Thanks again Art.
I will persist with the BMS 4550 further.
It is perhaps still running in at home studio levels - I read about one month may be necessary.
PA levels round be a few hours 😂
I can't cross much higher with the MF running out of puff pretty sharpish after 1000Hz.
BMS4550 at 1200Hz as you suggested knits well.
I will closely examine the response at the listening spot and see if there is anything more to make it pleasing enough.
Definitely closer.
I fast read your epic driver comparison test. Thanks for the link.
Amazing work!
I can't open the zip files with the Windows unzippers I downloaded in their store. Is there another trick?
My wife's usual evaluation request of Rihanna's -
On my old system (S2s just on 3uF caps), that vocal (played on Qobuz) brought her to tears, it sounds that good.
Wife's a singer and loves vocals like that.
The SH-50 alikes are some way short.
I hear it too of course.
It's a tough test of a revealing compression driver.
This was soon after the 2nd horn being ready so, it's improved a bit since..
The "trick" was in post #2:
The file type suffix < .zip > must be changed to < .mP3 >, they will then open with your mP3 app.
The surrounds and spiders on loudspeakers generally need to be stretched past Xmax to "break in", but I doubt the BMS 4550 Polyester diaphragm response will change in a month at home studio levels.
That said, you might get used to it's sound in a month
Thanks Art,
Too much skim reading on my part 🙂
I'm of a mind to try the Vitavox S2s on the back of the SH50 alikes.
Mounting the CDs as close as possible to the MF drivers.
The horn throat can be adapted to add that 1/4" extra width either side.
As always DIYAudio is a great source of info. Found this;
Source https://www.diyaudio.com/community/threads/two-way-synergy-horn.367812/#post-6532149
Which asks the question, where does the measurement start from? - the answer from that thread was from the CD diaphragm.
That would put a dampener on the strict 1/4 wavelength idea.
The S2s are pretty deep. Diaphragm near the rear.
However, then I read this
"It is my opinion that the 1/4 wave rule is not necessary and I believe that Tom's second patent removes it. One simply wants the drivers as close together and as far down the device as possible. Everything else is taken care of with EQ (to the extent that is possible.)" - gedlee
Source
https://www.diyaudio.com/community/...tanding-of-synergy-horns.309258/#post-5113817
How far does that rule bend?
Of course there is a picture of this.
Diaphragm well back from the CD exit.
Source
Not sure.
My S2s would need the Raal tweeters atop for the above 15KHz or so stuff.
I wouldn't mind that one bit if I got the S2 sound and the Synergy integration.
I could also turn / make a throat adapter to refit the BMS4550s for my outdoor events..
I see from these older threads many have contributed who have helped me with this thread - excellent stuff.
So, would it work?
Too much skim reading on my part 🙂
I'm of a mind to try the Vitavox S2s on the back of the SH50 alikes.
Mounting the CDs as close as possible to the MF drivers.
The horn throat can be adapted to add that 1/4" extra width either side.
As always DIYAudio is a great source of info. Found this;
Source https://www.diyaudio.com/community/threads/two-way-synergy-horn.367812/#post-6532149
Which asks the question, where does the measurement start from? - the answer from that thread was from the CD diaphragm.
That would put a dampener on the strict 1/4 wavelength idea.
The S2s are pretty deep. Diaphragm near the rear.
However, then I read this
"It is my opinion that the 1/4 wave rule is not necessary and I believe that Tom's second patent removes it. One simply wants the drivers as close together and as far down the device as possible. Everything else is taken care of with EQ (to the extent that is possible.)" - gedlee
Source
https://www.diyaudio.com/community/...tanding-of-synergy-horns.309258/#post-5113817
How far does that rule bend?
Of course there is a picture of this.
Diaphragm well back from the CD exit.
Source
Not sure.
My S2s would need the Raal tweeters atop for the above 15KHz or so stuff.
I wouldn't mind that one bit if I got the S2 sound and the Synergy integration.
I could also turn / make a throat adapter to refit the BMS4550s for my outdoor events..
I see from these older threads many have contributed who have helped me with this thread - excellent stuff.
So, would it work?
I can still remember what the above 15kHz or so stuff sounded like
Short answer- assuming the S2 throat exit is not reduced, it will work fine on your MEH, but will require a different time/phase alignment and EQ than the BMS4550.
One quarter wavelength is a 90 degree phase difference.
Two elements at 90 degrees phase difference combine to +3dB, within 3dB of "perfect".
Close enough for rock and roll..
The closer to 0 degree phase difference, the better the acoustical summation.
While physically in front of the HF driver, the acoustical low pass of the cone mids through their exit ports may result in their output being in phase alignment with the HF driver.
The physical distance alone does not insure phase alignment within 90 degrees.
In combination with passive electrical filters (and polarity reversal if required) the outputs may be brought into phase alignment.
As has been pointed out previously in this thread, passive alignment (and adequate driver protection at high output) is complicated.
The HF to the MF virtual distance can be adjusted with digital delay to bring it to 0 degree phase difference for a +6dB summation.
The distance between the mid ports is fixed by the horn geometry, if they are not within ~1/4 wavelength distance apart relative to the horn axis, off axis cancellation will occur, and can not be changed with any filter type or delay.
This distance limitation can still allow central ports with small diameter mid cone drivers, but with larger drivers like the 8" in DSL's SH95, or the 15"s in EV's MTS series of MEH, requires the ports to be located on the throat side of the cone to be within 1/4 wavelength distance across the horn throat in the crossover region.
The low coax mid-high drivers in the EV MTS series are crossed over to the 15"s at ~450Hz.
1/4 wavelength of 450Hz is ~7.5".
The distance between the coax 15" exit ports and the coax drivers is around double that, near a half wavelength, 180 degrees, near total cancellation...
That "cancellation problem" is easily corrected with DSP.
Art
Short answer- assuming the S2 throat exit is not reduced, it will work fine on your MEH, but will require a different time/phase alignment and EQ than the BMS4550.
One quarter wavelength is a 90 degree phase difference.
Two elements at 90 degrees phase difference combine to +3dB, within 3dB of "perfect".
Close enough for rock and roll..
The closer to 0 degree phase difference, the better the acoustical summation.
While physically in front of the HF driver, the acoustical low pass of the cone mids through their exit ports may result in their output being in phase alignment with the HF driver.
The physical distance alone does not insure phase alignment within 90 degrees.
In combination with passive electrical filters (and polarity reversal if required) the outputs may be brought into phase alignment.
As has been pointed out previously in this thread, passive alignment (and adequate driver protection at high output) is complicated.
The HF to the MF virtual distance can be adjusted with digital delay to bring it to 0 degree phase difference for a +6dB summation.
The distance between the mid ports is fixed by the horn geometry, if they are not within ~1/4 wavelength distance apart relative to the horn axis, off axis cancellation will occur, and can not be changed with any filter type or delay.
This distance limitation can still allow central ports with small diameter mid cone drivers, but with larger drivers like the 8" in DSL's SH95, or the 15"s in EV's MTS series of MEH, requires the ports to be located on the throat side of the cone to be within 1/4 wavelength distance across the horn throat in the crossover region.
The low coax mid-high drivers in the EV MTS series are crossed over to the 15"s at ~450Hz.
1/4 wavelength of 450Hz is ~7.5".
The distance between the coax 15" exit ports and the coax drivers is around double that, near a half wavelength, 180 degrees, near total cancellation...
That "cancellation problem" is easily corrected with DSP.
Art
Last edited:
Interesting that Art would post that phase wheel.
To the OP ( @Speedysteve7 ): look at the top of that figure from 305 degrees to 55 degrees (i.e., ±55 degrees of phase). You'll notice that the output sums to 1 dB or less of the full 6 dB of overall boost in output for having two mostly in-phase sources playing in phase at the same frequency, and within 1/4 wavelength of each other.
My comment in another thread on the SH-50 about the loudspeaker w/passive crossover being "mostly linear phase" recognizes this fact. Once you're within something like ±90 degrees of phase (average across a particular passband), the human hearing system will pick this out as "linear phase". When you play complex waveforms of music with peaks in their output over time due to multiple frequencies being played at the same time (i.e., harmonics of the individual instruments/voices, and a summation of the different musical voices playing together at the same time produce periodic and asynchronous spikes in the time-domain output of the music that the human hearing system picks up--and the sensation that is heard is something called "clarity", i.e. (from slide 17 of Griesinger's "What is Clarity?"):
Below you will find an excerpt from one of Toole's JAES papers on the subject of inferring subjective fidelity from objective measures. The black lines represent the phase response of the best loudspeakers tested in that study from the mid 1980s (i.e., using only passive crossovers), and the red line represents the type of phase response which is easily achievable using DSP IIR filters only today, especially MEHs:
Now you can see the degree of phase distortion present is typically passive crossovers (yesteryear's and also today's passive crossover loudspeakers), as compared to the easily achievable phase response from a typical MEH. That's a huge difference--and one that's quite audible if you clean up the early reflections in-room.
So when you have a loudspeaker that has several hundred degrees of phase growth from high frequency to low frequency, (especially those crossover networks having higher order crossover filters)...the human hearing system perceives that as a loss of clarity...like having too many early reflections to muddle the direct arrivals) and a shift in timbre. And asynchronous spikes in the waveforms that were recorded faithfully by the recording microphone are now effectively wiped out due to the phase growth of the loudspeaker passive crossover networks alone.
This effect is the greater part of the reason why "Synergy Horns" (the trade name) --a particular type of MEH implementation which pays attention to minimizing phase growth--has such a following: a nearly linear phase combined with a fully horn-loaded loudspeaker having inaudible modulation distortion sidebands--even up to levels well above 100 dB/1m, and controlled directivity down to the room's Schroeder frequency.
So the threshold of phase swing audibility is something like (on the order of) ±90 phase swings. Any phase swings greater than this starts to become audible (just as Tom Danley himself has mentioned before). Less than ±90 degrees of local or full passband phase shift, and the phase distortion is essentially inaudible (IME).
Just gazing at a phase response waveform to guesstimate what amount of phase growth is audible or inaudible based on visually what the observer thinks might be audible with his/her eyes--this isn't the requirement. The threshold of audibility of phase distortion is what's important.
JMTC.
Chris
Thanks guys, great insight.
As the shotgun bearing villain in Dirty Harry said, "I gots to know!"
This punk might get lucky, or it might bomb - we'll see.
Assuming DSP time aligning can compensate for the extra distance, diaphragm to MF taps, this morning being perfect indoor weather, I cracked on and modded the horn throat to be a perfect 1.5" / 38.1mm.
Will be fitting the spare screw on ring adapters, I turned / thread cut ages a
go..
Driver and Le Cléac'h horn (latter on hold for today, or longer?).
Trial pop on top
1" Ferrite BMS 4550 Vs 1.5" Alinco Vitavox S2
Can you tell which one I'm rooting for?😃
Gonna be interesting..
First one will be listenable this afternoon.
Will know soon enough.
As the shotgun bearing villain in Dirty Harry said, "I gots to know!"
This punk might get lucky, or it might bomb - we'll see.
Assuming DSP time aligning can compensate for the extra distance, diaphragm to MF taps, this morning being perfect indoor weather, I cracked on and modded the horn throat to be a perfect 1.5" / 38.1mm.
Will be fitting the spare screw on ring adapters, I turned / thread cut ages a
go..
Driver and Le Cléac'h horn (latter on hold for today, or longer?).
Trial pop on top
1" Ferrite BMS 4550 Vs 1.5" Alinco Vitavox S2
Can you tell which one I'm rooting for?😃
Gonna be interesting..
First one will be listenable this afternoon.
Will know soon enough.
Last edited:
All done and set up pretty well!
It's like staring into the abyss with low evening lights.
Back up in the right position - to my ears.
I can report, there's absolutely nothing wrong with the Synergy / SH50 concept, design etc. I just had to use a HF driver I love the sound of.
I could hear their potential, hence sticking with them and trailing
I enlisted the help of wifey’s ears too.
She thought she loved the S2s on Le Cléac’h, the latest hybrid Le Cléac’h/Synergys were a hit with her!
However, hearing the Le Cléac’h side verses the Synergy with S2 side, it was not so clear.
The Synergy side made the Le Cléac’h side sound a bit congested and even dare I use that word, coloured!
More horn somehow!
The Synergy side was clearer, cleaner, more revealing, accurate and crisper.
Still had that richer 1.5" throat, 3" diaphragm CD tone though. Neat trick!
Now with both sides dialed in I'm born away!
Man these things can boogie too!
Thanks for all the help guys!
It's like staring into the abyss with low evening lights.
Back up in the right position - to my ears.
I can report, there's absolutely nothing wrong with the Synergy / SH50 concept, design etc. I just had to use a HF driver I love the sound of.
I could hear their potential, hence sticking with them and trailing
I enlisted the help of wifey’s ears too.
She thought she loved the S2s on Le Cléac’h, the latest hybrid Le Cléac’h/Synergys were a hit with her!
However, hearing the Le Cléac’h side verses the Synergy with S2 side, it was not so clear.
The Synergy side made the Le Cléac’h side sound a bit congested and even dare I use that word, coloured!
More horn somehow!
The Synergy side was clearer, cleaner, more revealing, accurate and crisper.
Still had that richer 1.5" throat, 3" diaphragm CD tone though. Neat trick!
Now with both sides dialed in I'm born away!
Man these things can boogie too!
Thanks for all the help guys!
Cool.
And glad you're moving forward...finding what you like.
I know there's a camp that says all CD's sound the same once EQed the same.....but that's not my experience.
Lord knows, I work to get identical acoustic transfer functions for different CDs, ....on same horn / same speaker...and they do NOT sound the same.
I'm talking no smoothing mag and phase transfers, and polars that are super close. Identical measurements, as far as the usual criteria.
So I get it Steve
I've been away for a few days.
Not all bad to put some forced distance between brain and 'the sound'.
It has been quite an intensive time, from build and tuning, to modifying for the larger HF drivers and retuning etc.
Whilst I could hear the potential of the NO / few 1st order XO tuning method, I was not really comfortable with it.
Still learning the complex interaction between FR, driver alignment and resulting phase.
Spent some time listening to the my 'traditional'way of doing things - 24dB/oct slopes and phases aligned over the XO region.
This gives up its best sound, as you expect with the LF delayed in milliseconds, the distance from port (routed via baffles) to the HF diaphragm.
Measuring at 1m with acoustic absorbtion around between horn and mic.
The MF is also playing nicely like this. The delay being the distance from taps to HF diaphragm.
So far so good.
It sounds pretty good. Refined and hifi - is that 🙂 or 😧 though!
FFW to a minimalist XO approach (ala Chris, or as near as I can get)😃
Using a combo of 4uF cap and 6.8ohm resistor on the Vitavox S2s.. for some type of high pass XO..
No electrical XOs on the MF only PEQs, and 1st order high pass on the LF, PEQs where best to get flatish phase.
The MF is delayed as above / expected, but nothing I do and try can get the LF to play ball unless I have the MF and HF delayed more than the LF🥴
Forget the distance HF diaphragm to ports and everywhere around that. The phase is crazy out.
With the LF quite a bit ahead it plays ball, quite nicely!
I don't understand this, all help gratefully received.
Giving up at that, I volume levelled in DSP so I could compare the 2 setups - trad XOs vs Minimal XOs and phase levelled'ish.
The usual steep XO sound is detailed but quite thin in the MF. Not what the plots say..
The phase method is richer, disarmingly involving, but just sounds right - hard to describe. Less hifi more performance!👍
It's all that bit not fatiguing at all!
it's great to be able to switch been the two.
It's like having two systems in one!
Two toys, dog with two tails 😂
I know which I prefer!
I needed some time away to get this in perspective and re-energize.
ACDC Thunderstruck sounds amazing!
Will see what wife says with blind test on her favorite test tracks.
Not all bad to put some forced distance between brain and 'the sound'.
It has been quite an intensive time, from build and tuning, to modifying for the larger HF drivers and retuning etc.
Whilst I could hear the potential of the NO / few 1st order XO tuning method, I was not really comfortable with it.
Still learning the complex interaction between FR, driver alignment and resulting phase.
Spent some time listening to the my 'traditional'way of doing things - 24dB/oct slopes and phases aligned over the XO region.
This gives up its best sound, as you expect with the LF delayed in milliseconds, the distance from port (routed via baffles) to the HF diaphragm.
Measuring at 1m with acoustic absorbtion around between horn and mic.
The MF is also playing nicely like this. The delay being the distance from taps to HF diaphragm.
So far so good.
It sounds pretty good. Refined and hifi - is that 🙂 or 😧 though!
FFW to a minimalist XO approach (ala Chris, or as near as I can get)😃
Using a combo of 4uF cap and 6.8ohm resistor on the Vitavox S2s.. for some type of high pass XO..
No electrical XOs on the MF only PEQs, and 1st order high pass on the LF, PEQs where best to get flatish phase.
The MF is delayed as above / expected, but nothing I do and try can get the LF to play ball unless I have the MF and HF delayed more than the LF🥴
Forget the distance HF diaphragm to ports and everywhere around that. The phase is crazy out.
With the LF quite a bit ahead it plays ball, quite nicely!
I don't understand this, all help gratefully received.
Giving up at that, I volume levelled in DSP so I could compare the 2 setups - trad XOs vs Minimal XOs and phase levelled'ish.
The usual steep XO sound is detailed but quite thin in the MF. Not what the plots say..
The phase method is richer, disarmingly involving, but just sounds right - hard to describe. Less hifi more performance!👍
It's all that bit not fatiguing at all!
it's great to be able to switch been the two.
It's like having two systems in one!
Two toys, dog with two tails 😂
I know which I prefer!
I needed some time away to get this in perspective and re-energize.
ACDC Thunderstruck sounds amazing!
Will see what wife says with blind test on her favorite test tracks.
From your description "LF delayed in milliseconds" it sounds like you have delayed the LF by too much to be in phase on the same cycle as the mid driver, as there should be no more than a single millisecond time of flight difference between the two, even looking at the arrival time of frequencies above the desired pass-band of either.
Ah yes, my mistake.
I meant delay samples in my Najda DSP unit, not milliseconds.
It's running at 96KHz. I don't have the conversion for delay samples to mm to hand right now.
I've tried every delay in very small increments from the same point as the HF (just to see) to what would be reasonable beyond the MF point to the LF.
Can't get no satisfaction!🙂
Just as a reference point, in the two-way MEH cabinets I built, the HF had to be delayed a fraction of a millisecond to line up it's phase response with the LF after using EQ and 24dB/octave electrical filters.
With no crossovers or processing at all, the combined raw response of the HF/LF phase varied by only ~50 degrees in the acoustical crossover region:
50 degrees is nothing compared to the ~180 degree phase inversion of the bass reflex port output
With no crossovers or processing at all, the combined raw response of the HF/LF phase varied by only ~50 degrees in the acoustical crossover region:
50 degrees is nothing compared to the ~180 degree phase inversion of the bass reflex port output
Did some playing around this afternoon.
As per advice above👍, I took off all the XOs (bar on the comp driver), removed all PEQs and ran the Freq / Phase.
This is with the time delays that work best, blue plot.. the LF being earlier than the others!🥴
Red plot was with the MF out of phase, just to see..
I then tried the LF where you'd expect it to be, about 30-40cm behind the rest.
Only ever got this.. and worse
Tried less delay and more delay, way in excess of what's reasonable, in fine resolution.
The only thing that gets me on the same graph (not wrapping around), is with the LF earlier as previous post above.
With the LF some 46cm in advance of the HF, I'm getting near - phase plot not wrapping around.
I tried reasoning that the phase shift and physical distance of the bass reflex part was the cause - can't be surely?🤯
I'm used to long delays with the tapped horns.
Only using Holmimpulse feels like I'm rather in the dark..
Time to bite the bullet and get/learn REW?!
As per advice above👍, I took off all the XOs (bar on the comp driver), removed all PEQs and ran the Freq / Phase.
This is with the time delays that work best, blue plot.. the LF being earlier than the others!🥴
Red plot was with the MF out of phase, just to see..
I then tried the LF where you'd expect it to be, about 30-40cm behind the rest.
Only ever got this.. and worse
Tried less delay and more delay, way in excess of what's reasonable, in fine resolution.
The only thing that gets me on the same graph (not wrapping around), is with the LF earlier as previous post above.
With the LF some 46cm in advance of the HF, I'm getting near - phase plot not wrapping around.
I tried reasoning that the phase shift and physical distance of the bass reflex part was the cause - can't be surely?🤯
I'm used to long delays with the tapped horns.
Only using Holmimpulse feels like I'm rather in the dark..
Time to bite the bullet and get/learn REW?!
Last edited:
Yep, REW is a truly wonderful resource...both the program itself, and the manual/help files.
I can do so much, with so many options, it's quite daunting at times. (to me anyway)
I'd recommend you also try the donateware, dual-channel OpenSoundMeter https://opensoundmeter.com/en/
Real-time dual channel let's you see the mag and phase change on screen as you adjust processing filters, xovers, delays, polarites, etc.
Which makes it great for IIR work.
And Steve, pls know I haven't jumped in to try to help with the issues you've been experiencing....simply because I don't use either of the two routes you are playing with....
I've got my linear-phase routine down to where it is so simple, quick, and excellent performing ...that I just can't get into messing with IIR (no matter the approach.)
Just saying all this so you know I'm not ignoring your efforts.
Keep up the fine work! , and enjoy!
Thanks Mark!
One day I will get into the 21st Century 😂
I'm liking how far I've got with it on IIR though.
I test listened with no PEQs paring the drivers in to hi pass, lo pass and band pass.
Only applied a few to the appropriate spots looking at the whole plot of all 3 to get a decent level FR. Not many required really. Mostly tweaks to the MF.
It was remarkably good actually!
Not quite as pure / clean as the controlled slopes setup but on much material it's quite acceptable - this bodes well for the PA / outdoor use I have them in mind for further down the road.
A Behringer DCX2496 and 3 PA amps will do pretty well I hope, without the Nth degree hifi-neurosis in room set up..
Solved the LF needing to be displaced so far ahead in time, well at least I have a setting where it's where your expect it!
Phase switch on the LF and it's right where it should be by my calculations.
Silly thing is I already had a setup from the 15th of March like that, but lost my way somewhere in too many iterations and not enough Configuration Management🙂
Sounds right too! 🎶
Phase switch on the LF and it's right where it should be by my calculations.
Silly thing is I already had a setup from the 15th of March like that, but lost my way somewhere in too many iterations and not enough Configuration Management🙂
Sounds right too! 🎶
Last edited:
I've been throwing everything at this setup this evening.
It just sounds so good!
As if everything has snapped into focus!
The potential I has been realised.
All lower vocals are smooth and balanced, stunning insight into bass and mid bass detail. I could go on but I won't🙂
Thanks to @weltersys and @SpeakerBob for some wise insight that got me thinking along the right lines👍
It just sounds so good!
As if everything has snapped into focus!
The potential I has been realised.
All lower vocals are smooth and balanced, stunning insight into bass and mid bass detail. I could go on but I won't🙂
Thanks to @weltersys and @SpeakerBob for some wise insight that got me thinking along the right lines👍