DiyMobileAudio.com Car Stereo Forum banner

21 - 40 of 455 Posts

·
Registered
Joined
·
6,213 Posts
Yeah... several. The Helix software is just an RTA, you can't do any real troubleshooting or observe the time domain with an RTA. RoomEQ can perform autoEQ with both graphic and parametric filtering. It got a better overlay, average and graph managing ability. Quicker to work with (imo).

If you're just interested in the the overall FR, the Helix software works indeed. I'll present a few reasons why RoomEQ is better later on ;)
I get that RoomEQ is more powerful than the Helix software, but how do i then apply the settings that i view in roomEQ to my vehicle?
 

·
Registered
Joined
·
4,939 Posts
Discussion Starter #25
I get that RoomEQ is more powerful than the Helix software, but how do i then apply the settings that i view in roomEQ to my vehicle?
I will explain that in the next post... probably today ;)

Tapaaatalk!!
 

·
Registered
Joined
·
861 Posts
I have a Dayton umm-6 usb mic, which says its A weighted.

The options in REW is "C weighted" or "Mic or Z weighted", which one should be selected?
 

·
Registered
Joined
·
4,939 Posts
Discussion Starter #28
I have a Dayton umm-6 usb mic, which says its A weighted.

The options in REW is "C weighted" or "Mic or Z weighted", which one should be selected?
That's kinda weird. If you load the calibration file the weighing will be disregarded inside the range of the cal file. Choose Z otherwise.
 

·
Registered
Joined
·
861 Posts
That's kinda weird. If you load the calibration file the weighing will be disregarded inside the range of the cal file. Choose Z otherwise.
yes I have the calibration file loaded and "mic or Z weighted" selected. Thanks for this helpful guide as I just installed REW and was not sure where to begin! Thanks
 

·
Registered
Joined
·
4,939 Posts
Discussion Starter #30
Setup your system, Pt1: Interpret your measurements

OK. We had three parts so far. You should be a little more familiar with the software and how to do the "actual measurements" now but we won't jump into the advanced features just yet. First we need to understand what the measurement actually describes, what we can do with it and why we need equalization in a car. I'll try to keep it as simple as possible.

Even if this is a measurement tutorial I think that knowing some of the basic theory how we interpret sound is a good thing. It allows you to understand what you're doing and WHY you are doing it.

So, first some quick facts how we humans interpret sound, how we localize sound and why it's important.


[*] At low frequencies we localize sound based on the time it takes for the sound to travel from one ear to the other. This is called the "inter-aural time difference (ITD)". ITD is the main contributor for sound localization below ~800Hz.


[*] At high frequencies we localize sound based on the intensity (volume/level) difference between the ears. This is called the "inter-aural intensity difference (IID)". IID is the main contributor to sound localization beyond ~1600Hz.


[*] Between 800-2000Hz there's "gray area" where both IID and ITD contributing to sound localization and in the area 1000-3000Hz it's harder to pin-point where sound originates from than usual. Good to know when tuning.

Why is this important? Because if we want a good soundstage, with the vocals centered and a stable focused stage we need to consider both the difference in physical distance between speakers and the intensity difference altered by the interior of the car.

*Some headunits and basically all DSPs got a feature called 'Time Alignment' which is used to delay the speakers closest to you and therefore electronically compensate for the distance difference between left and right side speakers. T/A (Time Alignment) mainly affects the 'ITD frequency range'.

*Few headunits and many DSPs got channel independent equalization (L/R EQ). EQ is basically a "frequency selective volume adjuster", that makes sense to you? To compensate for the intensity difference between the left and right side we need equalization on both channels. As you might guess, the L/R EQ function affects the 'IID frequency range' to a great extent, but it does also affect the 'ITD frequency range' in a way* (I'll discuss this and the 'modal range' later on).

Basically L/R EQ and T/A go hand in hand. You need both for correct tonality and staging. That's why a DSP is so important, all car audio systems need processing of some kind to sound right.

See the picture below for an illustration:

rew18.jpg

OK. Let's return to RoomEQ again... If you have followed the previous instructions in measuring all individual speakers and several averaging points of each, you could end up with something like this:

rew19.jpg

This is just an illustration of the mess of trying to manage every graph at the same time (not smoothing the graphs add to the mess). If you want to avoid this, first deselect all graphs (see picture below) and then select the graphs you want to work with.

rew20.jpg

For this little example I selected the subwoofers (remember to set an appropriate graph frequency and level limit for easier viewing).

rew21.jpg

-------------------------------------------------

Various good-to-know facts before you get started.

## OK. I'm going to go through some various things before going into the fun stuff. First off:

There's a big misconception about crossovers!

[1.] So you enter your desired crossover points and slopes in the DSP and that's it, right...? Unfortunately, there's two "forms" of crossovers. One of the "forms" exist in the electrical domain and one in the acoustic domain. The ones you set in the DSP is the ELECTRICAL CROSSOVERS and they NOT the same as an acoustic crossover.

[2.] The ACOUSTIC CROSSOVER is the MEASURED crossover point and slope (as in 'measured with a microphone inside the car - in listening position'). This is "real" crossover point, the electrical crossovers can be set in any way you want as long as you reach the "optimal" acoustic crossover point/slope.

[2a.] Tip: Use different crossover points and slopes, left and right side to fix some of the frequency response errors BEFORE EQing the system. (We'll discuss this in detail later).

[3.] The actual (acoustic) crossover point is dependent on the relative level/volume of two speakers. IMPORTANT! See pictures below:

"Normal levels"
rew22.jpg

Midwoofers levels are offset by +10dB and resulting crossover point changes as you can see.
rew23.jpg

This might be an eye-opener for some ;)

So how bad can it be? Do we really need a DSP?

The picture below speak for itself. This is a typical door mounted speaker frequency response (fullrange, no EQ, same input level between L/R speaker).

rew24.jpg

15dB difference! Right in the smack of the lower midrange (vocal fundamentals lies here). That can't be good... I promise, it isn't. We can do "some work" by being creative with levels and crossovers but we're still going to need equalization here to level out the difference. As you might understand by now, an ordinary EQ is not enough, you need a channel independent EQ to compensate for such issues as in this example...

That's it for now... I have few more things I want to add but soon we're going to discuss the fun stuff ;)
 

·
Premium Member
Joined
·
2,414 Posts
Question. Should the Time Alignment be measured and determined before performing these room measurements and EQ with REW?
 

·
Premium Member
Joined
·
2,414 Posts
Doesn't matter. I do it before.

I'll will write up the next part soon, very busy at the moment.
Cool. I figured it best to do time alignment first.

I'm waiting for my Behrenger mic to arrive. Going to use my MacBook Pro with REW and an Allen & Heath Xone DB4. :D
It will be overkill on the soundcard, but should work. That way I don't need to buy something new.
 

·
Premium Member
Joined
·
2,414 Posts
My microphone is being delivered today. Really excited to start working with REW. :D

Going to do time alignment first, and then check the response and work on my EQ.
 

·
Registered
Joined
·
4,939 Posts
Discussion Starter #35
All right. Let's continue from my last post.

Setup your system, Pt2: Equalization theory

This is where I left off last time... If we look closely at the picture below we can see that there are peaks and dips in the response, it has a response deviation of 20dB from the highest peak to the lowest dip.

rew21.jpg
PICTURE 1 (Subwoofer Measurement, Sealed enclosure - Measured inside car)

A response deviation as large as 20dB is an issue, it WILL be audible and affect the tonal balance severely. As you may know, EQ is a feature to compensate for frequency response errors such as the graph shows, but can all response issues be 'fixed' with EQ and if not, why?

[*] As you boost with EQ, you will demand more more power from the amplifier in the EQ'ed area. A 3dB boost equals twice as much power and 6dB equals four times the power!

This goes both ways, if you cut with EQ you will demand less power from the amp.

[*] EQ will affect the speaker's performance in more ways than you might think. I won't go into detail of this but as you increase volume a speaker's (non-linear) distortion level will also increase. As EQ basically is a frequency selective volume control you will increase or decrease distortion if you boost or cut in a given area.

[*] Some areas can't be EQed, large dips in the response, especially those low in frequency 'in the modal range', should be left alone. They are related to the speakers locations, the listening position and the inner dimensions of the car. Problems around 60-80Hz and 120-160Hz are common in average sized cars (see "picture 1", the dip at ~65Hz is one of these issues). These areas can actually be viewed inside RoomEQ and this is an 'advanced feature' I will discuss later...

There can be cancellations in the response due to phase inconsistencies as well. Incorrect set time delay between speakers or incorrect polarity are common issues that can result in dips in the response. The response dip will not respond well to EQ in this case.

Before I describe how to use the autoEQ of RoomEQ/REW, you must understand the basics of equalization...

There are basically two types of EQ's available in most car audio DSP's and headunits. They are called "Graphic Equalizer" or "GEQ" and "Parametric Equalizer" or "PEQ" in short. A graphic EQ is easier to use, but also less powerful than the parametric type. Since parametric equalization is so widely used nowadays I will put some focus on that.

Graphic, "Semi-Parametric" and "Full-Parametric" EQ

A graphic EQ is the "normal" type of equalizer most people probably have seen in both home audio and car audio at some point. A GEQ has a varying number of EQ bands, which are centered around FIXED frequencies. The better units often got EQ bands with 1/3 octave spacing, i.e 200Hz, 250Hz, 315Hz, 400, 500Hz etc. Each of these frequency bands also got a FIXED BANDWIDTH, this means that if you choose to decrease 500Hz by 10dB, you won't simply lower just 500Hz, the frequencies around 500Hz will also be affected. That's why we call each EQ band "Center Frequency" . The area around the 'center frequency' is called "Bandwidth". The bandwidth is defined as "Q". A larger "Q"-value means a NARROW bandwidth (only the frequencies close to the center frequency are affected), a smaller Q-value have a WIDE bandwidth and will therefore affect a large amount of frequencies around the center frequencies.

parmet2.gif
EQ bandwidth (Q's)


*With a GEQ you will have FIXED center frequencies with a FIXED spacing and each center frequency will have a FIXED bandwidth (Q). This type of equalizer is easy to use but not as flexible as a parametric EQ.

*A parametric CAN be used in the same way as a graphic EQ. With a PEQ you gain the ability to CHANGE center frequency and bandwidth (Q). This type is more advanced to setup but is far more powerful if used correctly. A "full parametric EQ" will basically allow you choose whatever center frequency you want freely and a wide range of Q values. A "semi-parametric EQ" will allow you to change center frequency in a fixed octave spacing.

Example; You have an issue at 140Hz in the frequency response that needs EQing. With a good graphic EQ you should have EQ bands at 125Hz and 160Hz, unfortunately you won't be able to touch 140Hz without affecting the the frequencies around 125 and 160Hz as well. A parametric EQ allows you to choose a center EQ-band at 140Hz and the desired bandwidth (Q) to only attack the problem directly without touching the 125 and 160Hz bands.

How to EQ...

*Generally speaking, never use narrow band (high Q) EQ in the higher frequencies!

*Higher Q equalization/filters should only be used within the modal range (lower frequencies).

*Rather cut than boost. Remember the speaker strain and power requirement when boosting. I generally recommend not to boost more than 3dB.

*If the the frequency response doesn't respond to boosting an area by equalization, DON'T USE EQ there. You will only increase distortion.

*Don't be afraid to use large amount of EQ in problem areas, it's the acoustic response that matters.

Make sure levels and crossovers are optimized before EQ. IMPORTANT!

The optimization of levels and crossovers will therefore be be first step in setting up your system and I will discuss this in detail in PART 3. So you who have been waiting for the practical tutorial in RoomEQ will be happy next time ;)

Until next time ^^
 
21 - 40 of 455 Posts
Top