Hi all -
Continuing a discussion found both here https://www.diyaudio.com/community/threads/the-mysterious-danley-crossover.364891/ and https://www.diyaudio.com/community/...eh-synergy-project.409866/page-5#post-7631790, opening up a dedicated thread (in violation of @Patrick Bateman's first rule of Danley fight club
) to discuss the potential sound quality benefits of an external crossover on an SH50 vs their stock passive crossover on recent models. I realize this question is probably best fit to discuss with Danley themselves (and I've asked), and I realize this isn't really "DIY", but since there seem to be a good amount of members with direct experience with the SH50, I figured there could be some nice discussion.
Overview: I'm leaning towards a Danley solution for home audio, 2 channel mainly. I've tried several other highly rated (ASR) speakers but they don't seem to be getting me closer to a "Berghain in my living room" sound that I want to achieve. So, time to try an actual PA. I like Danley because I've heard them many times in a club setting (pretty much every new club in NYC is kitted with Danley these days), and the Synergy design promises great results when it comes to side/back wall reflections. Plus, they'll hold their value much better than most small-brand hifi speakers.
My choice is now between a boxes with passive crossover, or tri-amped ones built for external crossover. I'm concerned only with sound quality, not resilience. The question: are there known shortcomings with the SH50 passive crossover such that better sound quality in some dimension can be achieved going active externally? Assumptions: I'd be running the whole thing through a MiniDSP w Dirac for holistic room correction, so presumably this would take care of in-room frequency response. What's left seems to only be phase issues on individual drivers. Are there other aspects to consider?
Given the complexities of building passive crossovers, there must be some residual tradeoffs that Danley have made for their crossovers. In experimenting with these boxes, has anyone noticed what these tradeoffs might be?
Continuing a discussion found both here https://www.diyaudio.com/community/threads/the-mysterious-danley-crossover.364891/ and https://www.diyaudio.com/community/...eh-synergy-project.409866/page-5#post-7631790, opening up a dedicated thread (in violation of @Patrick Bateman's first rule of Danley fight club
) to discuss the potential sound quality benefits of an external crossover on an SH50 vs their stock passive crossover on recent models. I realize this question is probably best fit to discuss with Danley themselves (and I've asked), and I realize this isn't really "DIY", but since there seem to be a good amount of members with direct experience with the SH50, I figured there could be some nice discussion.Overview: I'm leaning towards a Danley solution for home audio, 2 channel mainly. I've tried several other highly rated (ASR) speakers but they don't seem to be getting me closer to a "Berghain in my living room" sound that I want to achieve. So, time to try an actual PA. I like Danley because I've heard them many times in a club setting (pretty much every new club in NYC is kitted with Danley these days), and the Synergy design promises great results when it comes to side/back wall reflections. Plus, they'll hold their value much better than most small-brand hifi speakers.
My choice is now between a boxes with passive crossover, or tri-amped ones built for external crossover. I'm concerned only with sound quality, not resilience. The question: are there known shortcomings with the SH50 passive crossover such that better sound quality in some dimension can be achieved going active externally? Assumptions: I'd be running the whole thing through a MiniDSP w Dirac for holistic room correction, so presumably this would take care of in-room frequency response. What's left seems to only be phase issues on individual drivers. Are there other aspects to consider?
Given the complexities of building passive crossovers, there must be some residual tradeoffs that Danley have made for their crossovers. In experimenting with these boxes, has anyone noticed what these tradeoffs might be?
Thanks for shifting the discussion to a new thread. The reason why I asked was because I was thinking that the other thread was going to become overwhelmed by the comments.
My assessment of the SH-50 passive crossover is that its only shortcoming is achieving ±1.5 dB amplitude response across its nominal 50-20,000 Hz passband (a characteristic which is easily EQed back to that level of flatness using careful EQ upstream). Here is an excerpt from Erin's review using a Klippel NFS (R&D) showing what I'm talking about:
This is nominally in half space/full space (i.e., there are LF effects of having a floor present at the lowest frequencies, but not nearly as much as the full 6 dB of rise in response would suggest). In my listening room in a mid-wall location, the SH-50 exhibits a much greater rise in amplitude response between 60-300 Hz due to quasi-quarter-space loading, and must be EQed flat again.
This is not the case in just about any other passive loudspeaker crossovers that I've encountered. Most passive crossovers usually have other issues, not the least of which is heating of passive components when playing the loudspeaker at "hand volume" (i.e., at or above 90 dB/1m). Other crossovers and driver voice coils simply do not have the efficiency of operation to avoid electrical resistance rise, etc. that plagues almost all direct radiating loudspeakers. Much more can be said on this subject, but I think that the key readers here may be familiar with this subject. You can check Erin's Audio Corner on this subject to see just how immune from heating effects the passive crossover and driver voice coils are. [A warning: I recommend the measurement plots at that site, but not the conclusions drawn by the article's author, which I think are quite off the mark. The Klippel NFS is a good piece of automated gear, however. YMMV.]
If you're think of using the SH-50s well above ~100 dB/1m continuously, I think you might start to see thermal effects begin to creep in, but for home operation, I don't really see it.
Beside the thermal stability of the SH-50 passive crossover, the other surprising characteristic of the network/drivers is that it has mostly linear phase response. To my knowledge, I know of no other loudspeakers with passive crossovers that exhibit this behavior. Below you will see this shown graphically:
What you see above are the overlays of the individual driver phase responses (compression tweeter = red trace, midrange cones = green trace, woofers = yellow-orange trace) with their combined responses (indigo trace), showing that the output of the SH-50 with passive crossover is minimum phase. I can't tell you how much this surprised this writer when I measured it myself, and began to scratch my head as to how Mr. Danley achieved this result. (What then happened is that I also discovered how to do it myself using DSP crossover IIR filters only.)
Folks, in my experience, this is the actual "secret sauce" of the Synergy series of loudspeakers from DSL over the prior Unity horns from Sound Physics Labs--linear phase using passive crossovers.
Additionally, you can verify the linear phase argument by looking at the excess group delay plot:
(the cyan trace only) and its step response (again, cyan trace only):
One really doesn't have a driving argument for using FIR filters with this loudspeaker design (passive networks).
So the above says that, if you can get the SH-50s with their stock passive networks, you will sacrifice little if anything in terms of audible sound quality performance. If you've got FIR filtering to lay on top of the passive response to tilt that linear phase response down to "flat phase response", that's certainly doable. Will you be able to hear the difference? Probably not.
YMMV.
Chris
My assessment of the SH-50 passive crossover is that its only shortcoming is achieving ±1.5 dB amplitude response across its nominal 50-20,000 Hz passband (a characteristic which is easily EQed back to that level of flatness using careful EQ upstream). Here is an excerpt from Erin's review using a Klippel NFS (R&D) showing what I'm talking about:
This is nominally in half space/full space (i.e., there are LF effects of having a floor present at the lowest frequencies, but not nearly as much as the full 6 dB of rise in response would suggest). In my listening room in a mid-wall location, the SH-50 exhibits a much greater rise in amplitude response between 60-300 Hz due to quasi-quarter-space loading, and must be EQed flat again.
This is not the case in just about any other passive loudspeaker crossovers that I've encountered. Most passive crossovers usually have other issues, not the least of which is heating of passive components when playing the loudspeaker at "hand volume" (i.e., at or above 90 dB/1m). Other crossovers and driver voice coils simply do not have the efficiency of operation to avoid electrical resistance rise, etc. that plagues almost all direct radiating loudspeakers. Much more can be said on this subject, but I think that the key readers here may be familiar with this subject. You can check Erin's Audio Corner on this subject to see just how immune from heating effects the passive crossover and driver voice coils are. [A warning: I recommend the measurement plots at that site, but not the conclusions drawn by the article's author, which I think are quite off the mark. The Klippel NFS is a good piece of automated gear, however. YMMV.]
If you're think of using the SH-50s well above ~100 dB/1m continuously, I think you might start to see thermal effects begin to creep in, but for home operation, I don't really see it.
Beside the thermal stability of the SH-50 passive crossover, the other surprising characteristic of the network/drivers is that it has mostly linear phase response. To my knowledge, I know of no other loudspeakers with passive crossovers that exhibit this behavior. Below you will see this shown graphically:
What you see above are the overlays of the individual driver phase responses (compression tweeter = red trace, midrange cones = green trace, woofers = yellow-orange trace) with their combined responses (indigo trace), showing that the output of the SH-50 with passive crossover is minimum phase. I can't tell you how much this surprised this writer when I measured it myself, and began to scratch my head as to how Mr. Danley achieved this result. (What then happened is that I also discovered how to do it myself using DSP crossover IIR filters only.)
Folks, in my experience, this is the actual "secret sauce" of the Synergy series of loudspeakers from DSL over the prior Unity horns from Sound Physics Labs--linear phase using passive crossovers.
Additionally, you can verify the linear phase argument by looking at the excess group delay plot:
(the cyan trace only) and its step response (again, cyan trace only):
One really doesn't have a driving argument for using FIR filters with this loudspeaker design (passive networks).
So the above says that, if you can get the SH-50s with their stock passive networks, you will sacrifice little if anything in terms of audible sound quality performance. If you've got FIR filtering to lay on top of the passive response to tilt that linear phase response down to "flat phase response", that's certainly doable. Will you be able to hear the difference? Probably not.
YMMV.
Chris
Chris, are you really saying you think the SH-50 with its passive crossover achieves linear-phase?
Uh yeah, ....what's a few hundred degrees of phase rotation...that's still linear phase, huh ? Lol
Depends on the synergy...it's H-V pattern.
It's easy on the SH-50...but gets tougher on less narrow horns.
Is that what you're getting?
https://www.diyaudio.com/community/...bandpass-mid-in-unity-horn.88237/post-2838793
Uh, you do realize that link is for a 4-way active speaker, don't you??
And that all discussion has been about the phase response of the passive SH-50?
So how does your link fit in?
Amen.I think a couple of you fellas are confused on the differences between minimum-phase, linear-phase, and flat-phase.
Dave.
Not that it should change the point.. in any case the J2 was made in passive and active versions, thanks for reminding me. I can't see which this one was. Perhaps the active was IIR?
I understand you're upset, but we should really keep this civil.
Maybe a version was out for a while with the mid-to-CD crossover being passive? Dunno...do you have a link? Looks all active nowadays.
I've always assumed a box like the J2, with twelve 1" CDs, twenty-four 5" mids, and six 18" lows would have current draws too high to allow economical passive xovers. Even for the mid-to-highs.
I looked for DSP presets for any of the Jericho models, in Danley's ASC-48 processor. Doesn't have any.
But other biamped SH models show the use of both IIR and linear phase FIR filters.
IIR is always going to be a given with live-sound boxes...no other way to handle lower-frequency work, without unacceptable latency.
However, once above several hundred Hz, lInear phase FIR filters/xovers become doable in terms of latency...and simply make for easier and better xover implementations, ime/imo.
DSL appears to think so too, given the presets I'm seeing.
Your understanding is incorrect
And I really can't see how I was being uncivil, by enthusiastically agreeing with another's post.
I think maybe this is a basic misunderstanding of what linear-phase is.
The DSL SH-50 data sheet shows a phase shift of almost 1000 degrees across the audio band. (As does the indigo-colored plot shown above.)
That's not linear phase. Not in my book anyways.
There's marketing talk and then there's actual data. You need to be able to separate the two.
Dave.
From Wikipedia:
Also, from the same source:
I would say that it is wise to be careful about inserting your own personal definitions for commonly accepted definitions, especially when the subject matter is engineering-oriented (such as signal processing).
Yes, I agree.
If you're talking about "zero phase" response, then I think we're talking about the same thing. However, there is a reason why linear phase is not defined as "zero phase".
There is also a good reason why I plotted the group delay: zero or constant group delay indicates linear phase. It should: the first derivative of phase with respect to frequency (assuming LTI) is how group delay is calculated.
Chris
Last edited:
I think a couple of you fellas are confused on the differences between minimum-phase, linear-phase, and flat-phase.
I admit to be confused!
I volunteer to ask the stupid questions. The following is part of a post i wrote some days ago in the tread that evolved into this thread:
"I am NOT an educated engineer, but have been reading about synergy horns for a long time now and am very slowly building my own. I have this idea of what a REAL synergy-horn is, given it is big enough to be "fullrange":
A: It is as close as you can get to a "perfect" fullrange-pointsource-loudspeaker-driver that reproduces the music as true to the original signal as possible.
B: Minimum-phase as a perfect dirac-pulse in the audible spectrum. Reproducing natural sound! Like a perfect fullrange-driver without a cross-over!
This is what I think is the goal Tom Danley follows, if I have understood him right. Like when he mentions Richard Hayser and his research."
I would like to know from you who understand "the differences between minimum-phase, linear-phase, and flat-phase".
Are statements A and B saying the same and do they make sense? Am I thinking right, expressing my self correctly. Please correct me if I am wrong, as I really want to learn.
Could one say that the SH 50 is close to be a minimum-phase devise in the region from say 80 Hz to almost 20kHz where it shows almost flat-phase?
How does the phase-response look for at minimum-phase devise in the audible spectrum?
At the risk of appearing as a total idiot i post this and see what happens.
Maybe I get smarter!Regards
Steffen
Even Dr. Floyd Toole seems to have trouble with this subject in his book (Sound Reproduction--Loudspeakers & Rooms, all three editions). While it's pretty clear to me that Toole would agree in principle with your statement, I think even he would hedge his answer in order to justify his apparent personal use of planar-type loudspeakers. (Also, Harman doesn't own the Unity horn patent--so I think he would tend to hedge his comments on this subject.)
I do agree with your quoted (above) statement, but I also recognize that I was probably complicit in helping to form such a viewpoint. (YMMV.)
Most of my posts on this subject focus on eliminating "all pass" behavior of crossover filters in loudspeakers (i.e., electrical filters that add time/phase delays within a multiway loudspeaker), and in the acoustic analog of that--i.e., in-room early reflections--as a way to much better subjective fidelity. I think that the "full range driver" crowd (some of which are found on diyAudio) are also after single driver-type loudspeaker electro-acoustical response. I just think that the MEH approach is simply a better approach, with no real downsides acoustically (objectively and subjectively).
So while your first sentence, quoted just above, may be problematic, the last sentence is much closer to what I've experienced in well-designed and dialed-in MEHs. I'm not alone in my opinion on this subject. The inventor--Tom Danley--basically says the same thing. I think he said it first, however.
Chris
Last edited:
It appears that you're "winding up" (i.e., to argue).
I just call them as I see them. If you wish to argue with the data that I presented in post #2, I can discuss how I took it, and how I read it. But anything further than that (like forum zero-sum games) isn't what draws me to diyAudio. I do try to help others when they have specific questions on matters that I believe I can help with. That's what I've done here...nothing more.
Chris
Using a synergy design with passive filters is IMO a massive compromise in regards to achieving the best overall performance in a given room.
The frequency band from 300 to 10k needs to be as smooth as possible. Phase linearity is not as important as achieving smooth phase without sudden jumps. Small abrupt swings are very noticeable in comparison.
Most synergy speakers by design are very complex and require lots of correction to linearize even without considering the added issue of in room FR correction. You'll always need some sort of EQ to fix these FR issues created by the room itself.
The in room LF peaks and dips alone are reason enough to employ DSP, so you might as well just do everything with DSP and have all the other benefits that come along with using it. I'm a pretty big advocate of using passive filters and keeping the signal path all analog. This is very limiting with designs requiring even the smallest amount of time domain related correction. The only issue IMO is the transient response of the filters ie pre-ringing and also signal to noise ratio. High sensitivity compression drivers require high signal to noise ratio converters and output stages to avoid hiss. This is usually where most types of DSP have issues. My typical choice of DSP is the RME brand. Their ADI2-Pro equipment is the best I've heard considering the cost. The filters and DACs they use are very transparent. Most pro audio type speaker processing boxes don't give audiophile grade sound. Mini DSP stuff is so-so IMO and not considered sufficient for higher end use.
The frequency band from 300 to 10k needs to be as smooth as possible. Phase linearity is not as important as achieving smooth phase without sudden jumps. Small abrupt swings are very noticeable in comparison.
Most synergy speakers by design are very complex and require lots of correction to linearize even without considering the added issue of in room FR correction. You'll always need some sort of EQ to fix these FR issues created by the room itself.
The in room LF peaks and dips alone are reason enough to employ DSP, so you might as well just do everything with DSP and have all the other benefits that come along with using it. I'm a pretty big advocate of using passive filters and keeping the signal path all analog. This is very limiting with designs requiring even the smallest amount of time domain related correction. The only issue IMO is the transient response of the filters ie pre-ringing and also signal to noise ratio. High sensitivity compression drivers require high signal to noise ratio converters and output stages to avoid hiss. This is usually where most types of DSP have issues. My typical choice of DSP is the RME brand. Their ADI2-Pro equipment is the best I've heard considering the cost. The filters and DACs they use are very transparent. Most pro audio type speaker processing boxes don't give audiophile grade sound. Mini DSP stuff is so-so IMO and not considered sufficient for higher end use.
Hmmm...a bit of nitpicking going on I think....
"Zero" phase is simply a special case of linear phase, when the phase is a flat line with a slope of zero across the spectrum....
which means nothing more than all fixed-time constant-delay has been removed. Fixed time being time-of-flight, fixed processing time, etc.
Remove the fixed-time delay, and you get zero phase.
Linear phase is the same damn thing, other than fixed-time delay isn't necessarily removed.
Phase will be a straight line again, but with a downward slope that equates to the constant delay applied to all frequencies.
BUT NOTE: It will only be a straight line when frequency is displayed on a LINEAR scale, not a logarithmic scale.
If and when you seen a straight-ish looking phase trace sloping downward, on frequency displayed log scale, like in post #2...
that is simply not linear phase. The straight line must be on a linear freq scale.
Linear phase has no relative phase rotation across frequencies, either with or without fixed delay removed.
Last edited: