The State of Speaker DSP

[Ralf11]

Ask not what you can do to your cables; ask what DSP can do for your speakers....

 

I think this is a very important technology, SQ wise.

 

Some speakers have it all built-in (Kii) and there are add-ons available too.  I am curious what DSP systems can now (beyond the obvious like freq. response).

 

One specific I seem to recall is that no or very few systems can do channels independently (e.g. if you have an asymetric room - like I do).

[Kal Rubinson]

47 minutes ago, Ralf11 said:

One specific I seem to recall is that no or very few systems can do channels independently (e.g. if you have an asymetric room - like I do).

Au contraire!  Almost all systems do each and every channel independently.  What they lack is the ability to correlate the correction in one channel with others with which they interact in the room.  

[Ralf11]

ok, thx (BTW, your Kii 3 review rekindled my interest in this)

 

interested in other abilities also...

[firedog]

There's the Dutch & Dutch 8c, which by all reports is a stiff competitor to the Kii Three.

There's also separate boxes like DEQX, which do all the DSP stuff with your speakers and/or DAC of choice (depending on model). 

[Daverz]

8 hours ago, Kal Rubinson said:

Au contraire!  Almost all systems do each and every channel independently.  What they lack is the ability to correlate the correction in one channel with others with which they interact in the room.  

 

Could you expand on this?  I'm not sure what you mean by correlate here.

[Keith_W]

There are 3 ways you can get DSP in your speakers: 

 

1. Purchase a speaker with DSP built in. Examples - Kii Three, B&O Beolab 90, Meridian 9000. The advantage is that there is little to no setup required. The disadvantage is usually lack of fine room correction. You can choose between a few modes (e.g. proximity to wall, bass level, etc) and that's it. Also, if it matters to you, you don't get to choose your own amplifier and you are stuck with whatever the manufacturer installed. 

 

The next two require either conversion of a speaker from passive to active, or DIY your own speaker. 

 

2. "DSP in a box", e.g. MiniDSP and DEQX. The advantage is that you gain more control over your speaker and you have room correction. You can spec your system to also correct for subwoofers and integrate the whole system. You can also BYO amps which is important to some people. The disadvantage is the learning curve. You have more opportunity to screw it up and end up with an awful sounding system. The other disadvantage is the relatively low processing power of these units, which limits the amount of correction you can do. Having said that, these units are usually powerful enough to be used in most systems. That is, unless you have many subwoofers and you need very long delays to correct multiple speakers. 

 

3. PC based DSP. In this case you use one software to generate filters (e.g. Acourate, Audiolense, Dirac, REW) which are then hosted by a convolution engine (e.g. JRiver, Acourate Convolver, Roon, HQPlayer). The advantage is that you are only limited by your skill and your computer's processing power and you can make really elaborate filters. You can even do things like MSED (Mid-Side encoder-decoder) and thus play with your soundstage, and even amazing things like a virtual double bass array. If you have a really complex system with multi-way speakers and multiple subs this is the best way to bring them all together. You can build the PC to be as powerful as you want, you are no longer limited by the hardware. Also, you can bring your own DAC's. The disadvantage is the sheer complexity and the learning curve, and thus you have even greater opportunity to screw things up. 

 

I have been running PC based DSP using Acourate and HQPlayer for some years now. To be honest thus far it has been a sideways move for me with some benefits (better coherence) and some downsides which I have been slowly chipping away at. The filters I make now sound so much better than my earlier efforts, but I am not quite there yet. More refinements will follow as I continue to climb the learning curve. 

 

Best yet, I have already spent all the money I need to spend. My system improves every month due to refinements with my technique, and these improvements are an order of magnitude greater than a DAC change (believe it or not!), let alone tweaks which I can barely hear (e.g. audiophile SATA cables). If you are an inveterate tinkerer of your system, and you actually enjoy making differences that really matter, going down this route will give you an immense amount of satisfaction. It was a giant leap of faith going down this path, but it has been totally worth it so far. 

[firedog]

DEQX has a great feature: buy one of their units and schedule a time for running the time/speaker correction and DRC. They will come online and run the program remotely for you so you have it professionally done in your listening room. 

[Kal Rubinson]

5 hours ago, Daverz said:

 

Could you expand on this?  I'm not sure what you mean by correlate here.

There are times when two or more speakers are reproducing the same sound (based on recording/mix).  As a result, correcting speaker 1 to be independently optimal and speaker 2 also independently optimal does not assure that, when playing them simultaneously, they sum properly.  For example, if they were somewhat out of phase at a given frequency, they would slightly cancel each other.

 

There are a few systems that will test adjacent pairs together after initial per channel corrections.

[goskers]

4 hours ago, Kal Rubinson said:

There are times when two or more speakers are reproducing the same sound (based on recording/mix).  As a result, correcting speaker 1 to be independently optimal and speaker 2 also independently optimal does not assure that, when playing them simultaneously, they sum properly.  For example, if they were somewhat out of phase at a given frequency, they would slightly cancel each other.

 

There are a few systems that will test adjacent pairs together after initial per channel corrections.

I would think that this same potential problem exists with a passive speaker as well?  

 

[Kal Rubinson]

5 hours ago, goskers said:

I would think that this same potential problem exists with a passive speaker as well?  

???? Sure.  Passive or active.

[goskers]

2 hours ago, Kal Rubinson said:

???? Sure.  Passive or active.

I should have elaborated a bit more. Is the scenario posed above where each loudspeaker is corrected individually using DSP unique to that scenario?

 

Wouldn't the same result be possible using a non-DSP based speaker?

 

if it is unique to that sceanrio can it be mitigated somehow?

[Kal Rubinson]

3 minutes ago, goskers said:

I should have elaborated a bit more. Is the scenario posed above where each loudspeaker is corrected individually using DSP unique to that scenario?

 

Wouldn't the same result be possible using a non-DSP based speaker?

It is common both to speakers with inbuilt DSP and to the external programs used with DSP and non-DSP speakers.

[aps]

On ‎24‎/‎04‎/‎2018 at 4:11 PM, Keith_W said:

There are 3 ways you can get DSP in your speakers: 

 

1. Purchase a speaker with DSP built in. Examples - Kii Three, B&O Beolab 90, Meridian 9000. The advantage is that there is little to no setup required. The disadvantage is usually lack of fine room correction. You can choose between a few modes (e.g. proximity to wall, bass level, etc) and that's it. Also, if it matters to you, you don't get to choose your own amplifier and you are stuck with whatever the manufacturer installed. 

 

 

Isn't it still possible to use DSP within the source to deal with overall room correction (in the event that on-board correction isn't good enough)? 

[Keith_W]

Yes of course it is