Kevin Haskins said:
Glad to hear that, Kevin. I'm currently doing a postdoc at IU in Psychological & Brain Science department. Got my Ph.D. at Ohio State. These days this new hobby occupies my mind that should otherwise be used for academic excellence!
I visited your website. You're doing this stuff as your full-time job? I envy you! BTW, the Erse caps look interesting. Are they good?
P.S. Bloomington is a good place to lose virginity!
Jay_WJ said:
Oops your bad....
Unlike you Mr. Kim, Chris D'Alessio builds and tests the designs he creates. The final designs Chris publishes on his website are based on the results of both listening and instrument testing, they're not just sims created by someone who's never built a loudspeaker in their life..
http://www.eldamar.net/audio/rs150mtm/#seas27tdfc
Hey Mr. Thomas H.T.G.,
Even with measurement, we have to tweak an XO using our ears. I HAVE good easrs trained by classical voice music. I'm convinced that sims can lead me to the point that will need my ears for minor tweaks. In this tweaking phase, sims can also be a useful tool. At this point, I'm pretty confident that I'm not going to fail even without a measurement setup. BTW, could you please get out of my precious thread? I don't want to see your post here any more. I MEAN IT. It's enough for me to see your name here and there.
Even with measurement, we have to tweak an XO using our ears. I HAVE good easrs trained by classical voice music. I'm convinced that sims can lead me to the point that will need my ears for minor tweaks. In this tweaking phase, sims can also be a useful tool. At this point, I'm pretty confident that I'm not going to fail even without a measurement setup. BTW, could you please get out of my precious thread? I don't want to see your post here any more. I MEAN IT. It's enough for me to see your name here and there.
Jay_WJ said:
I've been a member here for 6yrs so it's certainly no secret that I moderate on HT-Guide.
Since you've never built a loudspeaker in your life, isn't it little difficult for you to fine tune them 'by ear'..?
BTW if you want to limit participation in threads, you better start your own forum. I promise I won't join....
Come on guys...... no need for that.
Thomas is right though... .you have about as much chance of getting it right by ear as I do of giving you advice about modeling the female brain.
Our ears are easily fooled and even the best trained people are easily outperformed by a $0.50 Panasonic mic capsule calibrated correctly. ;-) Interpreting measurements is not always trivial though. That is something that takes looking at the measurements and trying to tie them back to whats going on physically. Thats part of the fun. That and trying to discover what things really matter and how much.
All the Erse parts are good. I like em! I'm going to be stocking some 630V parts soon. The Erse inductors are good too. Good company and they manufacture here in the USA (Ohio State) so I don't have to go through the contortions of dealing with currency brokers, customs and overseas shipping. Just a UPS truck away.
In terms of the business, its how I earn my cash but the IDEA of running an audio business is more romantic than the reality.
Thomas is right though... .you have about as much chance of getting it right by ear as I do of giving you advice about modeling the female brain.
Our ears are easily fooled and even the best trained people are easily outperformed by a $0.50 Panasonic mic capsule calibrated correctly. ;-) Interpreting measurements is not always trivial though. That is something that takes looking at the measurements and trying to tie them back to whats going on physically. Thats part of the fun. That and trying to discover what things really matter and how much.
All the Erse parts are good. I like em! I'm going to be stocking some 630V parts soon. The Erse inductors are good too. Good company and they manufacture here in the USA (Ohio State) so I don't have to go through the contortions of dealing with currency brokers, customs and overseas shipping. Just a UPS truck away.
In terms of the business, its how I earn my cash but the IDEA of running an audio business is more romantic than the reality.
Kevin Haskins said:
I agree with you. That's why I didn't want to design an XO relying SOlely on my ears albeit it's impossible. As you know the method I use is not without measurement---I use Zaph's infinite baffle measurement data which I found very accurate and consistent. So, the question is about the difference between infinite baffle data and in-box measurement. What matters is the accuracy of baffle step and diffraction simulation (bass simulation is much less important and can be predicted much better), and phase estimation via Hilbert transform. That's why I tested this method using many existing designs until I was convinced it's accurate enough for a simple 2-way design with a relatively low XO point.
BTW, there's no reason that it's difficult for me to have a measurement setup. I just want to try this for fun! Eventually, it's ME who'll build the speakers and tweak them to MY taste and be satisfied. Right? That's it. Wish me luck!
Yeah, there are not many choices for budget audio grade capacitors other than Dayton and Solen. I'll try the Erse. I like their +/- 3 % tolerance.
- Jay
Jay_WJ said:BTW, I DO believe that baffle diffraction is MUCH less complex to model with a reasonable amount of prediction error than a female brain. Don't you agree?
Yah got me there.
But your not considering the difference in impedance of the driver. That enclosure is going to present a different acoustical loading than the open baffle and change the impedance curve of your midwoofer. Also... the frequency/phase data is effected by more than baffle edge diffraction. I've measured drivers on the IEC baffle that looked one way, move them to an enclosure and you got a different situation that isn't always easily correlated to baffle diffraction, at least not in a way that I'd recognize.
Maybe you have this entire baffle diffraction thing figured out better than I but I have some pretty spiffy software. ;-) About the only baffle diffraction I'd be semi-comfortable simulating is the 2Pi-4Pi transition. Even that changes based upon loudspeaker placement so the best method I've found of voicing the loudspeaker is by trying it in different room s & locations.
The tweeter won't change from the IEC baffle measurement near as much. That will be a little more predicable in terms of models.
If you have a burning desire to write simulation software what we really need is someone to work on soft-part emulation. There is FineCONE from Loudsoft but its fairly limited. The other choice is a fully blown FEA with a mechanical modeling tool. Opportunity is knocking for someone with the math and computer skills to develop a good soft-part modeling tool that will accurately model cone, dust-cap and surrounds.
"But your not considering the difference in impedance of the driver. That enclosure is going to present a different acoustical loading than the open baffle and change the impedance curve of your midwoofer."
I'd be concerned if I did a 3-way and crossed a woofer low, but impedance change by enclosure should barely affect a range around the XO frequency I use for a 2-way.
"Also... the frequency/phase data is effected by more than baffle edge diffraction. I've measured drivers on the IEC baffle that looked one way, move them to an enclosure and you got a different situation that isn't always easily correlated to baffle diffraction"
That's exactly the effect of baffle diffraction. If all other conditions are the same except the baffle (infinite vs in-box), that's what we call the baffle effect. Right? Of course, there may be some other things that change (e.g., group delay) due to different cone movement in box versus free air, but FR around the crossover point of a usual 2-way design shouldn't be affected significantly by these factors as far as I know. That's why people can generally use the same design (XO and baffle) for different box alignments for bass. Do I miss something?
All in all, I've already considered many issues including the ones you mentioned, say, FR and impdance affected by on-axis vs off-axis and infinite baffle vs in-box measurement, phase prediction via Hilbert transform, acoustic driver offset, and so on. My conclusion is that for a 2-way with a low XO point, these things are pretty robust to prediction errors, at least for a particular design of mine.
I'd be concerned if I did a 3-way and crossed a woofer low, but impedance change by enclosure should barely affect a range around the XO frequency I use for a 2-way.
"Also... the frequency/phase data is effected by more than baffle edge diffraction. I've measured drivers on the IEC baffle that looked one way, move them to an enclosure and you got a different situation that isn't always easily correlated to baffle diffraction"
That's exactly the effect of baffle diffraction. If all other conditions are the same except the baffle (infinite vs in-box), that's what we call the baffle effect. Right? Of course, there may be some other things that change (e.g., group delay) due to different cone movement in box versus free air, but FR around the crossover point of a usual 2-way design shouldn't be affected significantly by these factors as far as I know. That's why people can generally use the same design (XO and baffle) for different box alignments for bass. Do I miss something?
All in all, I've already considered many issues including the ones you mentioned, say, FR and impdance affected by on-axis vs off-axis and infinite baffle vs in-box measurement, phase prediction via Hilbert transform, acoustic driver offset, and so on. My conclusion is that for a 2-way with a low XO point, these things are pretty robust to prediction errors, at least for a particular design of mine.
You should be fine with impedance at 1.5-2K. If it where a 3-way crossing over under 500Hz it would be a much bigger issue.
Setting levels between drivers is another consideration but if you keep the crossover simple that isn't as much of an issue. Just do it on the tweeter.
What are you using to model the crossover? Seems like a lot of work for something that is easy to do with a measurement. ;-) But hey... if your having fun go for it.
Setting levels between drivers is another consideration but if you keep the crossover simple that isn't as much of an issue. Just do it on the tweeter.
What are you using to model the crossover? Seems like a lot of work for something that is easy to do with a measurement. ;-) But hey... if your having fun go for it.
Hmmm...........
Tweaking by ear and simulation, i.e. predicting the effect of what you
are doing, is far more effective than playing around with components.
FWIW with my test CD with sliding 1/3 octave warble tones and 30
pink noise bands I can tell an awful lot without measurements.
Having a pop someone elses methodology is pointless,
though I do occasionally criticise lack of any sensible methodology...
Add something useful or don't bother.
RJB
IMO simulation before you start is the only to choose 3 way drivers.
I despair at post like "I've bought these drivers - 4 way c/o please..."
/sreten.
Tweaking by ear and simulation, i.e. predicting the effect of what you
are doing, is far more effective than playing around with components.
FWIW with my test CD with sliding 1/3 octave warble tones and 30
pink noise bands I can tell an awful lot without measurements.
Having a pop someone elses methodology is pointless,
though I do occasionally criticise lack of any sensible methodology...
Add something useful or don't bother.
RJB
IMO simulation before you start is the only to choose 3 way drivers.
I despair at post like "I've bought these drivers - 4 way c/o please..."
/sreten.
Measuring below 500Hz isn't difficult. Its just different. You use near field response that is VERY accurate. There is nothing wrong with simulation of it either but to say that you cannot get a good low-frequency response measurement is just wrong.
I think we are getting off on a tangent. There is nothing wrong with simulations. But it takes me 10 minutes to boot up the computer and take extremely accurate measurements for two drivers for both FR/Phase and the impedance curve. You can then import them into software for all the simulation you desire. My only point is that if your using someone else's measurements your trusting the most critical data gathering point entirely to someone else. Your only method of checking your work is purely by ear.
Since a measurement system cost so little in relation to the time/effort and expense of the components, why wouldn't anyone use them?
I think we are getting off on a tangent. There is nothing wrong with simulations. But it takes me 10 minutes to boot up the computer and take extremely accurate measurements for two drivers for both FR/Phase and the impedance curve. You can then import them into software for all the simulation you desire. My only point is that if your using someone else's measurements your trusting the most critical data gathering point entirely to someone else. Your only method of checking your work is purely by ear.
Since a measurement system cost so little in relation to the time/effort and expense of the components, why wouldn't anyone use them?
You also made a good point, Kevin.
I think it would be ideal to use BOth simulation and measurement. Sims can particularly be useful if one has a suboptimal measurement setup (a well calibrated measurement system is difficult to have for infrequent speaker builders). By borrowing good, consistent infinite baffle data from other hobbist (like Zaph) and running a simulation, you can at least avoid or correct some horrible errors or mistakes in your measurement. Also, sims can be very useful in a preliminary design phase as said above.
In my case, I know my use of FRDC tools is a bit more wild since I want to substitue simulation results for in-box measurement. This is why I wanted to take enough time and be very careful in my sims. This will be a good experiment, anyway. And I going to measure the system eventually to check my work.
-Jay
I think it would be ideal to use BOth simulation and measurement. Sims can particularly be useful if one has a suboptimal measurement setup (a well calibrated measurement system is difficult to have for infrequent speaker builders). By borrowing good, consistent infinite baffle data from other hobbist (like Zaph) and running a simulation, you can at least avoid or correct some horrible errors or mistakes in your measurement. Also, sims can be very useful in a preliminary design phase as said above.
In my case, I know my use of FRDC tools is a bit more wild since I want to substitue simulation results for in-box measurement. This is why I wanted to take enough time and be very careful in my sims. This will be a good experiment, anyway. And I going to measure the system eventually to check my work.
-Jay
Kevin Haskins said:
"Ears are easily fooled"??? Who's ears are you talking about? 99% of speakers commercial or otherwise will not fool me into believing that there is a real cello playing right there in front of me if I close my eyes. And I've never claimed to have "good ears". I will say though, that to me, most speakers sound like crap and I find them irritating to listen to for more than a few minutes.
Measurements can be a useful tool, but, their importance is INSIGNIFICANT compared to how a speakers sounds. In the end, our ears ARE what is important when it comes to speakers.
Jay_WJ's methods stand a good chance of getting him a design that sounds better than many speakers you'll find out there for sale on a shelf, I'd be willing to bet.
RE: ThomasW ~ if he's been a member on here for 6+ years, I guess that shows that the moderators on here are a lot more reasonable and tolerant than some *cough* *cough* forums that seem to be run by jerks?
Which do you think to be the best tweeters (no wavequide etc) to cross low? I'm thinking project like Harbeth Super HL5 which has 8" plus 1". and 0,75" for wolfs.
XO should be around 1500 Hz.
XO should be around 1500 Hz.
That may be true but you need quite low sensitivity tweeter ( at least if you compensate baffle step loss). And low SPL tweeter means small dome which can't be crossed low. Circle...
Which do you think to be the best tweeters (no wavequide etc) to cross low? I'm thinking project like Harbeth Super HL5 which has 8" plus 1". and 0,75" for wolfs.
XO should be around 1500 Hz.
This can easily cross at 1500 Hz:
Dayton RS28F
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.