Hi folks developing a synergy / meh / syntripp horn
Can dbx driverack or the Berhinger units dsp 8000/8024 ultracurve or deq2496 or minidsp 2x4 produce brick wall filters,
Due to the drivers been loaded by the horn not really bothered about phase/lobbing/nulls etc
So much will be acoustic filter. Port holes fruscom entry
At this moment in time for experimental purposes and learning wood angle cutting techniques the compression driver is celestion cdx1-1745 1” compression driver mated with one celestion 10” midrange tf1020 I can double these up as I have 4 but going to try a single first, possibly with some form of compression phase plug with very small back chamber.
Crossover point if my measurements are right aimed at around 1000hz
Low end at present will be covered by jbl 4638 cab (twin 15”) jbl 2035 drivers
This for home use so taking a bit of stick now n again and not prolonged blasting
Cheers
Can dbx driverack or the Berhinger units dsp 8000/8024 ultracurve or deq2496 or minidsp 2x4 produce brick wall filters,
Due to the drivers been loaded by the horn not really bothered about phase/lobbing/nulls etc
So much will be acoustic filter. Port holes fruscom entry
At this moment in time for experimental purposes and learning wood angle cutting techniques the compression driver is celestion cdx1-1745 1” compression driver mated with one celestion 10” midrange tf1020 I can double these up as I have 4 but going to try a single first, possibly with some form of compression phase plug with very small back chamber.
Crossover point if my measurements are right aimed at around 1000hz
Low end at present will be covered by jbl 4638 cab (twin 15”) jbl 2035 drivers
This for home use so taking a bit of stick now n again and not prolonged blasting
Cheers
Last edited:
About every 4-6 months someone comes in wanting to use DSP to avoid basic crossover design.
Stick to 1st through 4th order filters. Remember that the phase and amplitude matching is electro-acoustical, not purely electrical. Your crossover slopes and phase adds to the driver's to produce the final result.
Best,
E
Stick to 1st through 4th order filters. Remember that the phase and amplitude matching is electro-acoustical, not purely electrical. Your crossover slopes and phase adds to the driver's to produce the final result.
Best,
E
Thanks eriksquires but I have been using minidsp 2x4 on and off for about 3 years with 3 qsc 1450 amplifiers (biamped)
The goal of synergy multi entry horns relys on brick wall filters hence the original question.
What’s wrong with going dsp vs passive in your eyes.
Basic crossover design as you refer to is not simple in an synergy/unity/meh design thanks
Thanks Scott for the reference regards minidsp.
But the question still stands regards berhinger units and dbx driverack.. amended
Cheers
I personally have no strong preference besides physical complexity on the issue of passive vs. active filtering. I use a miniDSP on my sub and it works really well.
If there are benefits to high order filters in DSP land I am not aware of them. There's no filter I know of that is binary at a particular F, but there certainly have been some commercial applications using "Infinite slope" passive filters.
Best,
E
If there are benefits to high order filters in DSP land I am not aware of them. There's no filter I know of that is binary at a particular F, but there certainly have been some commercial applications using "Infinite slope" passive filters.
Best,
E
With miniDSP it is possible to realize elliptical filters. I already made 3rd order and 5th order designs.
See an example in the plots of a 5th order design of a 2 way system crossed at 1750Hz, the targets and the practical realized design.
The 5th order filters have a fall off in the stopband of 80dB/octave and the outband reduction is -55dB.
I do the design in Leap CrossoverShop with analog functional blocks and via a bilinear transform the coefficients of the biquads for miniDSP are calculated.
If no SPL equalizing for the drivers is needed, 3 biquads are needed to realize the basic elliptical filtering for a 5th order filter.
IMO elliptical filters sound very nice, but you have to make the design very accurate. Together with a friend I realized a 5th order 3 way design with very good sounding results.
So, if 5th order elliptical is brickwall enough, it is possible to realize in miniDSP.
Even a higher order filter is possible, but then the groupdelay will peak more at the crossover frequency, even too much.
See an example in the plots of a 5th order design of a 2 way system crossed at 1750Hz, the targets and the practical realized design.
The 5th order filters have a fall off in the stopband of 80dB/octave and the outband reduction is -55dB.
I do the design in Leap CrossoverShop with analog functional blocks and via a bilinear transform the coefficients of the biquads for miniDSP are calculated.
If no SPL equalizing for the drivers is needed, 3 biquads are needed to realize the basic elliptical filtering for a 5th order filter.
IMO elliptical filters sound very nice, but you have to make the design very accurate. Together with a friend I realized a 5th order 3 way design with very good sounding results.
So, if 5th order elliptical is brickwall enough, it is possible to realize in miniDSP.
Even a higher order filter is possible, but then the groupdelay will peak more at the crossover frequency, even too much.
Attachments
You can implement elliptical filters (or quasi-elliptical anyway) with passive filters too. 'Just' insert a pole zero notch in the stopband to accelerate the initial rolloff rate to your target acoustic transfer function.
Indeed, 3rd order elliptical passive filtering is easy to realize. I already made several such passive 3rd order filters. 5th order is possible too, but it becomes rather complex.
The behringer dcx2496 goes to 48dB/oct. I've no experience with the ultacurve units but I thought they were EQ only, no x-over...?
I am a huge fan of very steep crossovers,
but with caveats eriksquires mentioned,....it still takes work despite the steep slopes....
to get summed electro-acoustical phase and amplitude right.
If done correctly, steep slopes allow superior off-axis consistency IME.
Also a huge fan of FIR for the implementation. I'd recommend mini-DSP's 2x4HD.
For a crossover at 1000Hz, it easily has enough taps to brickwall.
I am a huge fan of very steep crossovers,
but with caveats eriksquires mentioned,....it still takes work despite the steep slopes....
to get summed electro-acoustical phase and amplitude right.
If done correctly, steep slopes allow superior off-axis consistency IME.
Also a huge fan of FIR for the implementation. I'd recommend mini-DSP's 2x4HD.
For a crossover at 1000Hz, it easily has enough taps to brickwall.
Why do you think you might want to use a brick wall filter?Straight over top of my head fast Fourier transform filters, above my pay grade
It is straightforward and possibly for no cost depending on your hardware to use a computer to create a multichannel file of a song where the channels are the brick wall filtered inputs for the drivers. You can then use this to hear why people don't use brick wall filters for speaker crossovers.
I don't think that works.
Maybe, it would sound the same, as a brickwall x-over......if the speaker had near perfect mag, phase, and time alignment....dunno, would be interesting to try.
But I feel pretty sure it won't replicate a physical brickwall crossover without those conditions in place.
Can you say why?
Read an audio file from disk. FFT the file. For each channel set all the frequency components to zero above or below the particular crossover frequency. Inverse FFT each channel to get the time signal for each channel. Create a multichannel file from each component channel plus zeros for the extra channels that are not used. Write to disk. Probably about 10 lines in most scripting languages with audio support.
Ok, I'll try.
Consider the speaker you are sending these split signals to.
It has a crossover in place, I presume...?
Let's just take a 2-way crossed at 1000Hz, with a typical low order slope.
Each side of the crossover is expecting a full range signal, for it to pass on its contribution to summing, above and below xover.. These expected contributions extend past 1000Hz for the low driver, and below 1000Hz for the high driver. A synthetic brick wall (if I correctly understand what you are doing) will truncate the signals the crossovers need to fullfill the expected above and below contributions.
Wouldn't you end up with a notch in the summed freq response at xover, that should be nearly the same as a well aligned polarity inversion?
OK, to solve that, say you take out the crossover. You then need, at a minimum, to align the levels, time and phase, of the non-crossed-over drivers. Who's ever done that?
Maybe, if a perfectly aligned speaker also had a brick wall xover in place, at the same freq used for the synthetic...then it might work 😉
Just guessin it might though 🙂
I think it is better to create brickwall targets for the X-over filter.
Then making a digital filter design inclusive the acoustical responses, in a way that the combined electrical-acoustical response is mapping on the targets.
Adding notches to an existing filter design needs an investigation of the sum at first.
Then making a digital filter design inclusive the acoustical responses, in a way that the combined electrical-acoustical response is mapping on the targets.
Adding notches to an existing filter design needs an investigation of the sum at first.
Paul, this really interesting to me. I've imbedded elliptical xovers into FIR files, but never thought about them in IIR biquads.
Some of what I do is live sound, and I've been looking for a way to achieve steep filters without the FIR latency..
This might help do that , thx.
Mark, indeed it is the most easy way to use IIR biquads to make the sum.
And if you prefer to make a phase linear system, you better do the phase correction on the sum response.
As I said, an elliptical design needs to be very accurate, that is my experience.
It is also very sensitive while tuning.
I started using elliptical filtering after hearing the Magico speakers.
They also use elliptical filtering in most of their designs.
??? The OP has a set of drivers in his speakers for which he wants to use brick-wall filters. His setup is active. The above enables him to simply bypass his current active crossover, if he has bought one yet, and check how the performance of brick wall filters are as crossovers. We could tell him why they are bad but this simple no cost test will enable him to hear.
- Status
- Not open for further replies.