do poor simulations sometime miss reality?

[freddi]

just curious if sometimes a less than optimum simulation will cause one to reject a driver-horn combination which would sound good and measure better than the sim?

[GratefulTony]

Good question-
Or perhaps a flattering simulation could make a cabinet seem better than it actually is?
Experts? How accurate are those simulations?

[hitsware]

Dunce and Merlin concur:

If you understand the 'simulator' well enough...
I.E. Know a little bit of the math AND
Believe in the results produced
(perhaps at the expense of a few experiments)
It will work !
Sorta metaphysical but true.
It would be a noble and enlightning experience
to write a simulator
pretty doable (I think)? in Excel
(I could never get the 'graphs' to work)

"If you write your own .... You can't complain about the product"
Frederic Karlsonasky ........ Mega

[peterbrorsson]

I'm no expert with sims or much theory behind speaker building. Just have listened to music!!!

IMHO the way of using/creating/reading the result from a sim matters more than just to have a sim for its own sake.

Add the fact that the best measuring speaker in an aneoic(spelling?) chamber, can sound like a chain saw with TBC, throw in a mist horn and descripton is complete in a "normal" room.

A little harsh description maybe but I have had some commercial speakers that were tuned for average rooms, not some artificial environment. Meaning that they were playing with the room, not against it. Stig Carlsson, RIP (Carlsson orthoacoustics) and Bo Hansson (Rauna of Sweden) tune their speakers in this way.

In reality this means a poor measuring in an aneoic chamber due to the fact they have a shallow roll off in the bass to compensate for the room gain. No boom, just tuneful fast bass. Only problem is that sales guys screams after ruler flat graphs!

I've got a feeling that nowadays most speakers has to be at least 1m from back and side walls. Add bass traps, diffusors and what else to make it function. Sure some people like to have it like this, just not my cup of tea.
Also I like to LISTEN and be MOVED by the music, not count the amount of nose hair of the artist as some of the modern speakers is far to analytic for my taste. Too flat graph?
Clinical, sure. Music, not.

And most important, it's the person listening that has the final word, not the sim or the graph.

Maybe this went OT in the end but I think it matters regarding the topic

[Nelson Pass]

The definition of a poor simulation is one that misses reality

Realistically, the best use of sims is as a starting point. You have
to follow it up with a build and with time put into adjustment
guided by measurement and listening.

It is possible to have a sim which does not measure out as
predicted.

Even if it does, that is not a guarantee that the speaker (or
other audio device) will sound as good as it looks.

Sims/Measurement/Listening are all tools that any good designer
uses in rotation toward improving the sound.

As Edison said, invention is 10% inspiration and 90% perspiration.

[Scottmoose]

Couldn't agree more. Simulation software is a useful tool, one of the most useful available, but it doesn't design an enclosure for you. You have to know what it does, and what it does not do, and how to interpret the results it gives. Optimum itself is subjective anyway, because what one person considers optimal, another will regard as substandard.

[MJK]

Some random thoughts :

A simulation will never tell you how a speaker will sound.

A simulation can be used to help make intelligent choices about a potential driver and how it might work in a potential enclosure. It will quickly assess and narrow down the choices to only a few.

If the simulation is done correctly, measurements will match the simulation results over the applicable range of frequencies.

A simulation is only as good as the person performing the analysis. A simulation tool will make a good "engineer" really good and a bad "engineer" really bad. "Engineer" is used to describe the person running the simulation because that is in essence what a lot of engineers do for a living, myself included. A simulation tool cannot make up for a lack of knowledge. You need to know how to run the siumlation, what assumptions were made in the simulation, and how to interprete the output.

Running simulation is not that tough, getting accurate results is that tough.


There are other options for designing and building speakers. You can do some rough sizing on the back of an envelope and make a quick sketch of what you want to build. This requires insight and experience. The results will be somewhat hit or miss and you may need to revise and rebuild the enclosure a few times learning as you go along. Eventually somebody who knows what they are doing will probably produce a good speaker design.

Personally, I am not that smart. To me, a simulation tool will short cut that process and allow virtual builds and results quickly so in the end you hopefully only actually build the speaker once. Is it fool proof, absolutely not.

[ronc]

Yes.

As Martin stated a sim is only as good as the program and the person doing the analysis. Sometimes its an obvious falt sometimes its not. Sometimes its out of the control of either the engineer or the program.
Example! Years ago at Boeing when there was punch paper rolls to run programs we had an automated rivet robot that would run great till a given stage then go haywire. It would shoot rivets into mid air away from the wing spar and was totally out of control. It was a definate mind stumper. A tech noted that a repair had been made in the roll approximatly in the time line where it went crazy. Well the end answer was when someone was making the repair the tape had been flipped over so all the controller recieved was jibberish.
Sometimes its not the program or the engineer, sometimes its an outside unknown value.

ron

[hitsware]

I think something like this would be a good stepping off place for
a homebrew 'emulator' :

http://www.peda.com/grafeq/

[planet10]

Originally posted by freddi
less than optimum simulation will cause one to reject a driver-horn combination which would sound good

The BIB.

dave