Member
Joined 2003
Harman room target curve lands somewhere in the middle of your two options above:
FWIW my latest creation is a 6.5" two way "bookshelf", shallow waveguide on the tweeter and some facets. I landed on a 2kHz crossover with the following power & di:
Full details on this speaker are here:
https://www.htguide.com/forum/forum/mission-possible-diy/953306-atari-it-means-to-hit-a-target
My main point is that there is a big difference between "hearing" something and actually "experiencing" something.
Sure, if I go into hyper super review mode, I can hear all kinds of things.
But how is that even close to being representative to just listening and enjoying music, especially in a living room with all kinds of stuff going on?
And even within that context, there is a difference between a small nuance and a clear difference.
A small dip or even peak of just a dB or so is a difference in a (very) subtle nuance at best (if you can already hear it), it doesn't make or break an entire system all of a sudden.
Let alone giving the impression that someone would miss out on something entirely.
well, I didn't read your discussion completely so sorry if also this post is off mark, just wanted to point everyone to listening tests to try what is audible and what is not to themselves, perspective to what is meaningful and what is not. The fact that some people can get to higher levels in such tests than some others means listening skill vary, and what is subtle nyance to one might be quite audible to some, so results provide perspective how to relate to things. If you got to high levels on some tests then a big difference to you might be subtle nyance to some one else on that regard. Or if you didn't get very high levels, then subtle nyance to you might be big difference to someone else. A perspective thing which you could use to your advantage as it is more knowledge of yourself, about your hearing / listening skill. I consider listening skill means to understand what I hear, which enables use of logic over what I hear.
Anyway, to answer your question it's helpful in this sense: even if I got into super hyper listening mode, the dips were still much harder to hear than peaks, no matter how hard I tried
same goes for casual listening, if I casually listen and feel something is off with the sound, it's much more likely a peak, than a dip, which logically means resonances potentially compete with music for attention and should be avoided so I could concentrate on the music. Or if I measure my speaker, see crazy dips and mild peaks, perhaps investigate the peaks first. Or, if one is concerned about lobing, perhaps it's not as important as it looks on the graphs, or if you want to hear it perhaps go into super hyper mode in order to evaluate whether it is important difference or not compared to casual mode, and so on. So the tests are something that gives perspective, helps to relate graphs to perception and provide foundation for logic in order to improve on things, to improve listening skill, and the system should improve with it.You are right some small deviations are not that important in the end, especially when there is nothing to compare against, hearing resets quite fast in this sense. I have been tweaking my system for few years and high mids is something that I seem to be very sensitive to, if too high it feels nasty, and adjusting DSP live even half db fix makes it better. Whether I hear it next day, I don't know, perhaps, perhaps not, current balance seems nice. But so did previous DSP preset, which I noticed was too bright trying the system in another house with very different acoustics, system balance changed quite dramatically. Rebalanced it there, and the new balance is fine at home as well. I don't know what was it, but if I didn't believe my own hearing when something is off, I'm quite sure the sound wouldn't get any better, just different. The tests gave me some confidence to trust some things on my hearing so it was quite fast fix to rebalance. I haven't yet figured out why the balance seemed so weird in the other house, what in the acoustics made it so apparent, or hidden in my own place. It's all just massive continuous learning opportunity
Last edited:
That is basically what I was trying to say with far to much text
In some of these cases "different" is even far to much of a strong word. (understatement)
We can all get super philosophical about that as well, because we have no idea what the intended sound reproduction should or must be (if even at all), or what the original mixing people had in mind (if at all).
Maybe that little peak was even intended as sound reproduction?
Who knows.
The only thing we know is what we personally like in the given room we have with equipment and system we use.
Reproducing "what was originally on the record" is by absolute definition not possible.
(what does that even mean? rhetorical question, don't answer)
Use the same system in a different room or move the same setup in the same room, and the whole reproduction quality has totally changed.
Anyway, this is going far to offtopic
YEah good sound in this sense is a moving target, but I have condenced everything to a relative target to reach toward: problem free sound, what ever that is.
Problem free sound contains the whole thing, like room, and of course the records and so on, but the recordings we cannot do anything about, except compensate if one wishes, consider it problem or not. Favorite records are great, no matter what the playback system, right
I think it's quite easy to detect which sound is due to system and which from recording. By system I mean the playback system, the room, and auditory system which varies in short and long term. Listening skill just helps to dig deeper, notice more subtle things within the system, but also quite apparent things I never paid any attention to. The most annoying things get noticed first, fix those and it's problem free system with the present listening skill and feels fun to listen to, nothing takes attention from the music Until one day one starts to pay attention to some feature with the sound, which means listening skill has developed and new challenge ahead try and come up what it is and what can you do about it. Hence I like to say our systems are only as good as our listening skill, listening skill needs to develop before our systems can.
good sound in this sense is a moving target, but I have condenced everything to a relative target to reach toward: problem free sound, what ever that is.Problem free sound contains the whole thing, like room, and of course the records and so on, but the recordings we cannot do anything about, except compensate if one wishes, consider it problem or not. Favorite records are great, no matter what the playback system, right
I think it's quite easy to detect which sound is due to system and which from recording. By system I mean the playback system, the room, and auditory system which varies in short and long term. Listening skill just helps to dig deeper, notice more subtle things within the system, but also quite apparent things I never paid any attention to. The most annoying things get noticed first, fix those and it's problem free system with the present listening skill and feels fun to listen to, nothing takes attention from the music
Until one day one starts to pay attention to some feature with the sound, which means listening skill has developed and new challenge ahead try and come up what it is and what can you do about it. Hence I like to say our systems are only as good as our listening skill, listening skill needs to develop before our systems can.
Last edited:
Hello hifijim and All,
I am attempting to sort out this 1.2 X the WL center to center thing in a way that it makes sense to me.
I believe that there are more variables at play than the crossover frequency wave length and the C to C distance.
When you apply the 1.2 X the WL center to center thing to the crossover between a woofer and mid-range driver it becomes a bit of a head scratcher. The center to center distance turns into 3 or 4 feet quickly. (a programmer’s sanity check)
When you make the assumption of point source drivers it adds a bit of perspective. *
Piston drivers are not point sources. Drivers have a coverage angle that narrows with increasing frequency, as the frequency increases they begin to beam. As the diameter of the piston driver decreases the frequency of beaming increases. In terms of frequency and octaves, as the frequency doubles the DI increases sharply above a critical driver diameter to frequency ratio.
Is 1.2 just a happy spot for the 6 1/2 inch driver and tweeter that you selected? What is the 1.2 C to C theory?
* Tolvavn.com/xdir has a widget that plots coverage lobes at a input crossover frequency and a input C to C distance between the drivers. Driver diameter is not variable that is included in the calculation.
Thanks DT
Woops... Sorry for the confusion, and you are correct: 1.2xWL at woofer/midrange frequencies (200 - 600 Hz) are huge, and unnecessary.
The best CtC distance is 1/4 WL or less. Going a little larger is still acceptable, up to about 1/3 x WL But if that kind of close spacing is not possible, then the next best option is 1.2xWL.
So at typical woofer-midrange crossover frequencies, it is easy to get 1/4 x WL or less CtC spacing.
Again, sorry. I should have been more clear.
j.
I am not sure I can explain it adequately. Three years ago I was reluctant to accept it because it looked so strange. Only after a lot of simulation, followed by building and testing a prototype, did I accept this strange looking layout.
I can show you an example of a speaker with and without the 1.2XWL spacing between mid and tweeter.
This speaker I made a couple of years ago, and it was the first one I designed with an emphasis on managing directivity from the earliest stages of design.
The baffle around the tweeter is small. It has 1.2xWL CtC spacing between the mid and tweeter, assuming a crossover frequency of 2k. And yes, it looks a bit odd. The crossover between woofer and mid is 300 Hz, and there the CtC spacing between the lower woofer and the mid is about 1/3xWL.
Here is the 6 pack charts for this speaker
The DI curve (dashed red line) has no big jumps or reversals. It stays within +/- 1 dB from 400 Hz to 6k. This allowed me to have a both a flat on-axis curve and a very nice sound power curve, which also means the ER curve and the Predicted In-Room curve (dashed black line) are very nice as well.
Now i will simulate what would happen if I positioned the mid higher on the baffle so that it would be touching the tweeter.
Now the DI curve has a 2 dB hump at 1.8k. The ER and PIR curves have a dip at the same frequency, and sound power curve has a wide depression from 1k - 3k. Notice the vertical polar response, this is where the change is. There is a deep null aimed at the ceiling at 2k, and this causes a drop in the ER and power curves. Now I don't want to make too much of this, it is a small change, and this kind of change has a small audible impact. Overall, the objective measured performance is still very good, and with the right positioning, the second hypothetical speaker could sound the same as the first. But I believe the first speaker (the real speaker) will be easier to position in the room, and less sensitive to placement.
Does this make sense? Sometimes I fret about providing too much detail, or not enough.
j.
Attachments
I take your concern seriously. Before I settle on this driver, I am going to do a subjective listening evaluation in the test baffle. This has the right width and edge profile, and the correct box shape for the internal midrange enclosure (as of now anyway). I will use DSP to provide some BSC and I will apply a notch at 6.5k to mimic the RLC notch I would use with this driver.
I didn't ask the original question, but you explanation was quite helpful to me.
I'm rooting for the CRC. I looked into it last year and the only opinion I found was wolf's, who had heard it once or twice but not used it himself. So it could be poor implementation. Since then I've seen two builds (seen on the Internet, not heard,) where it was the designers' favorite midrange, but I don't know how many speakers they have built. SB Acoustics' engineers seem to be pretty good so I'd love to see a week executed design. And better you than me be the Guinea pig
My thoughts as well. It is priced a few $ more than the SB15CAC, so they must think highly of it, otherwise they would just pull it from the production line.SB Acoustics' engineers seem to be pretty good so I'd love to see a week executed design. And better you than me be the Guinea pig
Hopefully my design is a "well executed" design, rather than a "week executed" design...
Hello hifijim,
Thanks for the reply.
The reason that I asked the question is that you do a good job of providing real data in addition to the theory.
The Tolvan widget shows serious nulls in the off axis coverage plot with 1.2 C to C distance. It is my estimation that the real driver diameters fill in the coverage nulls and provide a much wider coverage angle that may even be a smoother DI transition than coaxial drivers. * The the driver outputs sum to provide a, if you plan it, controlled coverage angle transition. I am of the opinion that crossover frequency and driver diameters matter.
Thanks DT
* https://tolvan.com/index.php?page=/xdir/xdir.php (my previous link had a typo, sorry)
I don't really agree with the 1.2x center-center from a practical point of view.
I am extremely familiair with the concept, very recently it was being brought up on a different forum as well.
This is based on nothing more than just a formula for the path-length difference = time difference between (two) sources as function of angle.
There is a specific point when these two sources cancel each other out = lobing.
With this specific factor in theory we get a more even "spread out" of the lobes, meaning a more even power response.
That number is all great in theory, but as soon as you change the delay by anything, like group delay from filters, that number will and shall change.
Not only that, but you really need to look at a 1 degree resolution to make sure you didn't make some deep null somewhere else.
Something you won't see with 10 or 15 degree steps.
(just simulate or measure that and you'll see what I mean)
It still totally dismisses the importance of the vertical directivity to begin with.
There are some extremely good loudspeakers out there, like a D&D 8C or so, and I have never heard a single person complaining about the vertical directivity.
More importantly, practically it just restricts the vertical listening window.
But as long as you position your loudspeakers properly, I don't see how that is much of an issue?
If you don't want it, just pick a nice coaxial unit, there are a couple of 5 inch ones out there.
I doubt if people are really gonna notice any significant differences though
I am extremely familiair with the concept, very recently it was being brought up on a different forum as well.
This is based on nothing more than just a formula for the path-length difference = time difference between (two) sources as function of angle.
There is a specific point when these two sources cancel each other out = lobing.
With this specific factor in theory we get a more even "spread out" of the lobes, meaning a more even power response.
That number is all great in theory, but as soon as you change the delay by anything, like group delay from filters, that number will and shall change.
Not only that, but you really need to look at a 1 degree resolution to make sure you didn't make some deep null somewhere else.
Something you won't see with 10 or 15 degree steps.
(just simulate or measure that and you'll see what I mean)
It still totally dismisses the importance of the vertical directivity to begin with.
There are some extremely good loudspeakers out there, like a D&D 8C or so, and I have never heard a single person complaining about the vertical directivity.
More importantly, practically it just restricts the vertical listening window.
But as long as you position your loudspeakers properly, I don't see how that is much of an issue?
If you don't want it, just pick a nice coaxial unit, there are a couple of 5 inch ones out there.
I doubt if people are really gonna notice any significant differences though
Last edited:
That is not how marketing works (that's not even a joke lol)
If you want to pick a 5 inch SB, go for the NBAC.
One of the best performing 5 inch drivers out there, even compared to MUCH more expensive drivers.
The other variants are not performing any better.
The NRX2 is close second, for slightly less money.
Member
Joined 2003
Overly simplified polar slice at a single frequency like Tolvan Xdir shows is a very poor method of evaluating speaker performance. Overall power & DI chart with real measured data is the best way. Fortunately evaluation of driver separation in VituixCAD is easy, just load in properly measured data, hammer in a crossover to perfection with active components, and adjust driver separation via y dimension. Change crossover frequency and repeat.
As far as the 1.2 - 1.4X "rule of thumb" goes, if it hasn't been mentioned already, it's important to note that this is a rule intended for a "standard" tweeter with flat faceplate, normal flat baffle and "standard" 4th order acoustic rolloff. Rule can change somewhat with complex baffle, waveguide/horn, etc. So best solution is just to aim for balance of overall Power & DI regardless of the situation. Main takeaway should be that placing drivers as close together as possible is often not the most ideal arrangement, and in some cases a very poor choice. Staring only at the simplified polar chart of Xdir you may come to different conclusions.
I agree the NBAC is a real gem. But I have already built two recent designs with the SB15NBAC or SB15CAC as the midrange driver. I want to do something different.
Well run companies are not that common. In the US, I estimate that at any given time, about 25% of publicly traded companies are "well run". And the list has a lot of turnover. It is very very rare for companies to be well run for long stretches of time (decades). Very rare indeed.
In well run companies, that is exactly how product placement (i.e. marketing) works. Well run companies value their brand highly, and it is an unacceptable risk to brand reputation to sell high priced turds just because they have inventory to work through... A well run company would sell its turds for a low price to customers who value price more than performance.
Well run companies are not that common. In the US, I estimate that at any given time, about 25% of publicly traded companies are "well run". And the list has a lot of turnover. It is very very rare for companies to be well run for long stretches of time (decades). Very rare indeed.
You assume that companies are aware of having turds.
There is also something like ignorance.
Truly believing in the idea that something is incredible, but it's actually just a plain old turd.
But anyway, I just read your message again and we mean the same thing.
I was just reading it wrong lol