I removed the Mid bass band pass X/Os in Najda and set my Crown XLS power amp up do them. It only has 4th order slopes.
I spent a while fine tuning the frequency points and evaluating the X/O in the Crown.
Not bad, and I did run a 4th order slope through the 15" driver mid bass horns.
When I changed to the 12", I settled on 2nd order.
The 12" has a better upper freq sound and is more pleasing on 2nd order.
It also goes just as deep with punch due to the longer horn / compressing radio.
So back with 2nd order, the original freq points, Najda doing the job.
The XTI range of Crown's offerings do indeed do 1st, 2nd, 3rd and 4th order.
The XLS I have, only does 4th.
As long as Najda is around no need.
Useful succession planning🙂
I spent a while fine tuning the frequency points and evaluating the X/O in the Crown.
Not bad, and I did run a 4th order slope through the 15" driver mid bass horns.
When I changed to the 12", I settled on 2nd order.
The 12" has a better upper freq sound and is more pleasing on 2nd order.
It also goes just as deep with punch due to the longer horn / compressing radio.
So back with 2nd order, the original freq points, Najda doing the job.
The XTI range of Crown's offerings do indeed do 1st, 2nd, 3rd and 4th order.
The XLS I have, only does 4th.
As long as Najda is around no need.
Useful succession planning🙂
For the Tapped horn subs the
Crown XTI series could be candidates.
Plenty of PEQ possible.
I only cut at 29Hz and a little at 55Hz = room modes.
Damping factor > 500 @ 20 - 400Hz ✔️
And all the XO slopes, 1st to 4th order available.
If the Cerwin Vega ever gives up an XTI could the next.
Anyone know if there are beefy amps with DSP logic that might have FIR filters instead of IIR / PEQ?
Many taps at low frequencies I understand and demands huge processing power = cost.
Crown XTI series could be candidates.
Plenty of PEQ possible.
I only cut at 29Hz and a little at 55Hz = room modes.
Damping factor > 500 @ 20 - 400Hz ✔️
And all the XO slopes, 1st to 4th order available.
If the Cerwin Vega ever gives up an XTI could the next.
Anyone know if there are beefy amps with DSP logic that might have FIR filters instead of IIR / PEQ?
Many taps at low frequencies I understand and demands huge processing power = cost.
Last edited:
Thought I'd try to physically time align the upper channels and see where the drivers / horns end up.
Tweeters to upper mids are already physically time-aligned.
Set all the time delays to 0, ignoring the position of mid bass to tapped horn, I could not physically move the mid horn far enough back or conversely the mid bass horn near enough to align
Some 40cm out!
The diaphragm of the mid horn would end up way behind that of the mid bass horn .
Checked everything several times.
Even wondered if the Crown amp with some DSP in it (not in use) was delaying things..?
Tried a simple big topping class D instead.
Nope. Same.
Any ideas?
I moved the mid / upper mid + tweeters quite a long way back and then time aligned in Najda as usual.
With those horns farther away it's a more polite sound.
Interesting.
Not had it like this since 2011!
Tweeters to upper mids are already physically time-aligned.
Set all the time delays to 0, ignoring the position of mid bass to tapped horn, I could not physically move the mid horn far enough back or conversely the mid bass horn near enough to align
Some 40cm out!
The diaphragm of the mid horn would end up way behind that of the mid bass horn .
Checked everything several times.
Even wondered if the Crown amp with some DSP in it (not in use) was delaying things..?
Tried a simple big topping class D instead.
Nope. Same.
Any ideas?
I moved the mid / upper mid + tweeters quite a long way back and then time aligned in Najda as usual.
With those horns farther away it's a more polite sound.
Interesting.
Not had it like this since 2011!
I suspect there is possibly a few effects at play here influencing the "polite" sound as you describe it.
Firstly, the phase delay caused by IIR filters creates a "high frequency first" bias, where the first part of the impulse response is predominantly high frequency energy, followed by mid, followed by low etc (I believe you already know this).
I recently tried messing around with very small delay changes on the HF of a synergy horn I have been designing, and found that differences as small as a third of a millisecond for a 1khz XO can (subjectively speaking) create a much more relaxed sound that I can listen to for far greater periods of time it's possible this setup simply has a "better" time alignment than before (I don't know the accuracy available for your DSP).
Secondly (and forgive me if I've missed something, I have followed this thread for some years now but have not read through it in a while), assuming you have not accounted for the inverse square law by simply placing the higher frequency elements further back you will lower the spl from each band at the listening position, biasing the low frequency bands which in my experience can also create a "polite" or warm sound to things.
Thirdly, by placing the HF sources further back you will increase the diffuse sound relative to the direct sound due to the side walls being "sprayed" a little sooner. This may appear to have the effect of taking the edge off of transients and thus introduce a "polite" sound.
Firstly, the phase delay caused by IIR filters creates a "high frequency first" bias, where the first part of the impulse response is predominantly high frequency energy, followed by mid, followed by low etc (I believe you already know this).
I recently tried messing around with very small delay changes on the HF of a synergy horn I have been designing, and found that differences as small as a third of a millisecond for a 1khz XO can (subjectively speaking) create a much more relaxed sound that I can listen to for far greater periods of time it's possible this setup simply has a "better" time alignment than before (I don't know the accuracy available for your DSP).
Secondly (and forgive me if I've missed something, I have followed this thread for some years now but have not read through it in a while), assuming you have not accounted for the inverse square law by simply placing the higher frequency elements further back you will lower the spl from each band at the listening position, biasing the low frequency bands which in my experience can also create a "polite" or warm sound to things.
Thirdly, by placing the HF sources further back you will increase the diffuse sound relative to the direct sound due to the side walls being "sprayed" a little sooner. This may appear to have the effect of taking the edge off of transients and thus introduce a "polite" sound.
Last edited:
Yes, I think you are spot on with the softer sound analysis, coming from the tweeters, upper mids & mids, now farther away.
It's easy for me to move them forward a bit and retime, until the sweet spot is found.
I do think my ears (brain) gets bored and a change in presentation is refreshing from time to time!
I remember physically time aligning the 4 channels (not including Tapped horns, when I had the 4 drivers all in a vertical stack.
Back then I had the classic convex shape with the mid bass at the bottom and the mids at the top nearer, then the upper mids and tweeters.
Can't get anywhere near that at the moment.
Perhaps the mids being so high were further away from me when sitting down (measuring mic), than I thought?
I only have PEQ filters on the tapped horns.
Have 3rd order on the high pass of the mid bass, 2nd order on the low pass.
The mids are 2nd order bandpass.
The upper and tweeters on 1st order high pass only.
Before, (12 years ago), when physically time aligning all were 2nd order passives..
I always measure 1st peak positive in Holmimpulse.
All help and insights much appreciated.
It's easy for me to move them forward a bit and retime, until the sweet spot is found.
I do think my ears (brain) gets bored and a change in presentation is refreshing from time to time!
I remember physically time aligning the 4 channels (not including Tapped horns, when I had the 4 drivers all in a vertical stack.
Back then I had the classic convex shape with the mid bass at the bottom and the mids at the top nearer, then the upper mids and tweeters.
Can't get anywhere near that at the moment.
Perhaps the mids being so high were further away from me when sitting down (measuring mic), than I thought?
I only have PEQ filters on the tapped horns.
Have 3rd order on the high pass of the mid bass, 2nd order on the low pass.
The mids are 2nd order bandpass.
The upper and tweeters on 1st order high pass only.
Before, (12 years ago), when physically time aligning all were 2nd order passives..
I always measure 1st peak positive in Holmimpulse.
All help and insights much appreciated.
Crown I-Tech HD's have the most FIR capability I've seen in a beefy amp.
Although I haven't canvassed the market lately, mainly because I don't think it makes sense to bundle a great deal of FIR capability in a prosound amp, given the typical inability of prosound to tolerate FIR delay.
Since you have such a high count multi-way system, I strongly believe you are one of the best candidates to take advantage of FIR xovers.
Here's a snip from a Crown I-Tech manual...https://www.crownaudio.com/en/site_elements/i-tech-4x3500hd-manual-english
Not only that for helping out a single xover region.
But also helps In a high-count multi-way, because how on earth can you keep low order xovers from spilling into sections past the two primary sections used with the xover?
In my experiences with 4, 5, and now 6-way builds, steep complementary linear phase xovers simply work.
And are much much easier to implement than any order IIR xovers. Ridiculously easier...(if the FIR delay is no problem.)
Cheapest, and perhaps best way to start experimenting, is to simply make lin phase xovers with rePhase, and put them into JRiver.
And also add min-phase EQs into the same rePhase xover files, for driver corrections.
Jriver, have licence. It's been about 5 years since I last used it!
I don't have a PC based front end to do the rest of the system justice, but as you say a viable experimentation bed, and then see where it goes..
Interesting.
I assume Laptop USB output into a Reclocker (Gustard U18 in this case) would work?
I don't have a PC based front end to do the rest of the system justice, but as you say a viable experimentation bed, and then see where it goes..
Interesting.
I assume Laptop USB output into a Reclocker (Gustard U18 in this case) would work?
You can put a FIR file for each driver section in JRiver, such that you really don't need a PC for a front end.
Just need a plain old multi-channel DAC to interface with JRiver.
What do you use a reclocker for? I have zero experience with such.
The Crown manual I linked gives examples, showing how the width the lobing range shrinks down to a narrow sharp dip with steep filters.
The narrow sharp dip occurs with either low order or high order xovers, so it's all gain/no lose with high order (other than delay)
The only thing misleading about their example imo, is that they use a full polarity reversal for the example
Otherwise, it's pretty much simple fact steep xovers narrow the frequency range of lobing...
It's a Gustard U18 reclocker.
Does a great job improving the sound of Sonore Ultrarendu renderer.
I assume with Jriver it would run on my Windows laptop as before.
I used to stream flac files from HDD back in the day through WaveIO and I2S into Najda DAC/DSP.
I could now use DLNA, Bubble UPnP app on my android device with Qobuz as I do now.
Makes Ultrarendu redundant 🙂
It's worth experimenting with. I'll have to get my head around it.
I've moved the mid / Upper mid / tweeter drivers to where they sound best time aligned - took a while with iterations.
Then listened with it all time aligned.
Then removed all alignment and switched between the two setups in Najda.
I've done this before years ago.
I'd forgotten the profound effect of totally aligned vs not!
On simpler music it's almost fine un-aligned.
Complex stuff mostly sounds a bit boxed in and disorganised.
Instrument placement is odd. Mid bass has less punch / tone. Vocals are positioned differently and almost of different character on some tracks.
It's like un-aligned, the mixing / master production was not done quite right 🙂 Not that of course, but it's the effect.
Time-alignment is definitely necessary 👍
Does a great job improving the sound of Sonore Ultrarendu renderer.
I assume with Jriver it would run on my Windows laptop as before.
I used to stream flac files from HDD back in the day through WaveIO and I2S into Najda DAC/DSP.
I could now use DLNA, Bubble UPnP app on my android device with Qobuz as I do now.
Makes Ultrarendu redundant 🙂
It's worth experimenting with. I'll have to get my head around it.
I've moved the mid / Upper mid / tweeter drivers to where they sound best time aligned - took a while with iterations.
Then listened with it all time aligned.
Then removed all alignment and switched between the two setups in Najda.
I've done this before years ago.
I'd forgotten the profound effect of totally aligned vs not!
On simpler music it's almost fine un-aligned.
Complex stuff mostly sounds a bit boxed in and disorganised.
Instrument placement is odd. Mid bass has less punch / tone. Vocals are positioned differently and almost of different character on some tracks.
It's like un-aligned, the mixing / master production was not done quite right 🙂 Not that of course, but it's the effect.
Time-alignment is definitely necessary 👍
Last edited:
I was searching and found some interesting old threads about JBL 2482 phenolic drivers.
Consensus / DIYaudio wisdom was that lowest freq would be 300Hz regardless of horn size/cutoff freq.
Full power was quoted at 300Hz however!!
I use a pair of Autotech composite T-200 horns - the only commercial ones I've bought and still use.
I've been using them on a 2nd order L-R high pass at 330 or 350Hz.
My reading peaked my interest, so I just measured my 2482s on the T-200 horns.
First 1m from horn, offset from centre.
Then from listening position some 4.5m or so, away.
On a 1st order, at normal to loud listening volumes the lowest I can get is 290Hz.
Lowering the freq in Najda DSP had no further effect as I deduce the driver / horn cut-off was dictating.
I do wonder if I could eek out some more on the T-160 horns.
They are 85cm long, so would fit okay.
I low pass cross the 2482s at 750Hz, so reckon the T-160 would do the upper frequency just fine?!
I quite fancy making a 4 or 6 or 8 sided pair of Tractrix 160Hz horns out of laminated up 4mm birch plywood. Mouth rounding would be done by thicker ply - like I did all those years ago with my conicals..
I think I would use plywood / bitumen sheet /plywood, as many layers as necessary for good damping.
Hmm, could be a fun project.
Am I "ut och cyklar" as Swedes say?
=
"Half a bubble off centre" in US speak 🙂
Consensus / DIYaudio wisdom was that lowest freq would be 300Hz regardless of horn size/cutoff freq.
Full power was quoted at 300Hz however!!
I use a pair of Autotech composite T-200 horns - the only commercial ones I've bought and still use.
I've been using them on a 2nd order L-R high pass at 330 or 350Hz.
My reading peaked my interest, so I just measured my 2482s on the T-200 horns.
First 1m from horn, offset from centre.
Then from listening position some 4.5m or so, away.
On a 1st order, at normal to loud listening volumes the lowest I can get is 290Hz.
Lowering the freq in Najda DSP had no further effect as I deduce the driver / horn cut-off was dictating.
I do wonder if I could eek out some more on the T-160 horns.
They are 85cm long, so would fit okay.
I low pass cross the 2482s at 750Hz, so reckon the T-160 would do the upper frequency just fine?!
I quite fancy making a 4 or 6 or 8 sided pair of Tractrix 160Hz horns out of laminated up 4mm birch plywood. Mouth rounding would be done by thicker ply - like I did all those years ago with my conicals..
I think I would use plywood / bitumen sheet /plywood, as many layers as necessary for good damping.
Hmm, could be a fun project.
Am I "ut och cyklar" as Swedes say?
=
"Half a bubble off centre" in US speak 🙂
I got sidetracked and completely forgot about the above..
Much more interesting has been making a 2nd order X/O for the Raal Lazy Ribbon tweeters.
Then there was moving further away from the mid bass to get more punch! Worked 👍
Then I leaned from @hornydude about time aligning the mouths relative to horn length Inc driver distance to diaphragms.
This was a revelation.
Much preferable to usual alignment diaphragm to ear method.
Then learning about the interaction and interference of room and horns relative to one another.
This spurred me to physically align all the mouths.
Basically as they would be mounted in a speaker front panel.
This meant bringing the mids, upper mids and tweeters nearer to listening spot by 40-50cm or more.
A bit of carpentry required for the supports.
The tapped horn subs and mid bass horns stay where they are..
I then retimed the drivers as usual.
A lot more delay required on the drivers I'd moved nearer of course.
I then applied the horn length algorithm (have created a spreadsheet so it's quick to get the new delay sample values for Najda), so the sound leaves the mouths at the same time.
This last critical step is the difference from my previous trials with physically aligned mouths, some 8 or 9 years ago!
I can now flick between usual alignment and the new way, but this time with nearer field mids/upper mids and tweeters.
Initial listening
- usual time alignment, help get me out of here and back to the old way.
Sounds harsh and hard.
- new horn length mouth exit alignment. Hmm that just sounds right, so uninterrupted and distortion free.
Lots of evaluation required, but marrying these two ideologies (aligned nearer field upper frequencies / horn length exit delay alignment) seems to bring both of their benefits!
The sofa being far away is great for bass punch, bringing the drivers right to the front of the mid bass horns essentially brings them near(er) field again.
Also interesting moving around the room - there are little or no variations in sound?! It's very uniform.
A picture worth a thousand words..
Much more interesting has been making a 2nd order X/O for the Raal Lazy Ribbon tweeters.
Then there was moving further away from the mid bass to get more punch! Worked 👍
Then I leaned from @hornydude about time aligning the mouths relative to horn length Inc driver distance to diaphragms.
This was a revelation.
Much preferable to usual alignment diaphragm to ear method.
Then learning about the interaction and interference of room and horns relative to one another.
This spurred me to physically align all the mouths.
Basically as they would be mounted in a speaker front panel.
This meant bringing the mids, upper mids and tweeters nearer to listening spot by 40-50cm or more.
A bit of carpentry required for the supports.
The tapped horn subs and mid bass horns stay where they are..
I then retimed the drivers as usual.
A lot more delay required on the drivers I'd moved nearer of course.
I then applied the horn length algorithm (have created a spreadsheet so it's quick to get the new delay sample values for Najda), so the sound leaves the mouths at the same time.
This last critical step is the difference from my previous trials with physically aligned mouths, some 8 or 9 years ago!
I can now flick between usual alignment and the new way, but this time with nearer field mids/upper mids and tweeters.
Initial listening
- usual time alignment, help get me out of here and back to the old way.
Sounds harsh and hard.
- new horn length mouth exit alignment. Hmm that just sounds right, so uninterrupted and distortion free.
Lots of evaluation required, but marrying these two ideologies (aligned nearer field upper frequencies / horn length exit delay alignment) seems to bring both of their benefits!
The sofa being far away is great for bass punch, bringing the drivers right to the front of the mid bass horns essentially brings them near(er) field again.
Also interesting moving around the room - there are little or no variations in sound?! It's very uniform.
A picture worth a thousand words..
Last edited:
Hello Steve.
Great to read how you have managed to improve your sound further.
I was wondering about the dimensions of your midbass horn, and your xo frequencies? It is my understanding it's the largest horn seen in the picture, while the sub is elsewhere. I recall you were driving the midbass with a 12".
Great to read how you have managed to improve your sound further.
I was wondering about the dimensions of your midbass horn, and your xo frequencies? It is my understanding it's the largest horn seen in the picture, while the sub is elsewhere. I recall you were driving the midbass with a 12".
Hi LewinskiH01,
Yes I have two tapped horns.
The mid bass horns (seen as the largest round Exponentials in the pic), are on 2nd order L-R X/Os.
Cross at 100Hz and 300Hz.
The mouths are 12cm diameter.
Horn length just over 1m. Mouth is 66cm diameter.
F12 in Hornresp is 89.43.
Don't require any EQ.
These horns were originally shorter with a 15" driver in.
Something I tried - wanted a bit more compact vs the original huge 12 sided conical horns I had in the beginning.
I made a 20cm throat adapter to extend and get the same lower frequency from the 12" as the 15" drivers.
The longer horn also gave more punch and better higher frequency tone.
Yes I have two tapped horns.
The mid bass horns (seen as the largest round Exponentials in the pic), are on 2nd order L-R X/Os.
Cross at 100Hz and 300Hz.
The mouths are 12cm diameter.
Horn length just over 1m. Mouth is 66cm diameter.
F12 in Hornresp is 89.43.
Don't require any EQ.
These horns were originally shorter with a 15" driver in.
Something I tried - wanted a bit more compact vs the original huge 12 sided conical horns I had in the beginning.
I made a 20cm throat adapter to extend and get the same lower frequency from the 12" as the 15" drivers.
The longer horn also gave more punch and better higher frequency tone.