seems to me that DSP always has the potential to stray away from that ideal.

Even if DSP is built-in on some other component like a pre-amp or pre-pro, there’s additional input signal processing. Such compromises are even more true if the DSP is its own component because then we’re talking an additional power supply and additional input/output cabling. All of which will contribute to very real potential input signal sonic compromises - to one degree or another.

Besides, DSP is dealing with the effects rather than the cause. If, without DSP, the input signal is already compromised (it is), then muddying or compromising the input signal further ought to be the last thing a performance-oriented type would want to do. Then again, I suppose for those who only care about personal preferences (not performance) then everything is fair game???

For better and worse, everything one should ever want or hope to hear is already embedded in the recording – which is a snapshot of the live performance in a moment in time and never to be duplicated again. For better or worse, the recording includes whatever music info it could capture from the live performance as well as all recording engineering. We can certainly doctor what music info is read in but we simply cannot improve on it downstream or after-the-fact. At least not while maintaining the same original fidelity.

I’d venture the best one could ever hope for with DSP is to create a phenomena of a sound that…

1. May seem more pleasing / less offensive (not to be confused with more naturally musical) but strays from the fidelity of the music info embedded in the recording.

2. Masks or band-aids sonic shortcomings elsewhere in the system. Which implies one has no intention to ever address root causes – especially if somewhat satisfied with their DSP tweaks?

If DSP is invoked, then there’s always the potential hassle of being compelled to tweak the DSP to certain individual recordings – since every recording is engineered differently. Which ought to be silly.

IMO, DSP is just one more example of dealing with the effects rather than the cause(s) – a strategy which most always if not always amounts to little more than chasing windmills.

 

[Tim Link]

IMO, DSP is just one more example of dealing with the effects rather than the cause(s) – a strategy which most always if not always amounts to little more than chasing windmills.

Some issues (effects) can actually be corrected with added complexity. Consider LP playback. You have to run it through the transform function provided by a phono stage to get it to sound right. As far as I know these all require a power supply to work well. And then there are crossover components in the speakers. It's true that a single driver can sound pretty good, and it's ideal to get rid of the passive network, but there's a huge price to pay for that too. A passive network is simpler than a digital one in terms of total parts and lack of power supply, but it doesn't work perfectly by any means, and a digital one can work better in certain ways.

Consider camera lenses. A peephole lens is very simple. It produces interesting effects. But if you add a lens you can get a lot of improvement in image clarity. Add multiple element lenses and coatings on those lenses and the whole thing can produce a purer image despite all the added transfer functions.

There are times when adding noise can actually improve the clarity and naturalness of an image or sound. Most of high end audio seems to be about intentionally adding some kind of transfer function to prevent the odd perceptual effects of straight wire. There's an oddness inherent to almost any kind of attempt to reproduce reality, and compensatory transfer functions are often sought after.

Going back to phono stages and records, a gramophone is fun to play with. The needle in the groove just vibrates a diaphragm and the sound comes straight out the horn. No electricity. Better than straight wire. Surprisingly loud. Not much gramophone action these days in the high end.

 

[Hear Here]

For some reason many members here believe that DSP = Room Correction. I was speaking of DSP for implementing digital crossovers for instance or format conversions. Room Correction is a separate discussion on its own and should be addressed as “Digital Room Correction” and not as “DSP”.

There are lots of reasons for Digital Signal Processing, including "room correction" Of course this is a misnomer as no signal processing will ever correct the room!

The point is that the OP, who initiated this thread said "I'm hoping/expecting better integration between my subs and mains". Thus, assuming that his subs were being adversely affected by his room's less than perfect acoustics, he was looking for what is often (though incorrectly) described as "room correction DSP". But correct me if I'm wrong in my interpretation.

Since so many people seem to think that messing with the signal in an attempt to improve the sound from a poorly selected or poorly set-up speaker system, is in my view misguided. To be fair though, this subject is far more prevalent on other forums than this one, as most contributors here have much better systems and the determination to get them set up for best performance. Not so elsewhere, but the question "Does DSP belong in State of the Art Systems?"

 

[Steve Rowan]

The NOT I guess confused some. I meant that I can clearly hear the differences in any equipment I can put in the chain before Wavelet.

I tune my room as best it can be without DSP turned on. I find the best place for the speakers and subwoofers in the room. Then add any room treatments as necessary. I listen for best placement for days or weeks. After that comes the final touch: the DSP. One can easily play the system without DSP with the press of a button on the Wavelet’s app. U

I am somewhat puzzled by the assertion that “expensive high end loud speakers” never need DSP or shouldn’t be contaminated by DSP.

I don’t claim to have heard every version of DSP that’s available. But I’ve heard and played around with several different systems. The complexity of many outweigh the benefits and leave you always wondering “did I do it right?”That’s not the case of the Legacy Audio Wavelet.

I cut my audiophile teeth at 16 with Dahlquist DQ 10s. By 19 I owned Acoustat 2 + 2s then Maggie MG 3s by 22 or so. Yep and somewhere around 20 I started working at a St. Louis High End audio shop where I got to play with Snells, Acoustats, all the Maggie’s of that era, Spendors, Mirage, and many others. Speaker placement and room acoustics play a major role in getting the best out of your system.

Of course I know that many record their systems and post the results here and other sites. To me trying to judge a system’s merits on an iPhone doesn’t make much sense. But if I have time I’ll record it someday.

Of course if you ever visit the St. Louis area I’d be glad to demo my system.

Several folks on this site have heard my system. Al Rainbow, Lukasz Fikus, Mr Peabody and a few others have visited. Mr. Peabody visits at least once a month.

 

[Tim Link]

I agree with Kal's response to the OP's question about integrating a subwoofer. My experience has been that extra A/D D/A conversion steps are not as good as doing all the processing in the digital domain, and then doing the D/A conversion just once. That said, I don't know that the extra A/D D/A steps are necessarily worse than a pure analog crossover. It still might perform better overall. It's just not fully optimizing the potential.

 

[sbo6]

That's absolutely right. Few high end systems include "room correction DSP" because they are likely to be feeding similarly high end speakers. The better the system (provided it's set up properly in a sympathetically furnished room), the less the need for DSP.

Also as you say, it's a boon to DIY speaker builders who don't have the R&D and testing facilities for multiple prototypes to get an ideal response from their home-builds. Here DSP will be of enormous help. It also helps with poor / mediocre branded speakers where corners are cut, although it still won't match a well constructed and designed branded one of much the same size.

The other place where DSP may be helpful is with Home Theatre multi-speaker systems where the owner is unlikely to have the skills to set up his system for best performance. It's not easy, so resorting to DSP is almost forgivable, particularly because in HT system the audio is only part of the performance and our brains are less critical of less-than-perfect sound if they are also processing the video sensation.

However, DSP can be beneficial without spoiling the delicate signal if it's used in an active system where only the bass frequencies are sent to a dedicated bass-only amplifier that serves the bass drivers. My own speakers have this DSP facility though the software is not intuitive so I am not even using this - let alone Dirac Live.

In your later post referring to analogue, surely no one who is a true analogue fan would ever use DSP in their system as it involves analogue to digital conversion, then DSP, then digital to analogue conversion? Better perhaps to ditch the analogue source and start off with digital and thus cut out the ADC. The more a signal is processed, the more it will suffer damage.

Others (users of DSP) will disagree of course!

IME the more revealing your system the bigger the degradation using DSP. It's a tradeoff, and sometimes it's worth it more so (again) for less revealing systems. I've used it in my office system, but in my main system, nope.

Also, there's this. https://www.stereophile.com/roomtreatments/883/index.html

 

[gleeds]

Those adverse to using DSP for bass room mode interaction have well-known options that can be used with varying degrees of effectiveness. That said, for most rooms/systems, this will require lots of painstaking work to achieve the same or similar results that can be had with a well-executed electronics room correction/calibration suite.

Full disclosure:
I work with AEquo Audio, an innovative high-end loudspeaker company from the Netherlands. AEquo has addressed the bass/room mode room interaction problem by developing an integrated and purely analog solution named ARPEC.

Later this year, AEquo will add ARPEC to their newest innovation, the Ensium loudspeaker, using a proprietary metal matrix composite cabinet material using advanced casting techniques to form a thin-wall cabinet whose resonant properties far surpass anything on the market regardless of weight and thickness utilized. When combined with ARPEC, we expect the Ensium to deliver bass performance exceeding anything on the market in its size and price class while offering the benefits of DSP without the drawbacks. Best of all, there is no need for additional boxes or elaborate measuring equipment other than the most straightforward measuring device available: our ears.

Below is a white paper describing the technology and its use.

Aequo Audio Technology

State-of-the-art technologies developed by Aequo Audio.

www.aequoaudio.com

Please scroll down the page to read about ARPEC.

 

[Tim Link]

IME the more revealing your system the bigger the degradation using DSP. It's a tradeoff, and sometimes it's worth it more so (again) for less revealing systems. I've used it in my office system, but in my main system, nope.

Also, there's this. https://www.stereophile.com/roomtreatments/883/index.html

Revealing is a tricky word. What exactly it means is not clear, but it doesn't exactly correlate with high fidelity. A passive network in a speaker will invariably produce audible and measurable signal degradation. But what does it do to transparency? I read a test years ago where a university was trying to split a signal into high pass and low pass, and then combine the two signals again so that the splitting and recombining process was not audible. Cost no object, they couldn't do it in the analog domain. DSP passed the test. Nobody could tell the difference, and I believe in that test they were actually inserting an extra A/D D/A step into the process. And it was just a Behringer pro audio DSP box. Now maybe if the listeners had been better trained they could have picked up on the digital conversion degradation. But still, they had no problem at all hearing the analog processing, with 100% accurate detection from just about anyone who gave it a try.
And yet, most high-end home audio speakers still have passive networks. And, most of the high end market seems to take them more seriously. My take is that it's an artistic target sound, similar to how there are still those who prefer film to digital photography. My dad used to say in the early days of digital photography "nobody knows how to make it look good yet." It's about more than accuracy. It's about artistic use of inaccuracy to create an effect in the mind of the seer or hearer. Analog creates its own effects, and these come to symbolize realism in our minds over time. I play video games that add a lens flare to the scene when the sun is visible. It's silly, but somehow that aberration adds perceptual realism.
And then there's higher frame rates in movies. Somehow they destroy something. They're too revealing, and yet the effect is that it destroys the feel that was meant to be conveyed. So they are not revealing of the mood as intended, but more revealing of the actual nature of what is on camera - a bunch of people on a set wearing makeup and reciting lines.

 

[Holmz]

Revealing is a tricky word. What exactly it means is not clear, but it doesn't exactly correlate with high fidelity…

But definition of the word fidelity, revealing correlated 99+% with fidelity.
It just fidelity doesn’t correlate well with what people find that they like, and it is not overly causal to “liked” systems.

Low speaker compression is probably more causal/correlated to “liking”, and perhaps the hard clipping of digital is sneaking in there?
Or the softer clipping of analogue.

 

[Tim Link]

But definition of the word fidelity, revealing correlated 99+% with fidelity.
It just fidelity doesn’t correlate well with what people find that they like, and it is not overly causal to “liked” systems.

Low speaker compression is probably more causal/correlated to “liking”, and perhaps the hard clipping of digital is sneaking in there?
Or the softer clipping of analogue.

That's interesting. What source are you referring to about the 99+% correlation of "revealing" with "fidelity?" I don't tend to think on those lines, as so many times a filter can be used to highlight things that couldn't normally be seen easily. NASA images come to mind. Often the high fidelity images are very bland and lacking in perceptual detail.

I would think low speaker compression would be high fidelity feature, and may be more revealing too. Although, compression can sometimes reveal low level details that would normally be masked. High fidelity would reveal only what would naturally be revealed to the unaided senses.

I'm glad you mentioned hard clipping in digital. I was going to bring that up because it's true, digital can clip in a nasty way if something isn't done to cause it to clip softly like analog. Soft clipping in analog is not automatic either. A lot of work was put into giving film a shoulder. My dad used to tell me that the early point and shoot camera technology was mostly in the forgiving nature of the films they used.

In digital photography, clipping is pretty much unavoidable in many outdoor scenes, so it's critical that a soft fade into the highlights occur. Easily done using curves. With audio, clipping can certainly happen during a live event, but I'm not sure that a professionally produced recording must necessarily have any clipping. I don't ever let any of my DSP processes get in to clipping. But then there's also the potential of loss of resolution in the quiet areas. Besides the noise floor becoming more apparent, which is also the case in analog, I'm not sure I've ever heard any of the effects of low bit depth in a quiet part of a recording that's been turned up really loud. I can see that in digital images, the blocking you see if you try to lift the exposure in the shadows. I'm not really sure what to even listen for to hear the equivalent audio effect.

 

[Holmz]

That's interesting. What source are you referring to about the 99+% correlation of "revealing" with "fidelity?" I don't tend to think on those lines, as so many times a filter can be used to highlight things that couldn't normally be seen easily. NASA images come to mind. Often the high fidelity images are very bland and lacking in perceptual detail.

I would think low speaker compression would be high fidelity feature, and may be more revealing too. Although, compression can sometimes reveal low level details that would normally be masked. High fidelity would reveal only what would naturally be revealed to the unaided senses.

I'm glad you mentioned hard clipping in digital. I was going to bring that up because it's true, digital can clip in a nasty way if something isn't done to cause it to clip softly like analog. Soft clipping in analog is not automatic either. A lot of work was put into giving film a shoulder. My dad used to tell me that the early point and shoot camera technology was mostly in the forgiving nature of the films they used.

In digital photography, clipping is pretty much unavoidable in many outdoor scenes, so it's critical that a soft fade into the highlights occur. Easily done using curves. With audio, clipping can certainly happen during a live event, but I'm not sure that a professionally produced recording must necessarily have any clipping. I don't ever let any of my DSP processes get in to clipping. But then there's also the potential of loss of resolution in the quiet areas. Besides the noise floor becoming more apparent, which is also the case in analog, I'm not sure I've ever heard any of the effects of low bit depth in a quiet part of a recording that's been turned up really loud. I can see that in digital images, the blocking you see if you try to lift the exposure in the shadows. I'm not really sure what to even listen for to hear the equivalent audio effect.

^That^ is a great analogy with images.

Maybe I am conflating resolution with fidelity.
You mention of NASA image processing seems like maximum entropy methods to bring out detail amongst the noise.