RTA-ing front stage to locate problems

[shizzzon 461]

I have yet to personally use an RTA for my own personal use so let's see if I get this right-

 

I want to use Pink noise because that is generated to play how our ears perceive sound.

 

 

When analyzing what pink noise looks like on a computer, it appears from 20hz-20khz, for every octave from 20hz, the amplitude drops 3db.

So, at 20khz, it's very low compared to 20hz but our ears must be very sensitive there(if you can hear it) to perceive a flat response.

 

 

So, when i go to RTA my front stage, is my goal to also (if it were perfectly possible) to achieve a full range reading exactly how pink noise is generated?  3db drop for every octave up?

 

In the end, I would assume that once one gets to that point as close to possible, you can either leave it set like that or, to the user's liking, start adjusting things from THIS point so you know exactly what you are doing.

 

What i will be doing is the following-

 

I have 4 8s, 2 tweeters in terrible locations..  The tweeters will be moved and replaced this summer anyways but i want to go ahead and do some RTA-ing so i can see at least how my 8's location is since it's not moving.

 

 

I will be doing a sweep on the 8s only.

I will be doing pink noise on the 8s only.

I will be doing a sweep on the tweeters only.

I will be doing pink noise on the tweeters only.

I will be doing a sweep on the whole front stage

I will be doing pink noise on the whole front stage

 

 

My midrange sweep consists of 100hz-5000hz

High range sweep consists of 1000hz-20000hz

 

The sweeps are around 5min long to slowly read small ranges. 

The RTA I will be using, the SPL-Lab RTA Pro, has an Octave adjustment down to 1\96th of an octave for look for very precise ranges in the frequency spectrum.

 

 

The amplitude range for this meter is 90-130db so i must play within this range for measurement results.

[altoncustomtech 773]

Damnit, I've got to get me a good measurement tool for doing the same thing.  I've not got any real dips in frequency response that I can easily pick out, but I have some peaks that need taming.  While I can get REALLY close, I've not yet been completely successful in taming the peaks completely.

 

I'm gonna have to look a little closer into this RTA.  

 

So far your plan seems solid to me.  I'm looking forward to your findings.

[shizzzon 461]

my x-over points are as follows-

 

8s- 400hz-3.5khz

tweeters- 4khz+

 

Slope, i cant remember.. it's either 18 or 24db.  i use a memphis 16-x03

 

I'm sure the response will look like chaos but i need to do this to learn and get some SQ in my system.

[///M5 2,833]

ONLY use the RTA to find anomalies. The sweeps won't help at all as the cabin is going to eff with those seriously.

Your crossover settings sound awful. I'd start by optimizing those. Make sure to include phase.

[shizzzon 461]

Ok. Im bout to have a batcb of pink noise tests uploaded in about 15 min.

I guess I wont show the sweep graphs.

All the pink noise pics were measured in stereo.

This is just to get a feel as what is going on.

There is about 397 things I can do without buying anything...

But based on what I post, I'll try and understand why things appear the way they do.

[ssh 924]

Anything you do now with an RTA is a complete waste because your current XOs and slopes are completely fucked. Work on the XOs and Slopes first.

[shizzzon 461]

Midrange only pink noise-

 

Tweeters only pink noise-

 

 

Full front stage only pink noise-

 

I don't know what's up with 10khz.. the computer i was using was hesitant around that area as it may not be strong enough to process live information.  Just a stock celeron.

[Quentin Jarrell 671]

I'd be curious to see what they looked like not installed in the vehicle

[shizzzon 461]

well, everything measured is off axis from the headrest.  

 

The tweeters will be a lot more on axis this summer but that's then.. i'm curious on just what's happening now for the time being.

[shizzzon 461]

The 8s at LEAST follow a pattern of pink noise... the tweets.. well, i'm screwed.. i dont even know where to begin with those...  unless their off axis is just retarded.

[Quentin Jarrell 671]

if they were tested outside the vehicle in approximated locations you would be able to tell if its just shitty speakers, cabin gain, reflections, xovers, power levels, ect.

[///M5 2,833]

Is that transient or averaged? What sort of windowing and what is your source?

[Impious 1,476]

One thing to remember.....a measurement is only good at the specific location you are measuring.  You might take one set of measurements, dial everything in....move the mic 3" to the left/right/up/down and the measurement looks like shit again.  And we don't "hear" from a single point in space like what you are measuring.  Your ears are on the side of your head about 8" apart, and your head isn't always in the same location.

 

Moral of the story.....you need to do spacial averaging if you want the measurement/tuning process to actually be useful and not just an exercise in futility.  If your measurement software can't do that, getting one that can would be a good idea.  ARTA I believe can do spatial averaging and is a free download. 

 

Second it would be helpful if your mic was calibrated or if you had a calibration file so you know if what you are measuring is actual system response or mic response.  You might be EQ'ing something that is an artifact of the measurement device and not the system.

 

Last, resolution to 1/96th is completely unnecessary.  We aren't very good at identifying high Q (narrow bandwidth) anomalies.

 

And like Sean mentioned, very helpful to know what type of measurement you are getting....if it's gated or not, and if it's gated what the window is.

[ssh 924]

Your XO points and slopes create "holes" in your FR. Between your sub XO point (speculate 80hz) to 400hz it's mayhem. At 130hz you're down by -30dB, and thats if you're using an 18dB per octave slope. 230hz, -15dB, etc. From the sub, you're down -21dB at 160hz with a 18dB slope. Since you're not using the same XO point between your mid and tweet you end up with another hole around 3750hz. Your graphs support this.

[Impious 1,476]

Also would be helpful to know what type of EQ you are using, assuming you are using one. 

 

But like the others have mentioned, fix the xover first.

[shizzzon 461]

Your XO points and slopes create "holes" in your FR. Between your sub XO point (speculate 80hz) to 400hz it's mayhem. At 130hz you're down by -30dB, and thats if you're using an 18dB per octave slope. 230hz, -15dB, etc. From the sub, you're down -21dB at 160hz with a 18dB slope. Since you're not using the same XO point between your mid and tweet you end up with another hole around 3750hz. Your graphs support this.

 

Yea i know this.  I wasn't but at the same time was expecting a mass collection of points, comments, etc to look over.  Too many to go through at this time though as i am going back out of town this weekend.

 

It will definitely take a while to fix the beginning of this process so i'll let it die for now until i can attack it.

 

Impious- i have no processor right now because i wanted to at least RTA first to determine what areas needed to be addressed.