[jhenderson010759]

Question for John :

Can you do the measurements, apply driver response correction, and add the crossover without inserting any of the time domain / phase correction?


Rick

If we could trick the DEQX software into thinking that the phase and speaker response measurements were perfect, then it would apply no corrections to the frequency and phase responses.  Then, the transfer function of its filter would perform frequency division for the different drivers only.  Got to love it! All the abstraction and complexity without any of the benefits!  

I don't see a way to accomplish this, but I used to do precisely this using SoundEasy and the DCX2496 prior to installing the DEQX.  Then, I'd use the DEQ2496 to implement parametric equalization to correct for low-frequency anomolies.  My chain was PC to Roland M1000 digital mixer (acting as a soundcard and digital attenuator) to DEQ2496 to DCX2496 to amps to speakers.  I also needed a mic preamp.  All of this gear has been replaced by the DEQX in my current rig, except that I have retained the M1000 so that I can continue to use SoundEasy.

[DSK]

The notch filter usually used with the Excel drivers is necessary with analog crossovers of low slope to prevent ringing, since even low level signal at the problem frequency is problematic.  Since the DEQX can implement an effective "brick wall" filter below the problem frequency, there is no output at the problem frequency and no notch filter is needed.  

For drivers that have uncontrolled response within their usable range (for instance some of the B&G planars) the DEQX will reduce output at the appro ...

Thanks for that. Does this mean that it doesn't really matter if the desired drivers have rising/falling freq resp curves ...the DEQX will simly flatten them without the need for the extra passive parts usually required? As long as the drivers have similar 'sound' (if any), does DEQX mean we can 'almost' throw together any drivers and have them blend well?

[John Ashman]

DSK, if the drivers have good dispersion and low cone distortion, they will blend as though there is no difference.  That's why it's important, in general, to use very rigid drivers that don't have harmonic distortion in their useful bandwidth.

[DSK]

DSK, if the drivers have good dispersion and low cone distortion, they will blend as though there is no difference.  That's why it's important, in general, to use very rigid drivers that don't have harmonic distortion in their useful bandwidth.

Thanks John. I've always lived by the code that "if something looks too good to be true, it probably is.", but so far I haven't read of anything that can be done by a passive xo but not by the DEQX.

Arguably, the only downside seems to be price. Yes, I can understand the argument that the price isn't that expensive for the benefits it provides, but it's still a chunk of change (DEQX + amps + cables). I guess competition will increase over time and prices will drop. I look forward to reading more reports from enthusiasts that replace the internal passive xo's (on DEQX suitable speakers) with the DEQX, and getting a better feel for just how much sonic benefit can be heard.

As they say ....interesting times  

[JoshK]

Couple Qs about the DEQX:

1) Can you use it full range, without the xo features, say while you are mucking around, or to compare the xo features to standard implementation?

2) Does DEQX have any way to correct for impedance curves? Presumeably to gain the best implementation of a speaker design one should try to obtain a relatively flat impedance response of the drivers, via Zoble traditionally.  I know you can man-handle the ultimate FR which somewhat alleviates the need for a zobel, but I can't help but think a Zobel would still be optimal if the DEQX can't correct for it.  I can't see how it would.

[John Ashman]

1.  Yes

2.  Not really.  Not sure why this would be necessary exactly.  As I understand it, a lot of the more wicked impedance issues are because of the passive crossover (?)  I should think with this out of the way and drivers directly driven, impedance issues should be nearly the problem they were in the past.  Plus you can tailor the amps to the drivers, you've got more real efficiency (Meridian suggested in one paper that having 3 or 4 75W amps is dramatically more efficient than a ~300W amp through a crossover).  

Perhaps a non-linear compensation become available, things like these will be addressed, assuming they need it.  I'm surprised servo control hasn't reached the midrange frequencies yet.

[Rick Craig]

Couple Qs about the DEQX:

1) Can you use it full range, without the xo features, say while you are mucking around, or to compare the xo features to standard implementation?

2) Does DEQX have any way to correct for impedance curves? Presumeably to gain the best implementation of a speaker design one should try to obtain a relatively flat impedance response of the drivers, via Zoble traditionally.  I know you can man-handle the ultimate FR which somewhat alleviates the need for a zobel, but I can't help ...

Josh,

Since the crossovers are active / digital there is no need for impedance compensation.

Rick

[jhenderson010759]

I just replaced the Scan Speak D2905/99000 tweeters in my North Creek Rhythm satellites with the LCY 130 ribbons (from e-Speakers.com).  Took about an hour to remeasure, then modify my DEQX configuration to accomodate.  

What a difference!  The SS was certainly a nice tweeter, but these ribbons are much more to my liking.  The horizontal dispersion is extremely flat over a far wider angular range, and the effect on the stereo image is nothing short of remarkable.  This is the sound I have been seeking since I jumped back into the audio hobby a few months ago.  

Having used SoundEasy in the past to dial-in my DCX2496, I find the DEQX software unbelievably simple-to-use.  True, it's nowhere as flexible or capable as SE, but for the casual hacker such as myself, it is wonderful.  v1.90 is out, but I have used v1.50 up to now.  

I am really beginning to enjoy the DEQX now.

On an unrelated note, I downloaded a white paper from DRA labs on using a second pair of speakers in conjunction with a digital equalizer to perform ambience recovery (http://www.mlssa.com/surround/surround3.htm).  I'm setting this up now.  Has anyone else tried this?

[DSK]

...What a difference!  The SS was certainly a nice tweeter, but these ribbons are much more to my liking.  The horizontal dispersion is extremely flat over a far wider angular range, and the effect on the stereo image is nothing short of remarkable. ...

Jim, just did a quick search around on these LCY ribbons. From what I read, they have far better VERTICAL dispersion than typical ribbons, but there is little written on their HORIZONTAL dispersion. I saw one comment that suggested they are not as good as typical ribbons in this regard.

Did you definitely mean that HORIONTAL dispersion is wider?

[jhenderson010759]

Jim, just did a quick search around on these LCY ribbons. From what I read, they have far better VERTICAL dispersion than typical ribbons, but there is little written on their HORIZONTAL dispersion. I saw one comment that suggested they are not as good as typical ribbons in this regard.

Did you definitely mean that HORIONTAL dispersion is wider?

Yes, I do mean that the horizontal dispersion is much better than my ScanSpeaks.  The LCYs are notable in that their vertical and horizontal dispersion characteristics are both quite good - unlike some ribbons.  

Here (http://www.e-speakers.com/products/lcy-components.html) is a link illustrating the off-axis performance of the LCY-110 (similar to 130 in this regard).  As you can see, 30 degrees off horizontal is virtually identical to the on-axis response.  And, 30 degrees of vertical is only down abut 8dB compared to on-axis.  

By contrast, here (http://www.aurumcantus.com/g2g3/g1.htm) is a link illustrating the off-axis performance of the G1 tweeters produced by Aurum Cantus.  I have some Leisure 5 bookshelf speakers in my den which use the AC G3Si tweeters and they sound great.  But, I did have to carefully select stands for them so that they'd match ear-height in my listening position.

[Rick Craig]

Yes, I do mean that the horizontal dispersion is much better than my ScanSpeaks.  The LCYs are notable in that their vertical and horizontal dispersion characteristics are both quite good - unlike some ribbons.  

Here (http://www.e-speakers.com/products/lcy-components.html) is a link illustrating the off-axis performance of the LCY-110 (similar to 130 in this regard).  As you can see, 30 degrees off horizontal is virtually identical to the on-axis response.  And, 30 degrees of vertical is only dow ...

Jim,

Have you ran the DEQX with the ribbons without a series cap? One thing I'm concerned about is that the test signal is more like a swept sine wave than a benign MLS signal.

Rick

[jhenderson010759]

Rick -

Per your suggestion on Madisound, I ordered a Solen 50 uF cap for use in this application.  I had it all wired up on a protoboard for use.  But, after a discussion with Andre at e-Speakers, I was convinced that I would never saturate the LCY's transformer with the small offsets I am seeing out of my amp (~3 mV DC).  So, I am currently running these tweeters without the cap.  

The only problem I ran into during setup with the DEQX was that I neglected to bump the start frequency of the log sine sweeps during the speaker measurement process.  The DEQX software defaults to a starting frequency of 750 Hz when measuring the tweeters.  However, the LCYs Fs is 1000 Hz.  So, I was sweeping through the resonance during the tests.  Thankfully, the sweep is rapid, and the ribbons were not damaged during the test.  Apparently, there is nothing beyond my capacity for error.  

I am currently crossing over at 1800 Hz @ 96 dB/oct, so I'm probably down ~70dB or more at the device resonance during operation.  Plus, crossing this low means that I avoid the hump in the SS 8545 response which comes into play at frequencies above 2.5K or so as well as the associated "beaming".

Perhaps after I listen to the current setup for a couple of weeks, I'll insert the cap and see if I can discern any differences.

[Rick Craig]

Rick -

Per your suggestion on Madisound, I ordered a Solen 50 uF cap for use in this application.  I had it all wired up on a protoboard for use.  But, after a discussion with Andre at e-Speakers, I was convinced that I would never saturate the LCY's transformer with the small offsets I am seeing out of my amp (~3 mV DC).  So, I am currently running these tweeters without the cap.  

The only problem I ran into during setup with the DEQX was that I neglected to bump the start frequency of the log sine ...

Thanks for the information. Have you found the DEQX user's group to be helpful? Any idea of how many members?

[jhenderson010759]

Rick -

I think that the public DEQX users group is useless.  However, the private DeqxBeta is very helpful.  While there are only ~130 registered users on that forum, but many are quite knowledgable and helpful.  When I received the unit, I read all (~900) posts and it really helped me become familiar with the unit.  It also made me happy that I had not purchased one earlier, since there were a number of bugs that were resolved by other innocent victims prior to my initial use.  If I want buggy software, I'll write it myself.  

Have you obtained one of these units yet?  I know that DEQX is anxious to have speaker manufacturers, such as yourself, provide canned PDC-compatible speaker measurements for each of their speaker configurations.  Are you interested in doing that?  I would imagine that you might be concerned about the stability of the PDC and it's manufacturer.  But, aside from the initial equipment purchase, I think that your investment would be minimal should your choose to provide correction files for your products.  And, if you could provide accurate, individualized anechoic measurements for each speaker and it's drivers, I suspect that many users would pay extra for the data.  So, while your revenue may be reduced by not bundling a crossover, it could be embellished by providing data.  

Another potential marketing avenue is to integrate suitable amplification and even an OEM version of the PDC into each of the speaker enclosures.  Then, you could sell entire families of fully digitally-corrected speakers, ala the NHT system, but with transducers and cabinet du-jour.   Seems like the market is ready for that.

[PEB]

If I may jump in here, I have wanted to produce speakers with onboard DSP filtering and amplification for years.  Just plug in the speaker, and plug in the signal, and voila...

The problem with this as I see it, is that it becomes an issue of marketing scale (once again).  If I knew that many audiophiles want a speaker like this (which includes a portion of the processing chain, and all of the amplification chain), then I would dive in to the OEM offering from DEQX.  But on the whole, they don't. Part of the fun for audiophiles is the mix and match process.  It is enough for them to decide which is the right speaker for them, let alone which is also the (non-optional) DAC, DSP filter, and amp for them, too!

Compared to DSP solutions, my passive crossovers are quite labor intensive to build, and I would like nothing more than to develop an advanced filter alignment in DSP, and ship that instead (perhaps by email!). Imagine the ease of upgrade should I develop a new alignment.

I, too, am expecting a DEQX soon.  Hopefuly, adapting to the interface will not take too much of my time, as I already have a good handle of the filter theory.  

I really wish they had an 8-channel unit, for a design that is on my drawing board.  6-channels may dictate a design change, because the DEQX (or DCX) is an integral part of this design.  I don't intend to even offer it passively.

[jhenderson010759]

Phil -

I certainly agree that much of the "fun factor" for most audiophiles is mixing and matching components.  Personally, I think that iterating through different amplifiers, CD players and other electronics and particularly cables is a waste of time.  The specifications on any reasonable modern gear is virtually perfect, compared to the inaccuracies introduced by speakers and the interactions of the listening room.  

So, how about developing a speaker system that is much more modular?  I mean, who among us wouldn't like to try various different drivers for a while to see if they sound better?  If the cabinet could be constructed such that the baffle were modular, then different drivers could be explored rapidly.  True, this is a formidable task, since structural integrity is essential for a good speaker enclosure, but this is not insurmountable.  I am imagining something like the PartsExpress enclosures, but much higher grade and in many more esoteric styles.  Perhaps with enough thought, the product could emulate entirely different principles - such as open radiator designs.  

My view is that given decent-quality upstream electronics, speakers and their interaction with the room are 98% of the sound of a system.  If a vendor could streamline the ability of end-users to assemble a customized system, there could be a substantial market involving recurring sales...

Your view?

[Rick Craig]

Rick -

I think that the public DEQX users group is useless.  However, the private DeqxBeta is very helpful.  While there are only ~130 registered users on that forum, but many are quite knowledgable and helpful.  When I received the unit, I read all (~900) posts and it really helped me become familiar with the unit.  It also made me happy that I had not purchased one earlier, since there were a number of bugs that were resolved by other innocent victims prior to my initial use.  If I want buggy software ...

I haven't purchased the DEQX yet but am close to it. I'm interested in using it as a design tool and possibly for consulting.

[Rick Craig]

Phil -

I certainly agree that much of the "fun factor" for most audiophiles is mixing and matching components.  Personally, I think that iterating through different amplifiers, CD players and other electronics and particularly cables is a waste of time.  The specifications on any reasonable modern gear is virtually perfect, compared to the inaccuracies introduced by speakers and the interactions of the listening room.  

So, how about developing a speaker system that is much more modular?  I mean, who ...

I agree with you 100%. The problem is that the high end audio press and many manufacturers have convinced the majority of audiophiles otherwise.

[PEB]

Well, I couldn’t agree with you more.  As a design engineer, it just makes perfect sense to me, if not to everyone.

I'm already doing something along your modular line of thinking, but on a simpler scale.

Last year was about developing 5 monitors using just two woofers, two tweeters, and two cabinets.  Since the baskets are the same for both woofers, and same for the tweeter flanges, the owner can even upgrade his S2 driver set for S5 without changing cabinets.

Part of my incentive was to have a decent product line from a simpler inventory.  All the monitors had to be 2nd order acoustic alignments ($10 phrase for sealed box), with F3 around 70-80 Hz, and Qtc of 0.7 (Butterworth 2nd order).  This was to make them ready for use with THX specs when combining as Small speaker setting on most HT receivers.

But, the monitors can also be actively crossed the same way with outboard crossover such as the DEQX, DCX, or BESL AAF.  So the monitors work equally well for HT buffs or 2-channel guys.

This year is about taking the modular concept one step further.  Passive crossovers are being moved outboard.  This speeds up the custom dial-in process of fine-tuning each crossover as mated to the particular driver set.  Plus, the owner can later upgrade to active power by simply unplugging the passive networks, and connecting directly to multiple amplifiers.  By using a bi-wire Speakon connector in the center of the back panel, a single cabinet can be used as left, sideways center, or right simply by rotating the orientation of the connector.  (Believe me, inventory control is everything to a small business!)  The new sub was designed with tapered sides and front panel for several performance reasons.  Yet, the customer can buy the monitors before the subs, or vice versa, and build up his system as budget allows.  (Grain matching is possible only for 4-cabinet sales on day 1.)

I am selling interconnect sets for active setups, plus a new bi-wire cable (with spades now, and spades/Speakon later) for the monitors, and another cable for the subs.  I like to help the owner plan out his equipment location and cable lengths to optimize the imaging, and that is why I build the cables to length.

Ultimately, I hope to have bolt-on custom side and front panels that are a bit more exotic, so that those who really care about custom finishes can have their cake, too.  Maybe next year.

As for what constitutes the majority of what we hear in the system, my personal philosophy is that *it’s all about the output system*.  IOW, give me the ends of your preamp output cables, and let me get to work for you.  I like to design the filters and speakers, make amplifier serving suggestions, and encourage room treatment.  Do what ever you like for the source.

Finally, the real beauty of DSP active filters to me is this.  I can develop a speaker to measure with flat amplitude, uniform off-axis response, and be phase coherent no problem.  But any customer can take that speaker as a baseline, and then tailor the tonal balance to his own system, room, and personal preference, and not offend me in the slightest.  It’s his ears after all!

Sorry for the commercially bent post, but I've given this very subject a lot of thought for a long time now.  I'm truly making business decisions based on my philosophy.

[Rick Craig]

I just noticed that there is an excerpt on the DEQX site from a NHT/Xd review that David Rich did for the Sensible Sound. David is a very reliable source IMO having been the technical editor for The Audio Critic. This is great news to hear. Can anyone fill me in on the rest of the review?


Rick