Speaker design

[EmergencyDpt]

Hi guys,
What's your favorite program for designing speakers? How do you test the actual speaker? I am building my Oliva's and would like to put together a graph on the response of the system and as I make changes to it, I want to have corresponding improvements in the data.
So what would you guys suggest. I have a fairly good microphone from sure, well at least it doesn't suck
I also have a program called SYNRTA that can give a graph of the room response. Is this a valid test of the speaker? I'm just using my laptop sound card to run the program on. Do I need a better soundcard? The laptop has a Realtek High Definition Audio card if I'm reading it right.
What if I took the speaker into the closet or other muffled area and ran the white noise program and did some comparisons. Do you think that would give me some meaningful data?
I welcome your input.
Thanks,
Spence

[just a guy]

What's your favorite program for designing speakers?

Depends on the project. I use mostly hornresp now for everything, even stuff I used to use winisd for. MJK's software has a couple of features that hornresp hasn't got, but I don't have it. Akabak does everything.

How do you test the actual speaker?

Depends what you want to test. Your program, as well as a few others will probably get you frequency response which is all most people care about. Other (free) programs to do this and more include arta (in a bundle including limp - for measuring t/s parameters), room eq wizard (very full featured for a free program) and others.

Depends how serious you want to get ($$$) and how acurate the results need to be. Laptop soundcards are probably not the best bet for ultra accurate readings. At the very least you will need to calibrate the soundcard and the mic (and make sure you know what the mic is capable of in the first place, or it could lead to misconstrued data. IIRC even the behringer measurement mic needs some sort of calibration)

What if I took the speaker into the closet or other muffled area and ran the white noise program and did some comparisons. Do you think that would give me some meaningful data?

Absolutely not. The data that you need will dictate the type of testing you need to do, but testing in a closet is not going to tell you anything except how the speaker tests in a closet.

If you want to measure the SPEAKER OUTPUT ALONE, measure using a program (and away from any boundaries like buildings, trees, fences etc - extra points for soft surface like grass below) like your SYNRTA (which I am not familiar with, I just assume it works adequately) or one of the other programs I mentioned above. There are ways to get accurate results inside but I am new at this too, and I'm not sure how to do it. This test shows how the speaker performs by itself (almost anechoic) and can be directly compared to any simulations that were done.

If you want to measure SPEAKER OUTPUT COMBINED WITH ROOM RESPONSE, measure in the listening room, from the listening position, with the speakers set up how they will be when finished. This test shows how the speaker operates in room, and can be compared to simulations that were done, and to the speaker output only measurements to see how the room is contributing to the sound, and what the in room response looks like.

No matter which program you use, you will have to become proficient with the features and settings and know to set up so that you are actually measuring what you intend to be measuring.

OTOH, I'll be doing some measurements soon with my cheap laptop soundcard and the built in mic. These won't be worth much on their own, but if I do white noise testing in an absolutely quiet ambient background outside at night, I can mentally establish how far off the measurements are (due to CHEAP measuring tools) and then I can (kind of) mentally extrapolate a corrected response. Like I said, it all depends how much you want to spend and how accurate the results need to be.

[Scottmoose]

Just to add, nearfield measurements aren't much use for any of the double-vent designs, because they need ~8ft distance before the wavefronts start to integrate.

For a quasi-anechoic response, you'd probably have to do it outside, with the floor heavily damped. In-room measurements are useful for yourself, but less so for anyone else, as all rooms are different.

[EmergencyDpt]

The reason I said closet is because with all the clothing in there, it's pretty muffled. Not much sound reflection. Outside is even better.

[just a guy]

I think I made a mistake when I said white noise, IIRC white noise = virtually unusable information (linear frequency representation). It's been months since I've seen these programs and don't currently have any installed, but I think pink noise is the one you want (exponential frequency representation).

But it's always best to make sure for yourself since I'm to lazy to confirm the accuracy of my statements at this time.

[EmergencyDpt]

Pink noise it is.

[just a guy]

Post #4
The reason I said closet is because with all the clothing in there, it's pretty muffled. Not much sound reflection. Outside is even better.

I know why you said it but it doesn't work that way. You literally cannot fit enough stuffing into a closet to simulate anechoic conditions. It's not possible. What you will be measuring will still only be a mildly stuffed tiny room and the results will be bad and they won't help you, unless you are only trying to determine broad trends in performance.

Anyway, you might as well try it for yourself. You will need to do hundreds of measurements in different locations before you begin to become proficient in the art and science of measurement, so measure everything you can, just so you know.

[EmergencyDpt]

Originally posted by just a guy


You literally cannot fit enough stuffing into a closet to simulate anechoic conditions. It's not possible.

You haven't seen my closet!