eq8 is a parametric EQ extension for Firefox, Chrome
https://addons.mozilla.org/en-US/firefox/addon/eq8/
https://chromewebstore.google.com/detail/eq8/bmbhmpfeabllbfcalamkbdknkaijpkja?pli=1
https://addons.mozilla.org/en-US/firefox/addon/eq8/
https://chromewebstore.google.com/detail/eq8/bmbhmpfeabllbfcalamkbdknkaijpkja?pli=1
A linkwitz transform is much easier and only needs one biquad.
You can in fact get any Q you want with any f3 you want.
Technically this also works with BR systems, but since the Cms and BL aren't linear, it's a very difficult to get something that also works at higher levels.
Because it will shift the parameters.
The same happens in closed box systems, but since they have a lower order system response, the error is smaller.
I am not sure how much someone can write on this stuff?
Since it's just basic response system manipulation.
So once you picked a certain Q and fs with LT, it's just a matter of combinations with other filters.
Downside is an higher group delay.
In fact the response will be identical to a BR system, only without the cone excursion dip.
So the real benefit is a bit gone.
These days it's just easier to create a target curve and let VituixCAD just find the right parameters for the biquad block.
Technically a linkwitz transform for low pass is also possible.
But I will ask for the paper, you never know if there is something in it
Since it's just basic response system manipulation.
So once you picked a certain Q and fs with LT, it's just a matter of combinations with other filters.
Downside is an higher group delay.
In fact the response will be identical to a BR system, only without the cone excursion dip.
So the real benefit is a bit gone.
These days it's just easier to create a target curve and let VituixCAD just find the right parameters for the biquad block.
Technically a linkwitz transform for low pass is also possible.
But I will ask for the paper, you never know if there is something in it
Yeppa, , and other than whatever driver is handling the VHF to above audibility, they are all low-pass filters too.
That's what I've been trying for the last year. Encouraged so far.
The Crosslite+ software i mentioned has nearly 50 window profiles to try, to help get the most out of tap availability.
It's target building capability allows any combo of any type filter, min, lin, or max phase. Or
match any curve such as another speakers acoustic output, or imbed a chosen equal loudness curve, etc.
DSP rocks, and it's only getting better !
Thx for that head start. Will study.
How to with a BR? With the port as a separate source, and an EQ bank that actuates both woofer and port?
You can just make any target curve you want and change the parameters accordingly?
Sorry, I am not sure what you're asking for?
EQ-ing vented-assisted means that there is both a Helmholtz resonator and a cone involved. In non-assisted BR systems, we usually only insert the combined port and woofer response. This however cannot be a basis for the optimizer function, because it occludes the separate radiators port and cone to VCad. So, would we then need to start with separate port and woofer measurements, include distance of port to woofer as well as orientation for phase, and then let optimizer do its thing? What is your approach?
I have the feeling that you maybe took a little shortcut here (I am often guilty of that too). It should be mentioned that it is indeed possible to equalise the final response to any Q and f3 but it can't always be done with the famous circuit published by Linkwitz. For some combinations of original and target parameters some more complex circuits have to be used.
I wanted to try to equalise a given reflex tuning to a 5th order Bessel one by using something like two Linkwitz transforms and a first order highpass but haven't found the time for it yet.
Regards
Charles
Are you suggesting a fourth order vented combined response is different to an equivalent IIR filter, I would tend to disagree?
You could achieve something quite similar by summing the outputs of a state-variable filter accordingly.
Maybe it is sufficient to sum just the highpass and lowpass outputs. If you have a filter Q of 2 and use the highpass with a weight of one and the lowpass with a weight of 0.2 you would end up with a shelf that has peak around 6 dB, followed by a notch slightly lower. If you need just a dip instead of a notch you can also add some part of the bandpass output. If the notch variant is sufficient you can just sum the output of ordinary sallen-key lowpass and highpass filters accordingly.
I don't have a simulator installed at the moment. Otherwise I would have given it a quick try.
Regards
Charles
Maybe it is sufficient to sum just the highpass and lowpass outputs. If you have a filter Q of 2 and use the highpass with a weight of one and the lowpass with a weight of 0.2 you would end up with a shelf that has peak around 6 dB, followed by a notch slightly lower. If you need just a dip instead of a notch you can also add some part of the bandpass output. If the notch variant is sufficient you can just sum the output of ordinary sallen-key lowpass and highpass filters accordingly.
I don't have a simulator installed at the moment. Otherwise I would have given it a quick try.
Regards
Charles
Last edited:
I sometimes cannot understand what you are after. Could you use more words please?
I thought my premises were quite clear: a PEQ which is applied to a combined response does not consider the system that is a bass-reflex port working together with a woofer. It only boosts the combined response as if it was linear. As port an woofer have specific slopes, Qs and efficiency, the effect of a PEQ on an assisted system is different from a PEQ that is applied to the combined response.
VCad bass EQ applied to combined response, nearfield woofer + port and spliced to far field (wrong):
https://www.diyaudio.com/community/attachments/sp68-power-di-png.954957/
Actual measurement result:
https://www.diyaudio.com/community/attachments/disagreement-png.956362/
If this is what you meant, yes: a PEQ that is applied to the combined response in simulation got nothing to do with an assist in a real vented speaker. The assist must be applied so that it effects both sources before summation in simulation.
Not just tend to disagree.
From a theoretical point of view, those are just both two identical minimum phase systems.
With the same phase, group delay and frequency response.
No, I just skipped a step or two to fast for most people
A Linkwitz-transform is only good for a 2nd order system with a 12dB/oct roll off.
So in order to get a 5th order filter, you just have to add other highpass filters accordingly.
In this case you need a 3rd order filter on top of the LT.
(or a 2nd order followed by 1st order, but that is the same thing)
A super and easy quick trick to do this in simulation software, is either work with a target curve.
Or otherwise when that is not available, first add a LT at like 0.1Hz (Q=0.707) and than add another 5th order Bessel (in this case).
The result is a perfect Bessel response down to 0.1Hz.
The rest is just calculating and summing Q-factors.
It has been a while ago that I did that by hand, mostly because I am lazy.
You can also just find the needed Q-factor combination in those Q-factor-filter tables.
So for a 5th order Bessel filter we need;
- LT with Q=0.5635
- 2nd order HP with Q=0.9165
- 1st order filter (doesn't have a Q)
Here is a random example (first one I had in the list)
Unfortunately you sometimes get a bit of a difference because of rounding errors.
Important note, this DOES NOT work in the newest WinISD, because of a major bug!!!
Obviously this can also be done with other programs like VituixCAD, I just find WinISD WAY faster and quicker with these kind of things.
In the end it's nothing more than just cascading filters.
A very important note from a practical point of view, is that you need to take very close intention not to boost to much to prevent clipping in either the stage or the power amplifier.
I personally take 6dB as a absolute max value, unless I know there is plenty of headroom.
This only happens when we extend the response of the driver obviously.
Last edited:
Btw, I very rarely use anything more than a 4th order total filter response.
Mostly because it won't give more benefits, in sense of max power and/or cone excursion, but the groupdelay goes up significantly.
For woofers that also do the low-end, I usually optimize to 3rd order filter (= LT + 1st order HP) even with a similar response.
Mostly because it won't give more benefits, in sense of max power and/or cone excursion, but the groupdelay goes up significantly.
For woofers that also do the low-end, I usually optimize to 3rd order filter (= LT + 1st order HP) even with a similar response.
I take sheeple's point that if the port were far away, and the box large enough to be modal then it could change.. of course normally this is not significant. That said...
The driver works on the box straight away, and so does the box on the port. They may be out of phase but there is no delay. They are minimum phase.
Last edited:
Also, my data was wrong, I mixed something up there.
But isn't there delay when the port is mounted on the back of a box?