35 replies to this topic

[Fury♫]

I'm seeing this term being thrown around a hell of a lot on this forum... some people are saying that RTA tuning is the key to "SQ", others saying it's horrible and lifeless...
My belief is that a lot of people here misunderstand it's use(s), or have been brainwashed with garbage for/against it.

So quite simply my question is:
What do YOU mean by RTA, and how do YOU use one?


Hopefully we can build a discussion on RTA and measurements, and how to PROPERLY use, and interpret the result.

[~thematt~]

I have a toolbox. In it are various types of tools, for various problems. The RTA is one such tool in my toolbox.

Understanding its limitations, I tend to use the RTA to obtain a visual representation of a summed instantaneous smoothed snapshot of the frequency response from the system. The RTA cannot decipher the time domain.

I find it works most effectively not to smooth the overall response, but to allow manipulation of my system to match the response for both seats, to all channels. It also allows me to see, in real-time, quirky responses that can be identified by my ears, but not specifically nailed down.

I do not tune with the RTA alone. Same as I do not build a house using just a hammer.

[SCorpion]

An RTA is used to collect information.

That's all there is to it.

It is up to the user to determine what that information means. Same as "just using your ears." It is a means of collecting information. It is up to the user to determine what that information means.

There is nothing else to it.

How to use a computer based RTA is and what it is used for is definitly a VERY large topic to discuss.

Is there anything you would like to discuss in particular fury?

[HEKYEH]

I have a pretty simple RTA setup. Omni directional mic connected to an external sound card then connected to my laptop. Sure, it's not overly high tech, but for about $400 it has me satisfied.

I have used it to see where there may be any major peaks or troughs in the frequency response in my system. It's not perfect, but it's a good starting point. It's also a good tool in helping me learn about another facet of audio.

[phunkdust]

An RTA should also help you learn and train your ears to what frequencies you're hearing... so you ideally go from "Gee that sounds a bit harsh, lemme pull out the RTA and find out what's up" to "Gee that sounds a bit harsh, I'll pull back 6khz a bit cos thats what it sounds like"

RTA is probably more useful for diagnosing phase issues in a car environment though.

Edited by phunkdust, 22 April 2009 - 11:17 PM.

[Cyberpunky]

An RTA takes a 2 dimensional picture of a 3 dimensional reality. As stated above it doesn't factor in time. Its just levels at different freqs. It can be a handy tool but isnt definitive regarding the fidelity of a system.
peace
Cyberpunky

[muzzy66]

I'll second what theMatt, Cyberpunky and others have stated here.

An RTA measures the frequency response of your vehicle, and that's about it.

Some RTA's can also measure phase as well as speaker T/S specifications (which can be handy) but the most typical use for it is to get an approximation of the frequency response of your system, within your car.

[~thematt~]

nidekcus, on Apr 22 2009, 08:01 PM, said:

I have a pretty simple RTA setup. Omni directional mic connected to an external sound card then connected to my laptop. Sure, it's not overly high tech, but for about $400 it has me satisfied.

You'll be surprised how much more advanced that setup can be, over an RTA boxed solution.

And more accurate.

[fuddbutter]

an RTA is an extension of your SQ penis..

[Fury♫]

Well lets define what an RTA is?
An RTA is a Real Time Analyzer, it allows us to measure the frequency spectrum of an audio system.

So how do we use an RTA?
It is generally accepted that we play pink noise through a system, and measure the output at a certain position in the car (as has been done in many previous

sq formats ie: IASCA). It is generally measured in 1/3 octave steps (which corresponds to the same bands on a 31 band equalizer).

This is unfortunately where RTA measurements become somewhat useless, and has given the use of "measurement tuning" a bad name. This way of tuning is

inherently flawed, and i'll touch on a few reasons why, and mention a few methods which, using an RTA, is much more effective.


So, why Pink Noise is flawed for RTA tuning...

Firstly what is pink noise?
Pink noise is a signal with equal power for all frequencies.
This actually gives us a response that drops -3db/oct from low to high, over the audio spectrum.
An image of this can be seen here:
http://en.wikipedia....se_spectrum.png

WikiPedia also has a technical explanation of it here for those interested:WikiPedia PinkNoise

This in itself means that tuning for a flat RTA Response using pink noise is flawed.
If we tune pink noise to "flat", then essentially we are creating a 3db/oct increase from low to high (inverse of the above graph essentially).
This is also a common reason why systems tuned like this can often sound bright.

Our next problem is reflections. As pink noise is playing all frequencies, it is exciting all reflections in the environment, and we are measuring both

direct and reflected sounds. This is also commonly known as the power response of the system. Equalizing this flat can make a system sound flat and lifeless.

What we actually want is a flat anechoic response. We still want the room to add a bit of "life" to the sound. This unfortunately cannot be measured with

pink noise.


Next we move on to the positioning of the mic. Ever moved your head from side to side and noticed the sound (tonality) changes?
This is no different from a microphone. Move the mic 2cm to the left, and you'll see you receive a different response. So how, or where do we place the mic to accomodate for this? Most people only take one measurement in one position, this can sadly give us an inaccurate result.
A bloke called Earl Geddes has done a fair bit of research into this (you can check out his JAES papers if you're really keen), and he's found that taking one measurement will give you an accuracy window of +/- 8db... this is pretty poor considering we're after the utmost SQ.
What's the solution you ask? We use a method called "spatial averaging". We take multiple measurements within the same area (moving the mic a few cm at a time), and average the results that we get. Geddes also found that a minimum of 6 samples in strategic posisions are required to give us an accuracy of +/- 0.5db. Much better!


Now what about the effects you have on the environment? What does your body do to the sound in a car? A lot of people recommend putting a dummy head (a foam replica of your head) next to the microphone, to replicate the effects your head has on the sound. You could go one step further and add an entire body, but it is more difficult (or you could sit yourself in the car as well). The human body can obviously absorb higher frequencies due to its mass and thickness, so this can obviously have an impact. How far you want to go with getting accurate results is entirely up to you.


Obviously not all of the above relate solely to pink noise, but to many forms of measuring.
I will briefly try and explain a few other methods of measuring sound, besides pink noise.

* Sine Sweep method: This method involves playing a sine sweep, and measuring the results. This is very good in giving us impulse response, as it can be measured over time. It can also allow us to measure reflections, as a reflection at a frequency will show up after the original sound has been played, and wont be masked like with pink noise.

* MLS Measurements: similar to the sine swept method, this can enable us to get quasi-anechoic measurements without having to place a driver in an anechoic environment. More on MLS: http://purebits.com/mlsteo.html

There are other methods as well but i'll leave it there as I have work to do

[syd-monster]

In the Traditional sence a RTA was a very popular product made famous in car audio by AudioControl.

R = Real
T = Time
A = Analyzer

Like the boys have said above, it measure frequency responce, in real time, that is whilst the system is playing.
Their in-box, plug in & use system looked like this.





These days you can set it all up & do much more on PC/Laptops




How-to here... (one of many ways)
http://www.diymobileaudio.com/forum/diyma-...100-dlls-5.html

[BMWTurbo]

I use ARTA....

http://www.fesb.hr/~mateljan/arta/

I have made cables to measure the impedence rise etc of drivers and used them to good effect also with the software.

The one thing I don't have is a calibrated mic. I am not super concerned with this jsut yet until I get confident enough to get replicable results from a tuning guide.


Just on MLS etc you can set the 'gate' length in order to filter out reflections.

[HEKYEH]

phunkdust, on Apr 22 2009, 11:12 PM, said:

An RTA should also help you learn and train your ears to what frequencies you're hearing... so you ideally go from "Gee that sounds a bit harsh, lemme pull out the RTA and find out what's up" to "Gee that sounds a bit harsh, I'll pull back 6khz a bit cos thats what it sounds like"

Yes, that's what I have found it quite useful for. It's a good way to learn what kind of sounds are at what frequencies.

[Cyberpunky]

fury, on Apr 23 2009, 02:41 AM, said:

Well lets define what an RTA is?
An RTA is a Real Time Analyzer, it allows us to measure the frequency spectrum of an audio system....

Lets define it properly, rather than in a half baked fashion(no offence).
It measures and displays the relative signal levels across a spectrum of audio, in real time.

Most of us have seen a spectrum analyzer on an mp3 player or other audio equipment. It may just have bass midrange and treble levels, dancing up and down. The difference in a 1/3rd octave RTA, is the levels are displayed in real time(or with such low latency[delay] that we can see any adjustments as we make them). The other benefit is being 1/3rd octave it allows fine enough resolution that we can see and then precisely adjust any problem area without effecting neighbouring frequencies, provided we have a suitable equaliser that has similar resolution or other capability to adjust the problem area.

It aids tuning because peaks at a particular frequency can be offensive to the ear espesh at higher frequencies, and it can also show us dips. Dips although influencing the accuracy of the response of the system are not offensive to the ear because they are by definition, the absense or a lower level output of a particular frequency or area.

Pink noise is the easiest source to use as it will quickly show us any peaks or dips in our spectrum. We may use EQ or crossovers(in an active system, to adjust any anomalies we find. Although a lot of ppl think EQ is the only thing used to adjust problem areas, if the peak or dip is caused by overlap or underlap at a crossover point, then adjusting these points is also an effective method.

There are other methods to find peaks/dips, such as using more boost to exagerate any peak, or more cut to do the same with any dip, with out using an RTA, but these methods require patience and often more expertise.

RTAs can be handy but are by no means essential, and at best, can only give us some guidance, on the way to a sweet sounding system.

peace
Cyberpunky

[fuddbutter]

hmmm i thought it meant Richard Clark Tonality Analyzer...