Does DSP belong in State of the Art Systems?

[Ron Resnick]

@PGA; I think you’re on the wrong forum . Don’t you know that analogue is good and digital is bad, because, because, well it just is . . .

Dear iansr,

FYI we have loads of digital aficionados here, many of whom are digital only. We also have SOTA analog-focused people who have chosen to pursue digital to the highest SOTA level.

 

[treitz3]

There is no wrong forum to discuss audio. So long as we are on topic with DSP?

Please continue...

Tom

 

[sparkie]

I really don't see DSP technology being useful beyond being a crossover. Its not like your living room is a stadium and you need a DSP to time align speaker clusters together.

Room treatment technology and methods applying them people should really focus on.

 

[PGA]

I really don't see DSP technology being useful beyond being a crossover. Its not like your living room is a stadium and you need a DSP to time align speaker clusters together.

Room treatment technology and methods applying them people should really focus on.

Respectfully disagree. At least in my case, I was only able to really tune in my system with DSP. And it’s not just about the bass. Getting the midrange and treble just right is equally important. The TAD R1s are great speakers and are nearly ideal transducers, but still DSP made a huge difference across the entire audible bandwidth.

My first attempt at room at room treatment was with ASC tube traps, including their huge bass traps. This did almost nothing for me and the room looked ridiculous. A Cello Palette EQ was very helpful with wide bandwidth adjustments, but it was useless for dealing with bass problems. I would say similar things about the Dangerous Music BAX. RPG is great reducing first reflection on the walls and ceiling, but my room needed more than that. And designing DSP filters with REW either created more problems than it solved or made little difference if applied in a more limited way.

Then I called Mitch Barnett and everything came into place. Audiolense is simply in a totally different planet vis a vis what you can do with REW. But it is a professional tool that requires skill to get the most out of it.

What is most surprising is that Mitch only requires measurements at one point in your main listening position. Yet the system sounds great pretty much everywhere in the room. The number crunching involved would have been science fiction 20 years ago, but now a $400 Intel NUC can run the convolution without breaking a sweat.

 

[sparkie]

Respectfully disagree. At least in my case, I was only able to really tune in my system with DSP. And it’s not just about the bass. Getting the midrange and treble just right is equally important. The TAD R1s are great speakers and are nearly ideal transducers, but still DSP made a huge difference across the entire audible bandwidth.

My first attempt at room at room treatment was with ASC tube traps, including their huge bass traps. This did almost nothing for me and the room looked ridiculous. A Cello Palette EQ was very helpful with wide bandwidth adjustments, but it was useless for dealing with bass problems. I would say similar things about the Dangerous Music BAX. RPG is great reducing first reflection on the walls and ceiling, but my room needed more than that. And designing DSP filters with REW either created more problems than it solved or made little difference if applied in a more limited way.

Then I called Mitch Barnett and everything came into place. Audiolense is simply in a totally different planet vis a vis what you can do with REW. But it is a professional tool that requires skill to get the most out of it.

What is most surprising is that Mitch only requires measurements at one point in your main listening position. Yet the system sounds great pretty much everywhere in the room. The number crunching involved would have been science fiction 20 years ago, but now a $400 Intel NUC can run the convolution without breaking a sweat.

Like I said nothing much more than a crossover DSP.
An example of some are the DBX DriveRack series. You wouldn't need the extended functions like feedback suppression. But they have been making these DSP processors for over 18 years.

DriveRack | dbx Professional Audio | English (US)

The DriveRack® series are a diverse group of processors that provide control of equalization and loudspeaker management. Used and appreciated in every kind of venue, from houses of worship to big-name world tours, the DriveRack series are the most popular loudspeaker management systems.

dbxpro.com

 

[sparkie]

My first attempt at room at room treatment was with ASC tube traps, including their huge bass traps. This did almost nothing for me and the room looked ridiculous. A Cello Palette EQ was very helpful with wide bandwidth adjustments, but it was useless for dealing with bass problems. I would say similar things about the Dangerous Music BAX. RPG is great reducing first reflection on the walls and ceiling, but my room needed more than that. And designing DSP filters with REW either created more problems than it solved or made little difference if applied in a more limited way.

Then I called Mitch Barnett and everything came into place. Audiolense is simply in a totally different planet vis a vis what you can do with REW. But it is a professional tool that requires skill to get the most out of it.

Treating rooms is important. But there are some rooms that are harder to treat than others. The more square and more smaller the room is, the harder it is to treat.

Btw, the best bass traps are not commercially made and if the room is very small, the trap ends up in a weird place because of the air space demand behind them needed to work properly. There are a lot of things involved in designing treatment for a room so its almost difficult to make a product that is going to work in every space. That is why you had problems.

 

[PGA]

Treating rooms is important. But there are some rooms that are harder to treat than others. The more square and more smaller the room is, the harder it is to treat.

Btw, the best bass traps are not commercially made and if the room is very small, the trap ends up in a weird place because of the air space demand behind them needed to work properly. There are a lot of things involved in designing treatment for a room so its almost difficult to make a product that is going to work in every space. That is why you had problems.

Yes. You would need very large bass traps to fix the bass problems I have in my smallish room (12x18x8.5). It’s just not a practical solution. What ended up working extremely well was placing the TADs essentially in the corners along the 12 foot wall and eliminating the bass boost with DSP. RPG treatment absorbed much of the reflections above 150-200hz. Imaging, stage width and depth was surprisingly good. Also the bass boost resulting from the corner placement allowed for 10+ db of extra headroom in the deep bass and a -3 db point of 17hz.

I’m not familiar with the DBX crossover you referenced above, but a quick read of the link seems to indicate it’s primarily for live sound in large venues. They don’t seem to have FIR capability. The Audiolense convolution filters I’m using have latency and more than 50,000 taps, and would not be applicable to live music. Mitch also provided low latency filters that I used for home theater, but these are not practical to use in a multichannel setup. So I ended up using DIRAC for TV.

 

[PGA]

 

[PGA]

View attachment 127925

This is the before and after response of the TADs in my room with the speaker close to the corners. Note these R1s are actively biamped using a Pass XVR1 set up by Andrew Jones. The LPF in the TADs for the woofers is bipassed. The internal crossover for the coax is still in place. The woofers and the coax each have a Benchmark amp driving them in mono.

 

[schlager]

After fiddling with my system for 40+ years I must agree with Analog Scott and Kal.

I tried everything, including having my room professionally treated with RPG abfusors, diffusors and bass panels. Also went through many speakers including Wilson, Cello, Dunlavy, TAD and Magico. And many DACs and amps from Spectral, Cello, Levinson, MSB, Berkeley and Pass. For the most part the improvements (or better said, differences) were subtle.

At the behest of folks at Berkeley, I tried implementing EQ in the analogue domain, using high quality pro gear. That helped, but nothing really got me to where I wanted until I had Mitch Barnett design a convolution filter that in my system runs on Roon.

Now with the convolution filter the system sounds truly fantastic, even after I sold all the expensive cables, power conditioners and ultra expensive Constellation amps and Berkeley DAC. I have a more humble Meitner MA3 and a Pass XVR1 feeding four Benchmark amps, all with Benchmark wiring powering modified TAD R1s. And I've never been happier.

DSP will not eliminate distortion, deep nulls, flutter echo and reflections, so paying attention to the room is important. You should have a system capable of playing reasonably loud without compression or distortion across the frequency spectrum. And also unless you want to spend many hours trying to learn a fraction of what Mitch Barnett forgot about DSP, you're probably better off staying away from DSP. But given the processing power that's available today, DSP done right can be truly transformative.

For me the acid test was when I invited over a dear old friend that is a big analogue buff, with vintage tubes and horns. I expected he was going to literally throw up listening to my 100% digital system. But instead he was totally amazed. Paraphrasing his comments after listening for a couple hours he concluded that maybe there was a thing or two he liked better in analogue, but there were many many things my system did much better.

Again you have to start with a low distortion system that covers the frequency range, you need to get the room as good as you can (but not to the point of going nuts with room treatments) and you need a professionally designed DSP filter made for your room/speakers.

Having said this, I use a Denon receiver as a pre-pro running DIRAC for home theater in this room. It triggers a custom XLR AB switch bypassing the Meitner DAC. DIRAC is quite good, it's not as good as Mitch's convolution filter on Roon, but for special effects on an action movie it is more than fine. And it images extremely well even without a center channel.

PGA, after 40 years of wandering in the dead and desolate Audiophilia wasteland, we welcome you to the community of true high fidelity sound quality

 

[Carlos269]

PGA, after 40 years of wandering in the dead and desolate Audiophilia wasteland, we welcome you to the community of true high fidelity sound quality

If only more could see the light instead of banging their heads against the wall.

 

[schlager]

Yep, the willingness for flagellantism within certain types of audio folks, surprises me over and over again, even with the just finished Easter holidays in mind

 

[Hear Here]

I wonder how many DSP users first asked themselves "Do I have the right TYPE of speaker for my room's features?" before resorting to DSP.

In my view, this is by far the most important decision and, if the right choice is made, may avoid the need for DSP. But there are other factors too and these should be taken advantage of before DSP.

It's what DSP stands for that condemns it for me (or at least puts me off) - Digital SIGNAL PROCESSING. Do we want our precious signal processed any more than is absolutely necessary? We shun tone controls, graphic equalisers, etc but seem to accept the far more aggressive processing involved if Dirac, RoomPerfect, Anthem, MARS, etc is allowed to interfere with the signal that we thought was just about perfect as it left our expensive amplifier.

It's often referred to as "room correction" but it's nothing of the sort, as it doesn't do anything with the room. It's simply signal manipulation (messing with) in an attempt to sort out problems that should be sorted by other means. If we all used DSP to "fix" our room's deficiencies, why should speaker builders go to great lengths to create a good measuring speaker?

As long as the drivers between them can generate all the frequencies that our ears can hear, why should designers be bothered to make their frequency response flat when we can achieve this with DSP. After all, the DSP software and the microphone have no idea what is causing non-flat response - it could be the room, or speaker, or even amp or source. It's stated job is to sort out all these potential difficulties such that our ears receive a flat response. Shouldn't we prefer good and carefully chosen equipment and great attention with setting up to achieve this.

Granted complex multi-speaker and AV systems may need skills beyond most owners' abilities to get right, but 2-channel should be within our capabilities without resorting to additional signal processing.

That's likely to put the cat amongst the pigeons, but observations welcome!

 

[Carlos269]

I wonder how many DSP users first asked themselves "Do I have the right TYPE of speaker for my room's features?" before resorting to DSP.

In my view, this is by far the most important decision and, if the right choice is made, may avoid the need for DSP. But there are other factors too and these should be taken advantage of before DSP.

It's what DSP stands for that condemns it for me (or at least puts me off) - Digital SIGNAL PROCESSING. Do we want our precious signal processed any more than is absolutely necessary? We shun tone controls, graphic equalisers, etc but seem to accept the far more aggressive processing involved if Dirac, RoomPerfect, Anthem, MARS, etc is allowed to interfere with the signal that we thought was just about perfect as it left our expensive amplifier.

It's often referred to as "room correction" but it's nothing of the sort, as it doesn't do anything with the room. It's simply signal manipulation (messing with) in an attempt to sort out other problems that should be sorted by other means. If we all used DSP to "fix" our room's deficiencies, why should speaker builders go to great lengths to create a good measuring speaker?

As long as the drivers between them can generate the frequencies that our ears can hear, why should designers be bothered to make their frequency response flat when we can achieve this with DSP. After all, the DSP software and microphone have no idea what is causing non-flat response - it could be the room, or speaker, or even amp or source. It's stated job is to sort out all these potential difficulties such that our ears receive a flat response. I prefer good and carefully chosen equipment and great attention with setting up to achieve this.

Granted complex multi-speaker and AV systems may need skills beyond most owners' abilities to get right, but 2-channel should be within our capabilities without resorting to additional signal processing.

That's likely to put the cat amongst the pigeons, but observations welcome!

You do of course realize that analog signals attenuate with distance and are susceptible to noise, while digital signals remain pristine and are immune to noise correct? Digital is not what you make it out to be and neither are the conversion to and from analog and digital. Need to let go of old preconceived notions and outdated mindset. We live in a digital world, our brains are digital, and with quantum computing things are going to get radically more interesting and sophisticated really, REALLY quick. Have a listen to convolution through HQPLAYER and then come back and report your findings of what you hear, it will wreck your world, as you currently know it.

 

[schlager]

Hear Here, you should learn to distinguish between minimum and non-minimum phase response. That helps a lot when doing EQ and in understanding room acoustics.

 

[PGA]

There is an old saying in sales "if a guy wants a green suit, sell him a green suit". It's very difficult to "get the room right" and since our hobby is someone else's business/livelihood, the practical business solution is to sell us expensive boxes and cables that don't do much to improve the system, but have a high tech look, have high margins and fit in most rooms. Very few (almost no) hifi dealers or manufacturers have embraced DSP, which is a shame given what's possible today, which is light years different than 10 or 20 years ago.

 

[sparkie]

They don’t seem to have FIR capability

For years the approach was Impusle response files and some have different speaker libraries, but they are limited. The newer approach is to characterize the speaker's IP with measuring it with a calibration mic. Of course there are ways to load new profiles and the older ones had supplemental software to make your own IPR file.

 

[tony22]

View attachment 127925

PGA, this is an impressive but also somewhat interesting (to me anyway) result. It looks like you’ve got boosts of around 5dB or more between about 50-80Hz. That seems like a lot of boost at such low frequencies. I thought that was a no-no.

 

[sparkie]

You would need very large bass traps to fix the bass problems I have in my smallish room (12x18x8.5). It’s just not a practical solution. What ended up working extremely well was placing the TADs essentially in the corners along the 12 foot wall and eliminating the bass boost with DSP

Using the old bolt area methods, the room is long, but only wide at one end because of the self cancellations. If you use the mode method you might be able to clean most of it up, But like I said before, most effective bass traps are customized for the installation.


Here is your room in a mode simulator : https://amcoustics.com/tools/amroc?l=12&w=18&h=8.5&ft=true&r60=0.6

 

[PGA]

PGA, this is an impressive but also somewhat interesting (to me anyway) result. It looks like you’ve got boosts of around 5dB or more between about 50-80Hz. That seems like a lot of boost at such low frequencies. I thought that was a no-no.

There are two sets of graphs, the darker ones are before DSP, the lighter ones after. I don’t see any peaks in the after DSP graph, just the dips that are caused by the room and a gradual Harmon curve slope