Im in the process of still tuning my new system. and came across something very strange tonight.

Firstly, my left and right channels are correct and are in Stereo for both the mids and tweeters

secondly my headunit I have to adjust the time alignment values manually (W200 + H701)
from the alpine manual, the furthest speaker - current speaker
then divide by speed of sound and x 1000 to get the delay on each driver

first question... would this formula apply to all speakers? as tweeters play higher frequency, so would that mean i'd have to take that into account with the distance between me and the speaker? i.e would the sound travel faster in some way? as its a high frequency?


2nd question
when i sit in the passenger seat, I have adjusted the speakers by the above formula, and the sound appears to come from the front of the dash on the passenger side which is great

problem is... on the drivers side, I have done exactly the same, as the above method, and the sound seems to be dragged to the left of the car?
yet i measured the distances and followed the above procedures
So have reversed these numbers and tweaked them by ear, which now results in the sound directly coming from above the steering wheel area which im satisfied with.

what would cause this? does the steering wheel and/or other objects effect the actual distance, reflections?
And if it was a phase issue, I would have the same problem in the passenger seat?


Have seen Timm3hs thread too, and it is an interesting read.

The speed of sound is the speed of sound. The only real factor that plays in its calculation is density. Temperature slightly, but mainly density. FYI, it is 343m/s at sea level (101.3 kPa) and 25deg C.

Frequency has absolutely no bearing on the speed of the sound. A wave hits your ears at exactly the same time, whether its 20Hz or 20kHz. What does change though, is the energy contained within that wave.......

You'll never have 'phase' delineation's in such high frequency (the tweeters) because the wavelength is simply too small for your ears to discern a difference.

So to answer your first question, that formula would apply to ALL speakers. All its doing, is converting distance to time, and then from seconds to milliseconds. By taking one speaker from another, all you're doing is calculating the relative difference in arrival times for each speaker.

To answer your second question, its all relative to the speaker that is farthest, and where you take the measurements. Plus, such a technique is only a guide and never nails it, because you'll notice that the distance between one ear and a speaker is different to the other ear..... so which ear do you use....?? Simply put, its a close guess to get you into the ballpark, so your ears can take over.

There are easier and more accurate ways to do it, but you'll need a microphone and a laptop!!

whats a laptop ??? *shrugs*

awesome thanks

I do have a microphone and a laptop currently
borrowing a m-audio preamp and behringer mic

If you have a Microphone and a laptop, its really simple to nail it first time.

Simply mute every speaker but one, and send impulse signals through the driver. Graph in the time domain, the response. Note when the response starts, and TA to that (relative to the other speakers). Note what direction the response moves, and change phase to that.

Seriously, takes like two minutes and I've always found it hits the right cue every time.

what software do you use to measure this with?

I used a trial verson of WinMLS to do it, was great. My trial period has run out now, so am trying to find another program to use to do it.

I use SMAART (http://www.eaw.com/products/software/index.html), but there are others.

WinMLS (http://www.winmls.com/), ARTA (http://www.fesb.hr/~mateljan/arta/) and Speaker Workshop (http://www.speakerworkshop.com/) are other choices. I really like ARTA, and Speaker Workshop is hard to use, but free.

There is also a fantastic add-on for Adobe Audition made by Angelo Farina (Aurora?) that works extremely well too. (http://pcfarina.eng.unipr.it/)

~thematt~ Have you bee able to get ARTA to accurately read impulse response? I was trying last night, but I can't see to get it to read acurately.

I would expect that it would require the loop back refernce like WinMLW, whcih I have built in to my cable, yet, when testing and changing the TA on the channel I couldn't get any variance in the impulse timing. With WinMLS I could clearly see it move as I changed the TA. But ARTA doesn't appear to perform the same.. I'll be reading up on it today.

I found the 'step' function in ARTA useful though, even though it seemed to be overly smooth on the low end response.

RTFM... I'll have another play tonight this time with some more boxs ticked etc, hopefully can get some useable data

x2 for Aurora.

I use it now with Adobe Audition 3.0

older versions of Aurora work with Cooledit (predeccessor to Audition) which I have seen for freeware download.

To follow up I managed to get ARTA Impulse Response to work well, for a free program it does all I require at this point in time. The only thing lacking is the ability to save data.

haven't had a chance to try anything due to the brillant weather in Auckland and the fact that I dont have a garage
but as soon as a I get a chance I will