Looking at the Alpair FR plots posted in this thread, it might seem surprising that such a big notch in the 7-9k range and a big peak past 10k can be ignored by most people yet studio engineers tasked with EQing on a daily basis won't be so surprised... The range above 6k doesn't have a lot of fundamentals nor is it easy for people to recognize specific pitches in the HF. Harmonics are splashed (or perceived as splashed) more than logarithmically spaced. For instance, if there is ample cymbal from 6k to 7.5k a notch from 7.7k to 8.7k and then more cymbal above, most people will hear it as a coherent cymbal, not as a cymbal missing it's body (or however someone might describe a similar notch affecting an instrument in the low mids). Peaks above 10k are often not terribly strident and just bring more air (but I'll get into the problem of this below):
I can see why dewardh is more concerned about notches in the 2k range and understand the logic, yet I think it's fair to ask speakers to be relatively unbiased reproducers (AKA reasonably flat). Why? Because large frequency dependent amplitude swings in consumer reproduction equipment means that studio engineers are more constrained to LCD engineering practices. This is a shame because many times I've had a mix that I enjoy very much on hi-end monitors and I know it won't work in the majority of consumer spaces because the emphasis i'm playing with will get blown out of proportion on crappy systems with bumpy FR.
Very much appreciate dewardh reminding everyone that playback SPL shapes our perceived spectral balance and that mic'ing techniques are usually artistic decisions rather than efforts to reproduce exactly what the hall sounded like in VIP orchestra seating. (Not to mention most live music includes amplification in some way, so whatever was 'natural' left the building with the talkies).
I'm in favor of better consumer FR to give engineers more room to let the music be the music (less dynamic range reduction, less eq, etc)... And to give engineers more room to let a classic performance that is slightly off go untreated because the artifact/essence of it sounds good given proper repro... And finally for us electronic musicians, to provide more room for interesting mixes that are pushed one way or the other purposefully.
LCD mixing/mastering sucks. Really! I could write for an hour on how sad it is to massage a mix that had life but was wiley and uncooperative into a properly balanced tune only to lose significant amounts of essence... It's a big topic and certainly doesn't belong in this thread so I'll leave it alone.
I can see why dewardh is more concerned about notches in the 2k range and understand the logic, yet I think it's fair to ask speakers to be relatively unbiased reproducers (AKA reasonably flat). Why? Because large frequency dependent amplitude swings in consumer reproduction equipment means that studio engineers are more constrained to LCD engineering practices. This is a shame because many times I've had a mix that I enjoy very much on hi-end monitors and I know it won't work in the majority of consumer spaces because the emphasis i'm playing with will get blown out of proportion on crappy systems with bumpy FR.
Very much appreciate dewardh reminding everyone that playback SPL shapes our perceived spectral balance and that mic'ing techniques are usually artistic decisions rather than efforts to reproduce exactly what the hall sounded like in VIP orchestra seating. (Not to mention most live music includes amplification in some way, so whatever was 'natural' left the building with the talkies).
I'm in favor of better consumer FR to give engineers more room to let the music be the music (less dynamic range reduction, less eq, etc)... And to give engineers more room to let a classic performance that is slightly off go untreated because the artifact/essence of it sounds good given proper repro... And finally for us electronic musicians, to provide more room for interesting mixes that are pushed one way or the other purposefully.
LCD mixing/mastering sucks. Really! I could write for an hour on how sad it is to massage a mix that had life but was wiley and uncooperative into a properly balanced tune only to lose significant amounts of essence... It's a big topic and certainly doesn't belong in this thread so I'll leave it alone.
Last edited:
Markus being Swiss should be able tell you that there is many kinds of green cheese. Now which one exactly? ;-)
In other news 8.6766×10^49 angels can dance on the head of a pin.
Markus, by good dispersion, do you mean smooth, uniform off-axis response? You can pretty much tell at what frequency a 4" driver is going to start becoming directional. I'd take a look at the Vifa TC9 or the Scan 10F ($$). Both are low distortion, and should have predictable response. They might not be as chaotic as some of the full range drivers.
Those 2 have been recommended several times already - and could be excellent candidates for the right application.
What I don't think is yet clear is just what metrics markus would deem as "good dispersion and high xmax", or the intended application - although apparently neither criteria have been met, or I'm sure we'd have seen a "thanks, gentlemen, that's exactly what I was looking for "
instead, there's been much else
What I don't think is yet clear is just what metrics markus would deem as "good dispersion and high xmax", or the intended application - although apparently neither criteria have been met, or I'm sure we'd have seen a "thanks, gentlemen, that's exactly what I was looking for "
instead, there's been much else
I agree with you 100% Chrisb
. It seems this post is going no were and its like a dog chasing it's tale 
. Well again I hope the guy that started this post finds what he is looking for? . Mr. Daniel
Last edited:
Take all available 3" or 4" drivers and look at their polar response and Xmax. There will be one driver that has the highes Xmax and widest dispersion.
Maybe there are drivers that are way better in only one of those two categories. Nevertheless it should be possible to rank them based on those performance criteria. What I've learned so far is that none of you has such data or posts it.
And no, manufacturer data is not adequate. They tend to post only limited data without mentioning how the data was measured.
Related to that point Markus, by what measure do you define Xmax? That's a serious question / point, since there isn't a single industry standard definition for it, most result in different figures for the same drive unit (sometimes significantly so), and very few manufacturers (or 3rd parties for that matter) state which they used. For e.g., a handful are:
10% THD
10% of any individual distortion type (whichever is highest)
VC gap height minus VC length
Per above, take absolute value / 2
Half of Xmech
There are lots of others too. It doesn't matter whether one agrees with the particular methodology employed to put a number to it, that can be argued ad infinitum, just to point out it's a rather nebulous term.
10% THD
10% of any individual distortion type (whichever is highest)
VC gap height minus VC length
Per above, take absolute value / 2
Half of Xmech
There are lots of others too. It doesn't matter whether one agrees with the particular methodology employed to put a number to it, that can be argued ad infinitum, just to point out it's a rather nebulous term.
Last edited:
Good luck on getting that from manufacturers, since as noted not all define it in that way (& you will need to establish some specific figures before it can be employed anyway). Also, not all that many DIYers will necessarily have the gear to make an especially good job of measuring it -especially when you consider that you're not likely to get consistent testing methodologies and / or equipment from them either, nor are they automatically likely to be using / have used your particular definition either.
This is unfortunately (one of) the problems inherent to driver testing -the lack of industry-standardised test equipment, methodologies and definitions. Which is not to suggest testing is irrelevant (far from it), just that there are limits / problems in relying on it unless you can ensure consistent methodologies from one unit to another. That's why Zaph's site is useful since while the equipment was limited in outright terms, there is a consistent baseline.
Last edited:
"Starts to compress" by how much? Measured how? It's a can of worms . . .
Distortion/amplitude is not a straight line with a knee on the end even in the best of cases, and for design purposes it's a "ballpark" figure at best anyway. The information needed for actual design is easily derived from the distortion curves, and using them each designer gets to make his/her own decision what is an acceptable low frequency extension for a given driver.
Not correct. Others posted their own measurements, and others still linked to various manufacturers data for their drivers.
So where's the "ballpark" data? As someone once said, "More data, less wank".
Zaph makes that point repeatedly on his own site . . . regularly warning against making "comparisons" outside a particular test group or to "external" sites, for the reasons you just cited. Of course there's some utility in such comparisons . . . breakup patterns and gross frequency response and distortion tend to be pretty consistent across various measurements systems . . . but looking for little squiggles or .5mm differences in xmax is not going to tell anything useful regarding how a driver will sound in service.
Correct and the reason why I've asked in a general fashion, "3" or 4" driver with very good dispersion and high xmax?"
- Status
- Not open for further replies.