Jump to content

Recommended Posts

One of the conversations in another thread discussing AA networks, I commented on the observation that if you carefully re-equalize your loudspeakers (and perhaps phase/time alignment) after changing passive crossovers--or just change out electrical parts like capacitors--that you can't tell the difference in sound from before the parts change. 

 

This phenomenon is apparently not widely known based on discussions in that thread. Many people apparently believe that certain components have a characteristic sound to them...sort of like an organismic view applied to audio components.

 

In another thread from over 5 years ago this topic was discussed at length.  The following three posts illustrate that discussion of the difference in sound from horn-loaded loudspeakers (a Belle Klipsch in this particular case) that are re-EQed and time aligned:

 

On 5/5/2014 at 0:18 PM, jorjen said:

Chris, what exactly do you hear/not hear when [time] misalignment is corrected? Just how significant a difference is it ?

On 5/5/2014 at 0:42 PM, Chris A said:

It sounds like I've got the wrong speaker between the Jubs when the delays are not corrected: a timbre mismatch.

 

This is most pronounced between the tweeter and midrange horns/driver if not corrected. Pianos don't sound right and female voices are also disturbed - perhaps "bigger than life" and not well focused within the soundstage image but simultaneously a narrow/thin soundstage image overall.

 

The midrange-bass bin crossover region sounds disturbed in timbre in the tuning fork "A-440" to middle-C region, and it gets worse/wider in its affected frequency band if using a gentle-slope crossover filter - such as 12 dB/octave Butterworth vs. 24 dB/octave L-R that I'm using now. The uncorrected sound is disjoint and nasal/horn-like in timbre. This is a bit more difficult to hear than the tweeter/midrange crossover distortion, but once you crank the delay correction back in and listen using A-B comparisons, it gets much easier to hear.

 

The soundstage image is disturbed where I can hear the center channel instead of it blending into one frontal soundstage image with the Jubs on either side (i.e., without delay correction). It also sounds like the Belle is very narrow in its apparent source width without correction, but blends very well when corrected.

 

The result of the two corrections is fairly extreme in my setup: I couldn't use the Belle between the Jubs without the delay corrections, IMHO.

On 6/20/2014 at 8:10 PM, Deang said:

Curious about the use of the word "timbre" in this thread. I don't see how a timbre match can be obtained since the Belles are still using phenolic diaphragms. Phenolics have their own character, they don't sound anything like titanium or aluminum -- so I'm guessing they wouldn't sound much like Beryllium either. I associate "timbre" with the character or sonic signature of the driver.

So this thread will illustrate the phenomenon of re-EQing to achieve certain loudspeaker sound signatures and give tips for those that are willing to move beyond typical "audiophile" organismic thinking related to passive crossovers and the exact materials used therein.

 

Chris

Link to comment
Share on other sites

At the time when I first placed the Belle between my Jubilees, I'd already tried out several center channel configurations, including:

 

1) phantom center

2) RC-62 (I couldn't stop laughing at how incredibly bad it sounded after buying one online and having it shipped to me)

3) phantom center (again)

4) Heresy I

5) Cornwall (1979 model)

6) phantom center (again)

7) Belle (stock configuration)

 

All of these sounded poor--including the Belle, like I was aware at all times that there was something playing that was different between the Jubs, so I took it upon myself to invest in another ElectroVoice Dx38 pre-owned, costing me $400(US).  After getting Dx38 set up, adding the two additional Crown D-75A amplifiers, and adding two zip-strip loudspeaker wires, I found that I needed to find the starting channel gains. 

 

I then scoured this forum's pages and found good references from Al K. (of ALK Engineering) and Bob Crites on what those starting channel gains should be, i.e., bass bin, midrange, tweeter channels.  I took a stab at it, and using only my ears and a 1/3 octave equalizer, I corrected the output and listened again.  It was much better, but still not really "invisible" between the Jubs. 

 

Then I re-installed Room EQ Wizard (REW) freeware and found my Behringer ECM8000 microphone and the small mixer that I use to provide phantom power to the microphone and to digitize the output of the microphone back to my computer via USB. [Nowadays you can buy a UMIK-1 that eliminates the separate mixer, providing USB plugin support directly.]  On the first upsweep, I could see many issues in the frequency response and the phase response of the stock Belle.  Here is a plot of the on-axis SPL and phase of that configuration at my listening position:

 

Belle - triamped without delay correction or EQ.png

 

You can't even see the phase above 800 Hz, even though the scale is zoomed out all the way.  So here is the first derivative of the phase curve (i.e., group delay), using a mathematical transform known as a Hilbert transform to show the "excess delay" of the crossover and time misalignments:

 

Belle Group Delay triamped without delay correction or EQ.png

 

It's pretty scary.  Any group delay above ~5 ms and 500 Hz is definitely audible. 

 

Chris

Link to comment
Share on other sites

After I got a little more EQ of the Belle, I managed to get it sounding much better than it had ever sounded, still without cleaning up all the phase misalignments and the microphone position:

 

Belle triamped with parametric EQing.png

 

It suddenly sounded like the Jubilees on either side of it, but you could tell that its "apparent source width" (ASW) was much narrower than the Jubs, and that it didn't have any low end bass, resulting in a timbre shift when playing pink noise for setting manual channel gain levels (using Audyssey). 

 

It stayed like that for about 2 1/2 years, until I decided to invest in a Beyma CP25 tweeter, a K-510 horn, and reuse one of my surplus K-69-A 2" drivers. 

 

Chris

Link to comment
Share on other sites

When I finally got the JuBelle set up (i.e., Belle bass bin, K-510 horn, K-69-A 2" compression driver, and a Beyma CP25 tweeter), it took me little time to start getting much, much better performance out of the center channel.  Here are the final plots of the on-axis SPL and phase:

 

JuBelle 3-way with CP25 and K-69-A.png

 

and the group delay plot, after I got all the phase misalignments out and the microphone positioning corrected:

 

JuBelle with CP25 @ 1 metre excess group delay.png

 

I finally had a center channel loudspeaker that I thought was up to the task--allowing me to walk from side wall to side wall across the entire width of the room just behind the listening positions without being aware of any change or localization of the sound at any point.  This was a spectacular experience!  The timbre was right on with the Jubs and the apparent source width was now more than adequate.  I was a happy camper--and felt as if this was the best combination that I could get into my listening room. 

 

(Little did I know that the forthcoming K-402 multiple entry horn [MEH] was going to trample the performance of this JuBelle configuration.)

 

Chris

  • Like 1
Link to comment
Share on other sites

Some comments from another forum: https://forum.audiogon.com/discussions/pass-labs-xvr-1-crossover

 

At its selling price ($5000US), its manual jumper discrete settings, and the fact that it's an analog active crossover without available delays except an all-pass filter, you might have some issues finding users among the horn-loaded loudspeaker crowd.  I could be wrong, however.

 

Chris

Link to comment
Share on other sites

While Schu waits for responses on the above piece of hardware, I'll recap the experiences above, i.e., the reason why I related those stories in this thread discussing "beyond passive crossovers":


1) By re-equalizing the loudspeakers after any changes to the crossovers are made (especially passive crossovers), timbre can be matched if you match the frequency response and phase of the loudspeakers that you're trying to integrate with. (With the Jubs on either side, it was easy to match amplitude/phase - just adjust the loudspeaker to have as flat a response as possible.)


2) Changing passive crossover components (such as capacitors) or even full crossovers...to active digital crossovers...the loudspeakers can be made to sound the same as before the change, assuming crossover filter slopes and phases are duplicated.  The "sound of certain materials of capacitors" I found to not be a factor at all. 

 

3) Using the "A/B" or "A-B" trace arithmetic function in REW also is useful in visualizing the differences between two captured traces to look for equalization and phase changes between crossovers or crossover settings.  If you don't move the microphone between crossover changes and re-running the REW up-sweeps, you can see the changes more clearly.  Also, Bill Waslo at Liberty Instruments has provided his freeware "Audio Diffmaker" application that also can be used to look for differences due to small changes in hardware on the amplitude or phase of loudspeakers. 

 

4) As the corrections for amplitude and phase/time delay are successfully adjusted for all loudspeakers in your setup, the realism and sound of the loudspeaker array becomes much more apparent and convincing...especially timbre and transient response.  Having your setup tuned in this way provides an avenue to unmastering your recordings for much greater realism and enjoyment than you might otherwise have believed possible.

 

Chris

Link to comment
Share on other sites

Chris,

 

I enjoy reading about your work, the great effort you've put into it, and the reporting of the results.

 

My recall is that PWK tried some delay and it must have been a pipe to bring the mid back into time with the woofer.  A tube on the tweeter seems un-workable.  He found IIRC that when re-equalized (was that via the little slide pot unit in the museum?) he found no difference.

 

Nonetheless, he didn't have the very precise processing and measuring units you are using.  If he did, he would likely agree with you. 

 

WMcD

Link to comment
Share on other sites

Physical time alignment can work if the higher frequency horn (whose horn mouth is further away from the listener), isn't occluded by the midrange or lower frequency horn mouths. 

 

Most multiway horns that are physically time-aligned usually have real issues in that regard that cause other listening issues...like diffraction in the crossover region if the horn centerlines are more than 1/4 wavelength apart, and coverage issues around the room from the occlusion of the larger horn mouths. 

 

1.jpg

 

Danley's multiple-entry horns are the only ones that are time aligned and aren't self-occluding...

 

Multiple_entry_horn.png

Link to comment
Share on other sites

I'm thinking that a first goal is to fabricate at least two more horns and assemble into corner speakers.  (For those that are not aware of this design, see the following thread for more information...)

The digital crossovers that I have on hand will be perfect for these home-fabricated units, although they need little delay--an all-pass filter is plenty since the physical unit is time aligned as-is. All that's needed is phase correction for the crossover network phase shift.  I found that the horn itself will resolve any phase mismatches internally, so you don't see the typical polar lobing outside the mouth due to woofer-HF phase mismatches.  This was a real surprise when I was playing with time delays. Only phase flattening across the woofer-HF pass band is needed via all-pass filtering, and even that isn't required, IMO.

 

The real need is for the parametric equalization filters and shelving EQ filters.   This is one instance that separating the woofer channel from the HF channel (bi-amping) is a smart thing to do since the effective sensitivity of the two channels are significantly different in order to achieve flat response.  [Danley uses gobs of attenuation on their HF drivers relative to the woofers for flat response, and that is why they're rated in the mid-90s (dB) in terms of sensitivity.]

 

So to answer your question: I think it would be an interesting exercise to design passives that approximate what Danley's passive crossovers do since time delay isn't needed for this horn configuration.   This work is definitely secondary on my list of things to do, however. 

 

I suppose that someone else that likes to design passive balancing networks with passive PEQ and shelving filters might like the challenge themselves.  I wouldn't think it would be difficult to reverse engineer a Danley Synergy crossover (assuming access to a Danley Synergy loudspeaker is available) but this is one area where Tom Danley himself is quite evasive in design details--never actually describing them.  I think that he views his "trade secret" IP as the crossover design itself, which isn't described in the current Danley Sound Labs patent (US 8,284,976 B2). 

 

A garden-variety active digital crossover with good PEQ/shelf filter capabilities is more than sufficient, however.  I plan to use the Yamaha SP2060 first for the corner horns, and a EV Dx38 for the surrounds.

 

Chris

Link to comment
Share on other sites

It has occurred to me in the past that the full-range K-402-MEH (multiple entry horn) would be a prime candidate for using miniDSP active digital crossovers since the prices are lower than equivalent passive networks having the same EQing capabilities (by several times).  The only thing to pay attention to, apparently, is the quality of the power supply to each unit.  In the past most people that attempted to use these complained about their noisiness, but newer information now points to the noisiness of USB power used, which is dependent on the supplying computer, not the miniDSP networks themselves.

 

I see this type of DIY using DSP modules as the logical extension of the DIY loudspeaker crowd--except using 21st century technology instead of 20th century technology (passive networks).  If you go over to diyAudio, you'll see that there is a very large contingent that's using miniDSP and other DIY DSP-based hardware/software. 

 

There is a reason why these folks are going that direction--namely significantly increased performance, lower price, and it is much easier to use (i.e., simply connect the components together and then dial it in using REW and a calibration microphone).

 

Chris

Link to comment
Share on other sites

13 minutes ago, tromprof said:

I have REW and a calibrated mic but I haven't spent the time to really figure it out yet.

 

If I can be of assistance, let me know.  I've tried to highlight the most common obstacles encountered in this and other threads so that others don't have to learn from the school of hard knocks.  Feel free to PM with any questions. 

 

I find that it doesn't have to take nearly as long as I've described above to come up to speed--perhaps a few hours to get most of the knowledge that you'll ever need to use REW and active crossovers extremely effectively.  I found that it's a lot easier for those that aren't engineers already--than learning DC circuit theory and using that knowledge to produce truly tailored passive crossovers for purpose.  It's certainly a lot cheaper nowadays doing the job extremely well. 

 

Chris

  • Like 1
Link to comment
Share on other sites

One of the interesting things about the K-402-MEH is that I could "cheat" on the phase compensation and push the group delay back down a bit by delaying the HF channel a little bit, without any apparent issues due to polar lobing (the crossover center frequency is 475 Hz):

 

K-402-MEH Excess Group Delay Minimization.png

 

and the corresponding SPL (on-axis), excess phase (white trace), and total phase (bottom red trace) plot:

 

K-402-MEH Phase+Excess Phase  Minimization.png

Link to comment
Share on other sites

  • Klipsch Employees
20 hours ago, Chris A said:

Physical time alignment can work if the higher frequency horn (whose horn mouth is further away from the listener), isn't occluded by the midrange or lower frequency horn mouths. 

 

Most multiway horns that are physically time time-aligned usually have real issues in that regard that cause other listening issues...like diffraction in the crossover region if the horn centerlines are more than 1/4 wavelength apart, and coverage issues around the room from the occlusion of the larger horn mouths. 

 

1.jpg

 

Danley's multiple-entry horns are the only ones that are time aligned and aren't self-occluding...

 

Multiple_entry_horn.png

time aligned?  how deep is the horn in the illustraltion?

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...