I am not a big fan of Magnepan, but I admit it makes some interesting sound. Also, one thing I want to add is, the manual of mine says it needs some refection from the back! I measured an evidence supporting the statement, because my space happened to be asymmetric. It's a living space with left side wide open. So, the left side one measured fairly well up to 15~17kHz, but the other, without the side reflection, barely measured to 13kHz or so. I tried many different things, and couldn't get the right side one to measure right... I think at least current generation maggies are better off with some close wall reflection, and looking at their older designs(virtually same mechanical structure), this wouldn't seem to be different for them, either.
Last edited:
Hi, possibly, directivity index on listening axis kind of shows how similar they are. Could be wide or narrower coverage, lower or higher DI, as long as DI is smooth, or complement to acoustics, what ever that is, perhaps optimized for something special. DI is always ~0 on very lows, and relatively high on highs, due to sound wavelength.
Room acoustics can be quite a wildcard, some homes and rooms have it fine while others worse, sometimes great effort can be made to improve and sometimes no effort is possible due to practical issues for example.
Here it's also important to understand context, what can be done with the room and what with the speakers, how is the positioning freedom. Also what one prefers, direct sound of the recording or local room sound, and there is stuff to tweak on both. Perhaps one wants to optimize for both, or just for one thing.
Anyway, my point was to open up why someone might be interested digging deeper on this stuff, while some are fine with almost anything. Mostly it's just listening skill, the more it has had a chance to kind of evolve, the more interesting the nitty gritty might become if interest is there, and the more important it is to at least try to understand phenomena behind stuff, acoustics, and of course the technology and all details on it, also to try and understand perception, and what role auditory system has in all of it. In general, reduce confusion over things.
Last edited:
"filler driver" seems to mean the philosophy or idea how you think, define or treat it, not what the driver is. Surely, one can do that smart approach as in the paper, but others can also think of it just as one of the drivers in the original plan and apply different design philosophy and methods. Am I understanding it correctly? If you don't think it's not a (good) midrange driver, is it because its most efficient(high SPL) FR is not covering midrange band wide enough?
Regarding "blindly pressing buttons", I know phase is very important especially in this crossover application, and I at least took care of that in the digital/electronic domain. One issue(probably most important) is that the summing happens in acoustic domain, so assuring correctness in electrical/digital domain is not enough. Unfortunately, I don't have any good professional measurement facility or any access to them, so I gave up that part. I might be able to take a look at the published phase response(if any) and try to compensate with that, but that would still be different from the actual response measured with the speaker. Lack of means and diligence(I admit!), I let the room correction take care of it.
Hi,
Sure they need reflection from frontwall ( theyr back) the issue is how long a delay (how far away the wall must be located).
I once tried to simulate the effect of a dipole with a reverb. I quickly hit a wall because sending an out of phase signal to generate ER needed was a real PITA but one thing was clear from initial test soundfiles: below 2m distance there is too much ER for direct sound not to be impacted. Painfully obvious.
Of course it's a matter of taste and music genre you listen to, but i'm used to low level of ER and this is definitely something i find bothersome to not have.
I usually listen to music which are not easy for loudspeakers either, in particular for open back system: electronic bass loaded, low overall dynamic on some passages, very high on other... doesn't fit well non boxed loudspeaker ime.
Issues with (room) reflection/diffraction can only be fixed through acoustic answer ( well at least past a certain freq range, LF and VLF are another story...).
Sure they need reflection from frontwall ( theyr back) the issue is how long a delay (how far away the wall must be located).
I once tried to simulate the effect of a dipole with a reverb. I quickly hit a wall because sending an out of phase signal to generate ER needed was a real PITA but one thing was clear from initial test soundfiles: below 2m distance there is too much ER for direct sound not to be impacted. Painfully obvious.
Of course it's a matter of taste and music genre you listen to, but i'm used to low level of ER and this is definitely something i find bothersome to not have.
I usually listen to music which are not easy for loudspeakers either, in particular for open back system: electronic bass loaded, low overall dynamic on some passages, very high on other... doesn't fit well non boxed loudspeaker ime.
Issues with (room) reflection/diffraction can only be fixed through acoustic answer ( well at least past a certain freq range, LF and VLF are another story...).
...which is a disc correction
Why it's called room? It changes the disc content
Why don't you look into free measurement software? You are flying blind without it. For example you could easily have a driver with the polarity switched and your DSP is wasting resources that could be easily fixed with a wire swap.
If you were building a clone how would you verify the design is correctly built without measurements to compare to the original?
Now if the original doesn't give you voltage drives per driver a set of measurements on and off axis and a system impedance curve I would look elsewhere.
Rob
Ok. But without trying to be pedantic the definition of something that generate from one side a positive pressure while the other side generate a negative one as well as no enclosure makes for the definition of a dipole in my book ( please keep in mind microphones are the same thing that loudspeakers at the other end of transducers and this is the accepted definition to what a dipole is).
Do you still have your measurements? I would be curious to see them ( and how you performed this kind of measurement too)?
I would not say off axis behavior was not something designers cared about during 80's, well maybe the majority of commercial consummers products might not have but from where i come it isn't totally true ( studio).
That said as can be read in Dave Smith interview linked, he didn't care about constant directivity but uniformity of off axis response, he did. Subtle difference but a difference.
We agree about means of managing directivity. The question being until which freq and with which shape it does matter? I would answer it's case related as different situations needs differents approach.
Yes, the quote was from Bruno Putzeys & Eelco Grimm paper that Andy (I think it was?) posted a couple of pages back, its a really interesting article.
It highlights the advantages of a wide baffle and explains their design philosophy.
One can still build a great 3 or 4 way active DSP based system with a narrow baffle, but think you will benefit from reading it, I know I did!
With all the different ideas discussed so far can you update us on your latest thoughts for your design?
Cheers
A.
Why don't you look into free measurement software? You are flying blind without it. For example you could easily have a driver with the polarity switched and your DSP is wasting resources that could be easily fixed with a wire swap.
If you were building a clone how would you verify the design is correctly built without measurements to compare to the original?
Now if the original doesn't give you voltage drives per driver a set of measurements on and off axis and a system impedance curve I would look elsewhere.
Rob
You are underestimating what the room correction can do.
I actually had my speaker delivered with external passive crossovers, and one of it had a wrong polarity at the output. I could notice the problem right away listening to it, not needing a measurement system for such an obvious thing. And there was no original design as a reference.
Last edited:
Oh, now I see where it came from.
. Unfortunately, it will continue to change, and I haven't reached a good point to share my updated plan. Please stay tuned.
Well no. The idea is to have a working speaker. Not one that relies on room correction to "fix it"
Depend on where and slopes involved it may not be as obvious as you think. I am not taking bass drivers here.
Rob
Rob. I am not selling speakers, and I just want a good sounding system as a whole, and it works for me, at least, or probably for many people. This is a WORKING "system".
It was only the polarity of one of 3 drivers of one speaker, and very obvious. I don't know what slope can hide such obvious things, and require a measurement system to detect the fault. I don't have a lot of speaker experiences, even including listening to ones, but it sounds like a faulty system.
It was only the polarity of one of 3 drivers of one speaker, and very obvious. I don't know what slope can hide such obvious things, and require a measurement system to detect the fault. I don't have a lot of speaker experiences, even including listening to ones, but it sounds like a faulty system.
One can use a tweeter as a woofer but the output will be insufficient except possibly for headphones. A filler driver only requires a low output at low midrange frequencies and so the driver has been designed to have a small surface area and small displacement in exchange for improving other driver parameters that are relevant to a filler driver. It can be used as midrange driver but the performance will be poor compared to a midrange driver. Similarly a midwoofer or a wideband driver can be used as a midrange driver but again the performance will be poor compared to a driver with parameters optimised only to cover the midrange.
Can I check. You intend to design a 3 or 4 way speaker without using a microphone? How does your room correction software/hardware determine the filters for each driver? Do you intend to play each driver in each speaker individually to determine the transfer function between it and the listening position? (There is going to be a problem with this). Or something else?
Which Satori model are you speaking about ? The MD60N like I plan to use (see below) ?
It shows a FS at 450Hz, so going to 500Hz is too low IMHO for linearity of FR and phase during the transition with the Woofer - If you intend like me to use passive crossover with gentle slopes (like on my 612SP 6 to 12dB asymptotic serial-parallel crossover).
I would not go under 1kHz (2xFS) with a 12dB/Oct. slope. I plan to use the MD60N between 2000-3000Hz and 7000-8000Hz, where its linearity stays very good from 0 to 15° off axis. As possible, I will try to stay far from the zones when the responses shows fast changes or linearity accidents. This is a prerequisite for a smooth, in-phase transition with gentle slopes crossovers.
Naturally, if it is about another larger dome midrange with a lower FS, or with a multi-amplification solution with active steep slopes and corrections, it's a different story !
T
Last edited:
To add to what @andy19191 is saying about "filler" drivers: I read a comment about this from someone, and I can't remember who it was (Jeff Bagby perhaps?). "A good midrange driver has a usable bandwidth of at least 3 octaves, and usually 4 octaves. A filler driver has a usable bandwidth of about 2 octaves". For some speaker designs, 2 octaves is enough to bridge the gap between the woofer(s) and the tweeter.
Regarding Magnepans. I have heard some installations sound bland and unimpressive, and I have heard other installations sound just fantastic. The really good ones had a listening position like this:
j.
Regarding Magnepans. I have heard some installations sound bland and unimpressive, and I have heard other installations sound just fantastic. The really good ones had a listening position like this:
j.
Yes : particularly with Space Ambient Electronic Music...
OK, I know : not everybody appreciate this kind of music... But Magnepan sounds great with that, enveloping you in the sound !
T
Didn't say you were. The point is a measurement system can help you get where you want to go a lot faster. You choose not to use one to help and verify set-up fine.
Happy with what you have now great!
Rob
Last edited:
Which Satori model are you speaking about ? The MD60N like I plan to use (see below) ?
View attachment 1317881
It shows a FS at 450Hz, so going to 500Hz is too low IMHO for linearity of FR and phase during the transition with the Woofer - If you intend like me to use passive crossover with gentle slopes (like on my 612SP 6 to 12dB asymptotic serial-parallel crossover).
I would not go under 1kHz (2xFS) with a 12dB/Oct. slope. I plan to use the MD60N between 2000-3000Hz and 7000-8000Hz, where its linearity stays very good from 0 to 15° off axis. As possible, I will try to stay far from the zones when the responses shows fast changes or linearity accidents. This is a prerequisite for a smooth, in-phase transition with gentle slopes crossovers.
Naturally, if it is about another larger dome midrange with a lower FS, or with a multi-amplification solution with active steep slopes and corrections, it's a different story !
T
Yes, I meant MD60N, and thank you for the relevant answer. I also thought 500Hz is too low, but graph seems to tell me 800Hz is okay for the sake of MD60N itself. Then I am reluctant to place a crossover point close to the frequency human ears are most sensitive to. I am sure good speaker engineers can even do the coherent and well-behaving passive crossovers, but I am just an amateur in that regard, and would like to avoid 800Hz if possible. I am now wondering if 600 or 700 is still acceptable to MD60N. Xmax of just 1mm is a bit concerning, too. It got only half Xmax of VM527, and about the same usable sensitivity(got to suppress the wide "peak" ranging 700 ~ 4000Hz), then somewhat bigger diaphragm area. Now VM527 looks slightly better...
I usually use very steep filtering(100+dB/oct), so my MD60N may be in a better situation than the one in your design. I'd still be careful to put some low frequency input power into dome mids, like you, but also want to maximize the use of the dome mid/mid-high. From 800 to 4000 ~ 5000 seem reasonable, but again, the concern of crossover in the most sensitive frequency range... I hope your last line is right(actually encouraging!), and 600 or 700 at low-end is okay.