Check out the frequency response of these

[digicidal]

Isn't that just a peak centered around 100hz combined with below 6db of BSC (which is understandable because people don't live in anechoic chambers)

I thought I did because I hung a blanket over my window.... man, now I have to rethink everything.

@ Dennis Murphy - I agree with that being midrange... however, in a random search of some definitions a couple of years ago, I found people placing it in all sorts of places - but then again I've always just considered everything below ~300Hz "bass" and everything above ~3KHz "treble" and everything else "midrange". Is there a standard definition of the terminology somewhere - I wasn't able to find anything like that - although I haven't searched for awhile.

 

[Dennis Murphy]

Isn't that just a peak centered around 100hz combined with below 6db of BSC (which is understandable because people don't live in anechoic chambers)

That's certainly an interesting point to raise. But if you interpret the rise at 100 Hz as a peak, and the true reference point at the minimum of the response curve, and remembering that there's a 10 dB demarcation window, then that would imply that Rick didn't invoke any baffle step compensation whatever, because the response at 1 khz is 6 dB above the lowest reference point. I doubt that Rick would claim that he hadn't adjusted to some degree for the baffle step rise. I don't know of any designers who claim that room response realities obviate the need for baffle step compensation. Room boundary reinforcements don't really come into significance until 100 Hz and below. The Baffle step takes hold several hundred Hz above that.

 

[BoredSysAdmin]

My point was that you can't accept factory plots at face value. The plots for the Selah are quite flat when you compare them with measurements for other speakers done at the same independent and extremelty well-equipped facility.

Check out the Wilsons some time.

Dennis, First I think it would somewhat inappropriate to accuse Harman of inadequate speakers measurements without having strong facts proving so. From what I gather they the industry pioneers regarding precise measurements and result validation plus strong believe in simple idea - precise speaker is a good speaker.

I know we can't compare mine (almost none) and your experience and knowledge regarding audio design, but even your own project seems to me a much better than the speakers OP mentioned.

I still don't get - how come 3-3.5db dip from about 130 to 700hz doesn't affect SQ?? I guess it depends in the end on personal preference and such coloration might sound pleasant for people

p.s: I didn't auditioned too many speakers, but I happed to listed to Wilsons a while ago. Both to Watt Puppy (25k pair) and 100k/pair model I forgot the name - but they there very big towers built from 3-4 separate pieces.
Not that they didn't left me impressed of their sound; They also made me wonder what sort of snake oil they selling for such copious amount of money - one could buy whole house for.

 

[GranteedEV]

and remembering that there's a 10 dB demarcation window, then that would imply that Rick didn't invoke any baffle step compensation whatever, because the response at 1 khz is 6 dB above the lowest reference point. I doubt that Rick would claim that he hadn't adjusted to some degree for the baffle step rise.

You managed to count the six ticks? I counted one and got bored. No wonder you can design awesome speakers.

I just glanced at it and assumed 4db of BSC and -2db in that region.

 

[Dennis Murphy]

I thought I did because I hung a blanket over my window.... man, now I have to rethink everything.

@ Dennis Murphy - I agree with that being midrange... however, in a random search of some definitions a couple of years ago, I found people placing it in all sorts of places - but then again I've always just considered everything below ~300Hz "bass" and everything above ~3KHz "treble" and everything else "midrange". Is there a standard definition of the terminology somewhere - I wasn't able to find anything like that - although I haven't searched for awhile.

I don't think there is. But bass just can't apply to frequencies above 200 Hz or so. One thing to keep in mind is that the highest note on the piano is only around 4 kHz. Most of what we consider treble is overtones generated by lower fundamentals. So any precise transition point from mids to treble gets tricky. But 2-3 kHz seems reasonable. That's the region that's best handled by a very small midrange or a tweeter.

 

[3db]

600 Hz is high bass? The orchestral tuning frequency is A=440 Hz. That's the highest string on a viola, the second highest on a violin. A standard male non-professional tenor voice can hit concert A a couple of times before turning red and faking a cough. That's solid midrange territory.
On most any modern 3-way, the midrange--not the woofer--would be handling those frequencies. Upper bass is more like 250 Hz at the highest. I'm not sure what's causing the little dip in the upper bass (240 Hz) of the Selah. I've seen that a number of times on several designs, including my own. It doesn't seem to have any audible consequences.

I've seen alot of graphs much worse than this and a few better than this. Relatively speaking its fairly flat curve. Thats why I brought it to the attention of this group. There will always be someone who can find something better but these are exceptions rather than the rule.

I brought this graph to raise a point here. Alot of people who chimed in here are throwing around terms that are in correctly defined... such as midbass. Maybe the group as a whole can decide which frequnecy ranges makup sub bass, bass mid bass, midrange and treble. That way when reading a frequency graph, people can get the values right. digicidal, that 3.5 db is still a stretch. Read it carefully. Also, if you assume 75 degrees off center as a typical listiening postion, you may want to correct that so you can enjoy your sound system a bit more.

 

[Swerd]

This thread could use some charts showing the frequencies of music:

http://www.har-bal.com/index.php?/frequency-chart.php

http://www.independentrecording.net/irn/resources/freqchart/main_display.htm

http://www.psbspeakers.com/audio-topics/The-Frequencies-of-Music

 

[jp_over]

questions

This thread could use some charts showing the frequencies of music:

http://www.har-bal.com/index.php?/frequency-chart.php

http://www.independentrecording.net/irn/resources/freqchart/main_display.htm

http://www.psbspeakers.com/audio-topics/The-Frequencies-of-Music

Rich,

One of those charts lists the bass (human voice) as going down to 100hz whereas the other lists it as going down to 60hz. I've also heard 80hz is the approximate low end limit of male human speech / vocals.

Any input would be helpful as I'd like to set my crossover (on the receiver) to allow my speakers to handle all the human voices and vocals.

Thanks!

 

[jostenmeat]

Rich,

One of those charts lists the bass (human voice) as going down to 100hz whereas the other lists it as going down to 60hz. I've also heard 80hz is the approximate low end limit of male human speech / vocals.

Any input would be helpful as I'd like to set my crossover (on the receiver) to allow my speakers to handle all the human voices and vocals.

Thanks!

I wouldn't worry too much about this you see, because dialogue intelligibility has to do more with the harmonics than the fundamentals, or so I've been told here after a lot of pestering.

I've never heard of anyone who can sing down to 60hz, but I knew a couple of guys who can sing below 80hz (very uncommon). IOW, don't worry about it, you probably do not own a single recording with anyone singing below 80hz, quite seriously. Unless you are into certain vocal works by Ockeghem, and even then your odds are low.

 

[digicidal]

Also, if you assume 75 degrees off center as a typical listiening postion, you may want to correct that so you can enjoy your sound system a bit more.

I stand corrected on the terminology... what I should have said was: "It seems like there's a pretty decent dip from the upper-bass to the lower midrange"...

Well, unless my geometry has completely failed me - no, I do not consider 75degrees off-center as a typical listening position. Even if the speakers and my listening position were an equilateral triangle, and the speakers had no toe-in at all... I would still only be at most 60deg off-center, right? Couple that with a recommended toe-in pointing to a point a foot or two behind your head and you're probably more like 5-10deg off-center to each speaker.

Am I missing something basic here??

@Swerd - thanks! That second link is very nice - including examples and effects for each range as is the PSB link. Also interesting that it shows 'midrange' as being as high as 6KHz and as low as ~240Hz (hard to really say because of the scale being in octaves but the legend lines being in Hz). Good to know.

 

[jostenmeat]

I stand corrected on the terminology... what I should have said was: "It seems like there's a pretty decent dip from the upper-bass to the lower midrange"...

Well, unless my geometry has completely failed me - no, I do not consider 75degrees off-center as a typical listening position. Even if the speakers and my listening position were an equilateral triangle, and the speakers had no toe-in at all... I would still only be at most 60deg off-center, right? Couple that with a recommended toe-in pointing to a point a foot or two behind your head and you're probably more like 5-10deg off-center to each speaker.

Am I missing something basic here??

I believe you would only be off-axis by 30 degrees, but to each* speaker.

After toe in, you may only have 5-10 degrees, or you may have zero degrees. I don't understand the reasoning, but I think I've come across statements that say some speakers enjoy more toe in, and others not so much? I prefer to be on-axis, and a person I respect and trust says to have the axes cross just right in front of your face. (I suppose that might make each speaker onaxis to the opposite ear?)

 

[digicidal]

I believe you would only be off-axis by 30 degrees, but to each* speaker.

Exactly... I was simply saying that even if they were straight ahead instead of targeting a sweetspot inbetween... the most you could ever be off-axis would be 60 deg (and at that point you'd be 0deg to the other speaker) - even though I would still expect that your couch probably isn't that long and you'd definitely have bad imaging at that point regardless.

Although, I've heard it both ways (crossing in front of or behind) I've also heard speakers that really sounded bad set up that way and required much less toe-in for proper imaging. I think (as with most things speaker-related) it not only depends on the speakers, but on the room and distance from them as to what the best toe-in angles are.

EDIT: Although now that I think of it... even if you were all the way off to the side... wouldn't you just be 45deg off on one and 0 on the other... since you're basically at a right triangle then with the 0deg speaker as the apex? About the only way you could get close to 75degree would be if they were pointed straight ahead and you were less than a foot or so away from them inbetween... or am I still missing something??

On the other hand, if 3db is suggesting that you have to look at the graph of 75degrees off-axis to find the swing I'm talking about... no.. then it's way worse off. I'm using the on-center graph (the top curve in the top graph).

Here's an annotated version for clarification - since I think I am failing to describe with words well enough:

 

[Dennis Murphy]

Thanks for the plot and labelings. It looks to me like the speaker fits within a 4 dB windown up to around 15k. That's not too shabby. When you're talking about measurements that flat, you have to listen to the speaker to form an opinion. Obsessing about a little peak here, a little dip there just doesn't mean anything.

 

[digicidal]

Thanks for the plot and labelings. It looks to me like the speaker fits within a 4 dB windown up to around 15k. That's not too shabby. When you're talking about measurements that flat, you have to listen to the speaker to form an opinion. Obsessing about a little peak here, a little dip there just doesn't mean anything.

Oh I agree... I've got some single-driver fostex bass-reflex that look horrible on both ends (but very nice in the middle). I was not trying to imply that these speakers would sound horrible or anything - nor even that a 3db 'dip' would create any form of significant deficiency in the FR during critical listening. After all if you are listening a referrence levels - a 3db difference will only amount to about ~4% and only then if you had an anechoic room in which you were listening to them. It's quite possible in a real room that you'd have a hump somewhere in that same region and they might actually wind up balancing out very nicely.

As this was in celebration of the chart itself - I simply compared it mentally when I first viewed it to this:

Which is the chart for the Ascend Sierra-1's - which I'm about to pull the trigger on and thus have the most familiarity with because I've been returning to their site often while continuing my research pre-purchase.

 

[3db]

Thanks for the plot and labelings. It looks to me like the speaker fits within a 4 dB windown up to around 15k. That's not too shabby. When you're talking about measurements that flat, you have to listen to the speaker to form an opinion. Obsessing about a little peak here, a little dip there just doesn't mean anything.

Compare that to this speaker for $12K....

http://www.soundstagenetwork.com/measurements/speakers/wilson_duette/


Now thats a bad curve regarless of price. Couple the price into this and one is selling snake oil.

 

[digicidal]

Compare that to this speaker for $12K....
Now thats a bad curve regarless of price. Couple the price into this and one is selling snake oil.

You got that right... and I've actually heard them and they don't sound bad at all - of course, that's without being able to do a blind shootout with anything else.

Of particular humor (at least I thought so) was "Chart 4 - Deviation from Linearity at 90dB and Above". At first this confused the crap out of me since it seemed to be indicating perfect linearity, which clearly was not being indicated in the FR plots... then the light came on. These charts simply state that the exact same crappy FR is present whether the speakers are playing at 70db or 90db - or at least I think that's what they're saying.

That's pretty hillarious, because most halfway decent drivers should have extremely similar FR within that range - now maybe if pushed over 110db you'd have something else entirely... but for the cost, I'd be utterly stupified if they couldn't reach referrence levels while maintaining their composure.

Although I don't feel it justifies anything as far as the Wilsons are concerned... I'm guessing there's about $1000 just in cabinet work in a pair of those... they are extremely complex cabinet designs - not that that it makes it better... just sayin'.

 

[digicidal]

I know we can't compare mine (almost none) and your experience and knowledge regarding audio design, but even your own project seems to me a much better than the speakers OP mentioned.

Is there an actual measured response plot for those speakers? All I was able to find in that thread were the estimated response plots, and only for bass response at that... which I would certainly hope were nearly perfectly flat. If not, by the time you factored in the actual physical realities of the specific drivers, and anomalies in cabinet construction, crossover network, etc... you'd probably be waaay off once measured.

Not that I think they would necessarily all be audible differences - but even the decision to use bondo as filler could potentially show up as a (very minor of course) measured difference if compared to a glue-only fabrication. Additionally with any driver there will be a tiny amount of variance between any two units rolling off the assembly line - at least until we have robots building everything in space for us.

 

[Dennis Murphy]

As this was in celebration of the chart itself - I simply compared it mentally when I first viewed it to this:

Which is the chart for the Ascend Sierra-1's - which I'm about to pull the trigger on and thus have the most familiarity with because I've been returning to their site often while continuing my research pre-purchase.

Hi What's the origin opf that plot--who did it? Thanks.

 

[jostenmeat]

Which is the chart for the Ascend Sierra-1's - which I'm about to pull the trigger on and thus have the most familiarity with because I've been returning to their site often while continuing my research pre-purchase.

I believe this graph (NRC) has been shared on a few occasions here for the Sierra1. 0, 15, 30 degrees, mic on tweeter axis.

http://www.soundstagenetwork.com/measurements/speakers/ascend_sierra1/

 

[BoredSysAdmin]

Thx, jostenmeat
I know we started this thread with one expensive bookshelfs which I still don't consider to have flat response, however I did found few graphs I liked much more:
Almost all PSB's:
http://www.soundstagenetwork.com/measurements/speakers/psb_image_t45/
http://www.soundstagenetwork.com/measurements/speakers/psb_image_t6/
http://www.soundstagenetwork.com/measurements/speakers/psb_synchrony_one/
http://www.soundstagenetwork.com/measurements/speakers/psb_synchrony_twob/
NHT Classic Three :
http://www.soundstagenetwork.com/measurements/speakers/nht_classic_three/
and:
Paradigm Reference Signature S1 v.2
http://www.soundstagenetwork.com/measurements/speakers/paradigm_signature_s1_v2/

plus another crown jewel:KEF Reference 201/2 - 6k/pr
http://www.soundstagenetwork.com/measurements/speakers/kef_201-2/

still much better than any wilson in both price and measurements