Hi XRK,
As you are using a miniDSP, try putting a notch at 5, 7 and 9kHz to smooth out the off-axis, as the breakup stays almost constant with angle. Try to optimize the 5-45deg response without worsening other angles.
Have a look at the time domain also, the result should be a compact impulse at (almost) every angle.
Could you plot the result with 30dB range?
As you are using a miniDSP, try putting a notch at 5, 7 and 9kHz to smooth out the off-axis, as the breakup stays almost constant with angle. Try to optimize the 5-45deg response without worsening other angles.
Have a look at the time domain also, the result should be a compact impulse at (almost) every angle.
Could you plot the result with 30dB range?
I will give that a try - thanks for the constructive feedback. I try not to do any EQ that is steep in the 500hz to 5khz range to preserve the phase linearity but I guess the 5khz is at the edge so probably ok. This should help the time domain.
Here is hor. polar measurements of a 4" mid in narrow baffle, and a 3-way with it. Sorry but I don't take the measurements of xrk's 5" FAST for valid. In my speaker 60¤ is -5dB at 3,5kHz and drops suddenly to -10dB at 4,5kHz. Cone profile and dustcap/plug play a role here too.
Attachments
Last edited:
Right - I have about the same relative drop at 60deg as Juhazi it is just the cone breakup doesn't happen until 5khz. The plot scale is different mostly because that was the size produced from a screenshot of a small 10in netbook LCD. I would have to transfer data to a desktop and reopen it there to get larger plots at 5dB steps.
Go easy guys, the PRV FAST design is pretty good, compromised yes but so is everything, let's not pick it to pieces. Particularly as we are discussing avoiding crossovers in the critical range, which this achieves.
You could exchange the driver for a BMR if your want wider dispersion. Or stick it in a waveguide fir better directivity but the size grows...
You could exchange the driver for a BMR if your want wider dispersion. Or stick it in a waveguide fir better directivity but the size grows...
As the speaker is mostly 'linear phase' device, adding a linear phase eq to correct any anomaly makes the system more 'phase linear', if it makes any sense. Look at teh time domain. Of-course you don't make any one angle perfect, but look at trends. THats why suggested 5-45deg and to ignore on-axis.
Yep. Agree that 1-4 kHz range is sensitive with bad implementations, but there's no reason to avoid if excellent implementation is possible. Maybe this is just playing with words...
But if the large bandwider driver is a cone : does not it suffer from two big defaults : those heat issues we spoke above and the lake of Sd in the low to push air... at least one of this worst default is the complex break ups it has to deal with with wide band it has to reproduce like the BL deals between the micro dynamics in the highs and the macro dynamics mouvement of the cone in the low. Following this philosophy I believe the mid bass has to be a standalone to not suffer of the Bass and not to be a brake to the more dynamics signal above it, but as it needs to be fast and push air (80/100 hz to 500/800 hz ?) in the same time, yes it has to be a standalone and to have a Sd like a 8" to 12" !
Maybe one of the sucess of the CD driver is their low Sd which allow them to be more piston like ? Does a Magneplanar or ESL avoid the defaults of a wide bande cone driver ?
Maybe one of the sucess of the CD driver is their low Sd which allow them to be more piston like ? Does a Magneplanar or ESL avoid the defaults of a wide bande cone driver ?
Last edited:
Thi may have already been mentioned. This is why folks in the Fullrange forum cross below 400Hz in a FAST implementation with a fullrange driver handling the critical 500Hz to 5kHz band where all the phase and spatial point source. coherency is important.
The problem is not being able to get high SPL levels with a typical fullrange driver. Recently rather sensitive and high power wide band drivers are changing this. Take for example the case of a 5in driver that goes from 300Hz to 15khz with up to 225watts rms and 95dB sensitivity. Couple that with a woofer XO below 400Hz and you have a system that doesn't have any XO in the critical human hearing range.
http://www.diyaudio.com/forums/full-range/259293-prv-5mr450-ndy-fast-applications-2.html#post3995448
I made a loudspeaker like that with Fostex FE126E in OB and Beyma SM115/K in BR for my friend. Xover frequency is at 290Hz - it's a 2dB gridd.
It makes for a simple xover, measures well and sounds better than most of the loudspeakers that are on the market for fraction of the price. Here are the measurements on axis, 30 and 60 degrees in BLH cabinet:
An externally hosted image should be here but it was not working when we last tested it.
Here is the response with Beyma bass added and xovered at 290Hz (that nasty dip at 200Hz is floor bounce as GM calculated 30-40 pages ago 🙂 ):
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Most of the full ranges that my friends and acquaintances have are Fostex, Lowther, Sonido and there are few vintage Philips, RFT and Lorentz full ranges. I saw couple of Supravox 8" FR but never had the chance to listen to them.
But i must say that none of them measured any of speakers that they own. Friend of mine and owner of the Fostex FE126E is the first one that has FR drivers at home and he let me measure them 🙂
But they do sound nice because of lack of shout that is present in most FR drivers out there - not that i saw measurements but i listened a number of it and these little things sounded essentially as a very nice and clean midrangers. I measured them after listening them for a couple of hours. That's my favorite approach because i love to see what do i hear, how do i interpret it and how measurements relate to my hearing. I sad to the guy that there are no shout and highs are a bit hot for my taste before measuring them and measurements later proved me right. The big peaks at 7 and 14KHz didn't kill me because we listened them slightly off axis.
Other guys that are into FR drivers 8", 10", 12" and 15" that i know don't measure them and if they do, they do not post measurements because they like the legends and myths about their speakers better than facts - that most of them are shouty and have poor on and off axis response.
But i must say that none of them measured any of speakers that they own. Friend of mine and owner of the Fostex FE126E is the first one that has FR drivers at home and he let me measure them 🙂
But they do sound nice because of lack of shout that is present in most FR drivers out there - not that i saw measurements but i listened a number of it and these little things sounded essentially as a very nice and clean midrangers. I measured them after listening them for a couple of hours. That's my favorite approach because i love to see what do i hear, how do i interpret it and how measurements relate to my hearing. I sad to the guy that there are no shout and highs are a bit hot for my taste before measuring them and measurements later proved me right. The big peaks at 7 and 14KHz didn't kill me because we listened them slightly off axis.
Other guys that are into FR drivers 8", 10", 12" and 15" that i know don't measure them and if they do, they do not post measurements because they like the legends and myths about their speakers better than facts - that most of them are shouty and have poor on and off axis response.
Last edited:
I hear this so much that I have to point out the fallacy. Not all compromises are the same. Compromising on a $100,000 salary is not the same as compromising on a $50,000 one. This comment seems to try and put all compromises on an equal footing and that simply is not the case.
That's a pretty big "If". You'd have to show me how ones does an "excellent implementation", i.e. one that does not have lobbing errors, etc. Sure along one axis you can do fine, but this is an over simplification of the problem. I accept that excellent crossover can be done for coaxial units, but the frequency response of the drivers in this case tends to suffer. KEF has done some excellent work here as my recent measurements show.
I was referring to IIR filters vs FIR filters. I only have IIR filters on my miniDSP and putting a high Q notch or peak introduces a phase wrap typically. FIR filters in EQ can be designed to have a notch without wrapping the phase - although they have their own issues with audible pre ringing used to effect the flat phase EQ. Put too many of these IIR induced phase wraps in the 500Hz to 5kHz range and it sounds phasey - wierdness that makes one cringe and want to leave the room despite the FR being flat.
A "phase wrap" electronically, which is exactly what is needed to "unwrap" the acoustic phase from the high-Q peak. Net result is no peak and minimum phase. Speakers are not "linear phase" but they do tend to be minimum phase.
- Status
- Not open for further replies.