I dont think it is a matter of using too much EQ, its how and where it is used. if you are pushing the bass bands all the way up,. then you are gonna run out of signal for higher freq bands.

as an example, if you boost 40hz 12 db, then your signal level for that band will dictate how much signal the midrange will have because your bass will clip first. this will reduce your midrange signal level by 12db. If you have an EQ with 9volts of signal this may not be as critical as it is if you have 2-4 volts.

If you are making adjacent bands at opposite ends of boost/cut, then you are not doing it right either (meaning if you have 1khz boosted 6db and then 2.6khz cut by 3db.) you want a nice smooth transition.

my suggestion with an EQ is to set all the bands to zero, then play pink noise and fix your hot spots with cuts. once you have it decently flat, then start fixing things by ear, you should be able to tell. just take a picture of your EQ (or save it if you have a DSP) so you always have a baseline to go back to.

 

I rarely boost anything. If I do boost, I limit it to 3 db. Some of my cuts, though... I've cut some frequencies by as 12 to 15 db, or even more. I don't want to overdo it, but it seems that my car is quite peaky in some areas.

 

I assume you have an RTA? if you do, then check that huge cuts like that are actually cutting. I have a hole at 8khz that must be a reflection/cancelation because no amount of boost fixes it. more I boost, the more it cancels and output on the RTA stays the same.

 

I don't have an RTA. I'm trying to squeak by with test tones and an SPL meter... and also my ears. I have a few issues like yours, where boosts and/or cuts don't provide a measurable difference. What do you do in these instances, leave that band flat?

 

This is a common problem in tuning by ear. What happens usually is one tunes with one piece of music until it sounds good. Then another piece of music goes in that has energy concentrated in places where the original piece didn't and the annoying frequencies get turned down. then, on to a third piece where more stuff gets turned down and so on and so forth until there's no more midrange.

Buy a measurement Mic for 60 bucks at parts express and download REW from hometheatershack.com. This is by far the cheapest and best upgrade for anyone's system.

 

The whole idea of tuning is that everything should sound good. Measuring your way to a proven house curve will give you a good base. For 95% of the folks this may be enough. But for the balance 5% who want to tweak further, to get the best sound across all well recorded media, the house curve is starting point.

Tuning by ear is perfectly valid. You need to train your ear to pick better or worse with 80% or higher accuracy at 1/3 oct resolution. You need to need to listen, pick an issue localize it and then pick the 1/3 oct band to correct. Correcting an issue can involve tweaking multiple frequencies.

The Harmon How To Listen app is about training your ears. I'm not sure how you can deny the validity of tuning by ear.

 

Most people can't do it effectively. They spend hours upon hours tweaking things back and forth and making things different, worse and better, back and forth spinning their wheels. For less than $100 you can get a measurement setup ready to go and after reading a few guides can have your car moving in the right direction. Tuning by ear HAS a place, but it's really best when you've already got things dialed in and are just tweaking to your own personal tastes.

 

<sqnut> What I read in Andy's post is that most people do not tune by ear *properly*. I didn't read that he was completely discounting the validity of tuning by ear - when it is done correctly. You yourself pointed out that setting a baseline against a proven house curve is step one. Final tweak with trained ears & controlled methodology is step two.

Step one requires a measurement setup or auto-EQ. In my opinion this is the most important step. This is not in conflict with the points Andy and T3mpest were making. I do agree, many use EQ 'by ear' incorrectly (either by not doing step one first, or by doing step two incorrectly and endlessly) They would be much better off only doing step one and leaving it alone.

 

Yeah I thought Andy was discounting the whole ear tuning thing. I definitely think you should get your measured response right before even attempting the tune by ear.

 

getting a waterfall plot to indicate problem areas is more useful than straight line RTA, I'd suggest going the extra and getting good analyzer software or at the least, try to make it to a GTG and maybe coordinate a session of tuning at beer cost, or peanuts.

 

So, I shouldn't be afraid of 12 db cuts, if that's what the tools and my ears tell me?

 

A 12 db cut would certainly be unusual. The only place I cut that much is about an octave above and below the pass band. Even here it's more to do with trying to get a true 4th order acoustic slope and reduce audible distortion beyond the drivers pass band.

I'm using the eq to do two things, balance for L/R and to set the overall response. Given these objectives, the deepest cut I have at any one band is about 4-5db, that's the max. This is in the pass band of the driver.

If you're going by ear and feel that you need to cut something 12db, then chances are very high that you're cutting the wrong frequency. If you're measuring and seeing a 10db peak then you need to look at the resolution you are measuring at. So a measured 10db peak at say 8khz @ 1/24 oct resolution, does not mean that that you need to cut that frequency by 10db.

Keep in mind that your hearing sensitivity is ~ 1/6 oct and you can correct only at 1/3 oct. So go back and measure and at 1/6 and 1/3. As a rough thumb rule, measure at 1/6 to identify the frequency and at 1/3 to get an idea of how much you need to cut. That 10db peak at 7.15kz @ 1/24 oct may just be a 4db bump at 1/6oct and a 2db bump at 8khz @ 1/3. So cut 8khz by 2-3db and that should be enough.

Hope this helps.

Arun

 

I wpuldnt be afriad of them, unless only 1band needs it and no cut at all on adjacent bands.

 

I ordered a mic for my phone from parts express today. Not ideal, but should be a helluva lot better than what I've been rolling with. We shall see soon.

 

12dB cuts or even more are quite common in tuning a car. Of course, if you eschew using an analyzer, then you'd never know this.

the reason I say tuning by ear is a bad deal is because:

1. Few people can correlate what they hear with the tools. This is what Harman's training attempts to address.

2. Different pieces of music concentrate energy in different spectra. IF you have a massive peak at 1k, but the music you're using to tune doesn't have any content there, you won't find anything to fix. Then, when you play a piece with a saxophone or Amy Winehouse singing, you'll reach for the EQ and turn down 1k too much because there's more info concentrated at 1k than in lots of other music.

My suggestion is to tune with the analyzer. Then listen. Then verify what you THINK you hear with the analyzer. Then retune with the analyzer.

Think about this one. Let's say you were asked to turn the volume up by 3dB. The first time, you had to choose the 3dB using your ears. The second time, you could use an SPL meter. Which one would be more accurate?

I'm just trying to be helpful here, as usual.

SqNut, you're really frustrating.

 

SqNut, you're really frustrating.

Well, it's not by design, if that is of any help. I respect you and I continue to learn from you. I'm not trying to be a contrarian but I would really like to discuss this topic with you. Just want to clear the context. I know I'll be wiser in some way post this discussion.

12dB cuts or even more are quite common in tuning a car. Of course, if you eschew using an analyzer, then you'd never know this.

Depends on how you're measuring and correcting. If you're measuring at 1/12 or 1/24 and you have control over a thousand pots across the spectrum and the objective is to flatten all measured peaks, then yes one will be cutting 12db or more at quite a few places.

But you don't have to flatten everything you can measure, only that which you're hearing. Ear level measuement in a room will also give combing artifacts. Look at it in a high enough resolution and you will have your 8-10db peaks there as well. But that doesn't seem to bother us. One doesn't feel the need to whip out an eq and start tweaking.

At a resolution of 1/3 oct where one can correct (on most processors) or 1/6 ish which is what we're basically hearing, the situation looks very different. Heck, my bit10 maxes out at +/- 12db and both my ears and measurements tell me that I don't need to cut or boost anything in a drivers pass band by 12db. Beyond the pass band sure I'd cut 20 db if I could.

An analyzer is as common as an rta app on the android, as a very basic tool.

1. Few people can correlate what they hear with the tools. This is what Harman's training attempts to address.

I wish I had access to this tool seven years ago. The ability to correlate what one is hearing with the relevant frequency/range to correct takes time. I tried the tool about 2 years back after ~5 years of tuning by measuring and by ear and one could pick the frequency boosted or cut or be within +/- 1/3 oct 9 times out of 10. If I had the tool earlier I would have learned quicker.

The point is that correlation, can be learned. Either with the tool or by just stumbling along till things start falling in place. Either ways it is not something that one learns in a day or a month or even a year. It is a painfully slow process. But it can be learned. 80, 160 and 200hz all add extra bloat / fatness / boom to the sound, the three are very different and distinct sounds. If its bloated cut 80, if its fat cut 160 and if its boomy cut ~200.

2. Different pieces of music concentrate energy in different spectra. IF you have a massive peak at 1k, but the music you're using to tune doesn't have any content there, you won't find anything to fix. Then, when you play a piece with a saxophone or Amy Winehouse singing, you'll reach for the EQ and turn down 1k too much because there's more info concentrated at 1k than in lots of other music.

When you have the right balance in the response it doesn't make a difference if you're listening to Amy or Immogen Heap, they will both sound accurate. Tuning is done across a wide spectrum of music. Acoustic balance across the 10 octaves makes everything fall in place. If one has a huge peak at 1khz, one will hear the sound thin out. Too little of 1khz and the sound is dead with no dynamics, too much and the sound will thin out and sound like a cheap radio.

Think about this one. Let's say you were asked to turn the volume up by 3dB. The first time, you had to choose the 3dB using your ears. The second time, you could use an SPL meter. Which one would be more accurate?

Sure measuring is easier and more accurate. But the whole point of tuning by ear is that the measurement devices that count are those which are fixed to the sides of our head. The ear will tell you if a cut of X is too much, too little or just right. The ears are the measuring device. Assuming the ability to correlate is in place.

I'm just trying to be helpful here, as usual.

Now you're embarrassing me .

One has to start off by measuring, get the timing right, get the L/R balanced, level match, then measure and dial in the appropriate house curve. What next? Is that the best the car can sound? Do we now need to measure 50 additional things.

At the end of the day it comes down to using timing and response and the ears to go beyond that. Measurements will take you so far, but beyond that you have to go by what you're hearing. You will do some basic measurements even when you're tuning by ear, just to make you haven't thrown something out of whack. But if you have, chances are your ears will tell you that even before you pull out the mic and laptop.

Arun

 

Interesting direction this is going in.

I took the side of "cut what you need to" to fix a problem. My car is super loud in the 160-250 range and I have to use 10-12 db cuts in there to correct.

I have tried several different curves and have settled on one and have been tweaking it for months on end.


I feel I am getting there as I listen to a wide variety of music and all of it sounds pretty damn good with just 1 or 2 clicks +/- to bass/treble and sub level matching to the music.

 

Afraid to say mine once resembled a happy face


...good times.

 

That's ok, the settings on my mids look like a sad puppy face