I have been playing around with the sound meter test and using a test tone CD i made up.
If you would like to do your own tests and post the results here , then here is the link to the test tone file.
Its a 46mb zipped file with 75 wave tones ranging from 42hz up to 10,079hz.

http://www.4shared.com/file/10875266/4840f...42hz-10000.html

Heres a link to a thread by powerband,it shows some helpful hints to get u started with your own tests.
http://www.caraudioaustralia.com/forums/in...showtopic=79624

Heres a link to a thread on phase and its effects on SQ
http://www.caraudioaustralia.com/forums/in...showtopic=78965

Any SQ related test is welcomed...........cors
Also i have created a template
Click to view attachment

Nice work zion, I'll add this link in for people that want to get serious with their Radio Shack meter.

http://www.rivesaudio.com/software/softframes.html

Also has a nice chart available for free download.

thumbs up corey...


good to see youve dedicated the monkey business to the car audio business and not the getting in trouble business

don't forget also, if you use the 'C' curve on the RS meter, to allow for the roll-off at the top and bottom of the frequency range.

I'll dig up a graph....





so what you see is the C curve drops off below 60Hz (down to -7dB at 20Hz)
also above 3kHz, a more gentle slope (-3dB@ 6kHz, -6dB@ 10kHz, -12dB@ 20kHz)

you need to correct this by adding the dB 'lost' at each frequency to your own measurement.

so, say you measure 96dB at 10kHz, that would really be 102dB, since you corrected the -6 from the graph.

somehow when u download the testtones they are grouped from 1 to 9,instead of starting with 42hz,eg the first one is 111 and they continue all mixed up till 9969 or something.Some how u have to re order them before u burn onto disc...Any brainiacs here can tell how to do that?.................cors

rename the tracks in order with a number at the start..
01. 42Hz
02. 63Hz
03. 80Hz
04. 100Hz
...
10. 315Hz...
etc.

Yes the zero is very important and if you go into hundreds you need two, like so - 001. What happens is the computer starts with 1 then goes to 11 then 111 etc

Good idea Zion and you put a lot of effort into your graph. I might post something up in a coupla days or so


cheers

AFAIK, that rives test tone disc has tones that compensate for this - will let everyone know when I get mine. Damn bank and their damn credit card expiry dates...

gee, I hope they label the tracks clearly then - most unusual if they do...

Click to view attachment

EDIT# yep my system has a supercharged midbass spike from hell#
ive set all my PEQ back to 0 for this test.
I thought i would set my TA first.
not sure where to start .........cors

I've tried a few more things and I have to get around to scanning the graphs but they look nothing like yours

What kind of sound level meter do you use Zion?

$80 jaycar sound meter.............cors

im changing the 160hz band to 125hz and making it -5 sharp
changing the 1.25k band to 1.6k and making it -3 meduim sharp

Click to view attachment

looks like you got something out of phase 1.4kHz and 3kHz left... midrange speaker perhaps?

and also 250hz

i checked my wires they all look + to+ and - to- .............cors

Zion, invert that one midrange and have a listen

They mean acoustic phase, not electrical...

As Fury mentioned, try inverting one of the speakers electrical phases, and see its response acoustically.

Zion can you reverse the phase in your HU? That's what I did yesterday and it is the magic that I've been looking for. Check my shark bite thread

cheers


Doug

i only have reverse sub phase............cors

Damn! that response graph is jaggedy!

Been pissing around with home audio lately and the response readings from a listening room are so linear compared to a car that it's disgusting! , and those HIFI nuts still complain! Good to see car audio nuts who spend time living with, and taming some seriously nasty response curves. Bring on the Luurve!

the 250Hz is likely cancellation, or interference from the steering wheel or something.

try boosting it +3dB. If it doesn't increase by 3dB then it's a null so you can ignore it until later if you have a spare PEQ.

acoustic phase - I find that inverting one of the mids is necessary (not sure why, might be something about the size of the average car... strange though, every car I can think of has benefitted from it.

I have the Dick Smith meter so I thought they'd be similar but my graphs come out like a rainbow. The reason for this is my meter is only accurate from 300Hz to 6kHz (or whatever) Here's a couple of the latest readings. Check the info on the graphs for more info

Click to view attachment
Click to view attachment

Comparing these graphs shows that the dips and peaks are due to cabin gains and nulls (is that correct terminology?) because in the first graph there are no crossover points

I hope there's something in there that can help

EDIT: The info on the graphs can be a bit confusing, read it carefully

cheers


Doug

i changed one side and i cant tell, someone else post there problems maybe i can learn from your mistakes .........cors
BTW i have gone down to 400hz crossovers, will post up a test sheet soon.

ideally, would the graph look smoother and flat? or would it be in a rainbow shape? cheap and effectice idea here. i dont know SFA about tuning etc, but this has me somewhat interested

ideally it would be flat and smooth, with a little roll-off like in the C-weighting curve of post#4. This assumes you are using a sould meter with C-curve.
Ideally it would have a slight rise below 500Hz, but generally smooth and even.

in-car response is seldom linear on a graph, due to the nature of the measurment in this case. The sound meter will add or subtract reflections from the main signal depending on arrival time to the meter's microphone.

The two ways to avoid measuring the reflected sound are using swept-sine method or gated noise method.

Ideally you would set the crossovers first to suit the speakers, then TA to suit the locations, then EQ as a final 'fix' for the resonances and nulls which are still annoying.

Keeping the volume down a bit would surely help too?
My first readings were almost into the 90's and it was terrible to listen to. Now I have the volume down so the highest readings I get are about 75dB. Have to find a very quiet place though which is hard to find. You find a place away from traffic and then a plane goes over every ten minutes!

It's occured to me that Zion must be using C-weighting and I use A-weighting (because its what shiny car suggested) I tried C-weighting before but it didn't make any sense to me but I might try it again.

im sure its on A weight but ......cors

well, A-weighting will give you a massive bump in the middle - like what your graphs show.
C-weighting will give a flatter looking graph, but really the idea is to get an un-weighted graph, so you really see what is what.

turning the volume down can introduce it's own problems, but should be not too bad. If you're going to use A-curve, I wouldn't go much below 90dB@1kHz though - you start to lose too much at the bottom end.

i had a look at the meter and i use c weighting on fast read and high meter level..............cors

fast or slow - you want to take the average reading, not the peak reading, as pink noise will always have peaky bumps in it - the average is what you set the level with.

i also noticed some tones load up in db when given time?............cors

It's quite funny actually - try searching on a thread i did years ago.
It was called "How Pink is your noise"

Ah yes! check my latest reading using C-weighting and compare to previous ones to see the difference

This graph shows the effect of reversing the phase on one driver. It doesn't matter which I reverse, midrange or midbass and still comes out the same I'm low passing the sub off the amp though which may make a difference. (Also bit of a problem the black and blue pen look much the same online)
Click to view attachment



Yes I find they can also go down. Takes a bit of time to decide what to pick sometimes. What I do is plot the graph in pencil then go through it again and dot the points in ink, then erase the pencil. By the time I've plotted a graph I've gone over it a dozen times (Thats why I'm really keen to get this over and done with!)
Here's a graph in the process
Click to view attachment


Here's another thing I graphed. Set the crossovers on one driver to "through" top and bottom and turn everything else off. I did this to my LH Midrange and Midbass. It shows you dips and peaks that are caused by the cabin and not by crossover points. Both drivers come out very much the same.
Click to view attachment


What occured to me after seeing those results was to set the x-over at 160Hz. I thought it might bring the peak down a bit. Btw I think in the first graph I had a bit of gain on the sub
Click to view attachment




http://www.caraudioaustralia.com/forums/in...c=47070&hl=
Mmm I think I remember reading that. At the time I thought to myself "I'm not gonna get into it that deep"

Here I am a year and a half later...up way past midnight...searching for that glory!