PDA

View Full Version : System EQ & Setup



ralphs99
11-24-2004, 08:55 PM
I'm wondering what you DIY guys are using to EQ your systems with. I've been struggling to date with 1/3 octave noise fed into the speakers with a measurement mic positioned at various distances.
This is giving me results that don't seem to match what I'm hearing. Even allowing for the effects of room reflections it seems less than ideal.
I guess what I need is some sort of system that measures amplitude response before the arrival of the first reflection. I know there are such systems out there, but I can't afford to be spending thousands, so I was wondering what the collective Lansing brain could suggest!

Cheers,
Ralph

Robh3606
11-24-2004, 09:45 PM
Hello Ralph

I am using a Behringer 8024 in RTA mode for measurement with the Behringer measurement microphone. I have run it against a couple of software packages and in 1/3 octave mode it matches right up. I use Urei 839 cut only's for the EQ. This combination works very well for me. I measure from my primary listenning position and set-up the system as flat as I can through the midrange. I also try to match the L+R speaker as close as I can too as far as frequency balance and level. I really don't get any surprises and am able to get a very good over all response graph with the smoothing the RTA does. Attached is in room both channels driven with pink noise. The HF roll off is from the 2344 horns and the levels bellow 60Hz are where the subs run. If I run them flat bass/VLF sounds a bit anemic.

What kind of problems are you having?? New room?? New Speakers?? New EQ'S?? Rooms can really be a PITA. I was lucky that my basement was a shell when we moved in so it was "mine" and the stereo was banished there. That was a good thing cause my livingroom is aweful compared to the basement. I was able to set-up the speakers and listening postions for the best sound in room. Not the best use wise. Without that kind of flexability it can very difficult to work around WAF and other issues like the livingroom couch and the bay window. What kind of space are you in?? Maybe we can help you.

boputnam
11-25-2004, 12:03 AM
I've been struggling to date with 1/3 octave noise fed into the speakers with a measurement mic positioned at various distances. Hey, Ralph - know our friend Ian...? He might could help, here.

Two (three) things might be giving you a wobble.

- How flat and reliable is your mic? Not a good place to spare expense.
- Are you able to measure against a signal post the pre-amp? The best results come from comparing the pre-amp output signal to the acoustically measured signal, rather than EQ'ing to pink noise acoustic output, only. Not doing this could be introducing some subtle differences in your resultant EQ.
- You are using "pink", right?

I was able to get "good" results with an AudioControl SA-3054. Best results were obtained using SmaartLIVE in FFT "Transfer" mode, with Earthworks M30 mic. I was able to justify the expense 'cause I use it 2- to 3-times a month for our live gigs...

Ian Mackenzie
11-25-2004, 03:21 AM
Sorry I not much of a knob farmer but I consider matching broadband driver sensitivity (as a womans tits) critical for optimum presentation.

Ian:eek:

johnaec
11-25-2004, 06:35 AM
One thing to be aware of is that the human ear hears sound far from flat, mainly depending on the volume level. I assume you're familiar with Fletcher-Munson equal loudness curves? Many people will EQ a system flat, then wonder why it sounds so bad. I usually start with a flat EQ as per the RTA, then adjust the lows and highs accordingly to take the loudness curves into account. Here's one of many brief commentaries about this on the web: http://www.allchurchsound.com/ACS/edart/fmelc.html

Frankly, I get the best sound when I manually adjust the curves band by band, using my ears and a discreet frequency sound source, such as the Neutrik Minirator. It puts out single frequencies one at a time centered on the ISO 1/3 octave frequencies, and I just go through them until each one sounds equal to me. This has always resulted in better sound to me than using pink noise with an RTA, though I often go through a series of procedures to check against each other. Granted, the result may not be optimum for all listeners, but I know it's accurate for my personal use. Doing live sound, (band), I stick much closer to the averages...

John

ralphs99
11-26-2004, 12:32 AM
Thanks for the feeback everyone. I am constantly amazed at the quality of this forum.



Firstly you can always trust an Aussie to bring down the tone of the conversation! (but you're right....):)

Rob: That's a nice smooth looking curve on the RTA. Can you tell me what the vertical axis is?

Initially I was thinking about just EQing everything at the listening position. Then I thought it would be better to EQ the speakers to flat using a parametric EQ, and then making hopefully minor adjustments at the listening position with a graphic. I figured that every decent speaker manufactured aims at having a flat response and that it would make my life easier if I decided to move the speakers to a different position and so have to re-EQ everything.

I measured that the critical distance in my small living room is about 1.5m for frequencies above about 300Hz, so to get a flat direct SPL from the speaker I need to measure at distance smaller than this. At the listening position I am more into the reverberant field so the touch-up EQ for the room should only need to (maybe) compensate for the power response and null out the worst of the low-frequency peaks. At least that's what I thought would happen....

Bo: I'd love to have your setup!

The speakers are a home brew setup with 2420 on 2344, E110 with paper domes in 0.5cuft boxes, 2231 reconed with 2235in 5cuft. I recently bought a DBX drive rack and now have lots of filters to play with! Up until now I had little in the way of EQ and mostly adjusted the driver levels with an active crossover and added some passive HF boost as per the crossover in the 4430's. I didn't worry too much about the rest of the peaks and dips in the amplitude response and just enjoyed what I had.

The signal source is an Audio Precision CD with 1/3 octave pink noise that I have measured electrically and done an FFT on. The corrections required are less than +/- 1dB. The measurement mic is a sennheiser insert with published response of +/-2dB up to 15kHz but not calibrated. The electronics is flat. The variablity in source and mic response is clearly a factor, but not enough to give me the results I'm hearing.

The EQ required to get a flattish response varies a bit depending on the distance the measurement mic is placed from the speakers so I average measurements made at 0.25, 0.5 and 1m, on axis with the driver I'm measuring. Listening along as I'm EQing I hear a lot of harshness from the horns. I thought that might be because of the +6B DI of the horns.

John: After reading your reply makes me think that it's partly due to the Fletcher-Munson effect. I am well aware of the effect at low frequencies but never really gave it much thought in the mid-band. So that's one mystery solved! However it still doesn't explain the subjective difference between one 1/3 octave band and the next. There are significant differences in the harshness between frequency bands. The meter doesn't indicate more than a couple of dB different, but it sure sounds different. Maybe the diaphrams aren't up to speed anymore. I've been looking at some other posts that claim huge improvements with new diaphrams. Maybe their old diaphrams displayed similar problems. I'd be very interested to know what the subjective or objective differences a new set of diaphrams make. Mine are the original 70's tangential surround. I was hoping in the back of my mind that I could EQ my way out of having to buy new diaphrams, but I think I'll have to bite the bullet and buy a set of D8R2425's. (can't justify the cost of 2421's)

Given that the Fletcher-Munson effect is causing part of the extra harshness, then if I EQ by ear I'm going to end up with an EQ setting that is similar to the Flecther-Munson curves in the mid band. Not flat at all! Should we really be EQing speakers this way? Or are we so used to hearing mostly-flat speakers that this is what sounds 'correct'? I will try out some other speakers I have and see if I can hear similar mid-band peaks and dips. What are everyone's thoughts on what the target is? Flat at the speakers? Flat at the listening position? Fletcher-Munson response? I feel like I've uncovered a Pandora's box of pshyco-acoustic questions!:confused: (couldn't find a happy face with question marks above it's head)

Ralph.

Ian Mackenzie
11-26-2004, 02:09 AM
Ralph,

I might be able to help you out here.

The level matching does matter around here.

Imho there are too many 2307/2308 & 2405 bashers on account of Tin Ear level matching but that's anoth story. Non better known to than the Pope himself Bo the L100 basher who communes locked away sitting on a bar stool most Saturdays front of a set of 4345's down on the SF peninisula CA.

The harshness you hear maybe the original diaphragms in your 2420, that would be a worthwhile upgrade, I did this not that long ago when I bought my 2420 SH.

I might be interesting in selling my D8R2425's as I plan to upgrade to something else soon. Pm me if you are interested.

Old aluminium diaphragms go off eventually and have a characteristic harsh tone. The newer diamond Ti versions are more robust and last longer, and the prevous aluminium diamond versions are very nice but rare as hens teeth.

As to the eq, without a time gated FFT its a bitch as you need to distinguish the drivers individual responses apart from early reflections. To be honest unless you have a known reference and know the speaker is flat to start with the graphic/RTA is only useful for room correction imho but non the less a very useful tool for final eq at the listener position.

I use a windows based software package called Winairr that does everything adequately for this purpose without breaking the bank incl MLS and pulse using a standard duplex soundcard and for less US$100.00.

If you search arround you can find threads where I have posted pages of curves. Below is a previous curve I ran when I owned a set of 2344's, since departed and now somewhere high in the lofty hills of the Sierras CA.

The 2344 is a lot of fun and can be tweeked up with the best hi end caps for startling results.

Enjoy.

Ian

Robh3606
11-26-2004, 07:24 AM
Hello Ralph

"Rob: That's a nice smooth looking curve on the RTA. Can you tell me what the vertical axis is?"

Thanks! But asside from the 2344 up top no. I can only guess. I have never really worried about it the speakers and inline in a vertical array with 24dbL/R so never thought of it as an issue. You do get some sit down stand up but they are tall and you end up on the "wrong side" of the 2344. Still sounds OK but its better up closer to the horn/midrange axis. Was always more worried about the power response and tried to use the criteria set up in line with Drew Daniels as far as crossover points, driver diameters and horn directivity. Idea being I could eq them and not totally mess them up on axis like the example in the paper on the 4430/4435 developement. Off axis it looks good with the HF rolling off as expected. So I think I did OK with it but to honest these are never going up the stairs and into the yard to find out;) When I run a sine sweep the room modes show up as the pinknoise can't excite them. They are as expected and fortunately very narrow only a couple of hertz wide. So that graph may look flat but all is not perfect by any means.
Ian and Bo are right using the RTA in room does have it's limitations. There is no substitute for a "real" measurement system that gates the room out. That said if you don't have one or can't afford one it's better than nothing. I have seen lots of people say they do there EQ's only by ear and I know for me it would never work. I need that reference point. Final tweeks are by ear but the overall gets set with the RTA.


"The speakers are a home brew setup with 2420 on 2344, E110 with paper domes in 0.5cuft boxes, 2231 reconed with 2235in 5cuft. I recently bought a DBX drive rack and now have lots of filters to play with! Up until now I had little in the way of EQ and mostly adjusted the driver levels with an active crossover and added some passive HF boost as per the crossover in the 4430's."


We have very similar set-ups. I have E145's, 2123 and 2416 with 2344 in an all active set-up. Subs, LE-14A'S, under the 145's. Use a charge coupled version of the 4435 compensation on the 2344. We are doing the same thing. I use the RTA to set driver levels. Above your upper crossover point to the 2344 you should be very good without any EQ.


"Given that the Fletcher-Munson effect is causing part of the extra harshness, then if I EQ by ear I'm going to end up with an EQ setting that is similar to the Flecther-Munson curves in the mid band. Not flat at all!"

You doing your EQ at your listenning level?? You know the E110 could be part of the problem. The 2123 is supposed to be flat. Well mine are not I have to notch them at 800hz. They have a rise there. 2122 does the same thing take a look. Where are you having a problem through the midband?? What are your crossover points??


Rob:)

ralphs99
11-26-2004, 07:01 PM
Hi Ian,

Sorry, I didn't get your comments about 2405 bashers at all. Maybe they were intended for a different audience....

I'll send a PM regarding the diaphrams. The more I think about the problem, the more the diaphrams seem like the probable culprit. I have to get over the mindset that speaker components are not consumables!

Your comment about Aluminium diaphrams having a characteristically harsh tone was interesting. The other threads I've looked indicate that the consensus of opinion favoured Aluminium over Titanium for smoothness. Or are you saying they become harsher with age? Also, what was your impression after replacing the diaphrams in your 2420's?

Thanks for the tip about Winairr. I'll check it out. I'd love to be able to produce graphs like yours!

Hi Rob,

I think you're spot-on about needing a reference point to do the EQ. I can EQ by ear, but its a very slow process. Cross-referencing measurements with what I'm hearing is the way to go for me.

That system of yours must really kick in the upper bass region! Where do you cross over between the 145's and the 2123's? Also, did you find much difference after adding a polarising voltage to the HF capacitors? From a scientific point of view I can't see the logic in it. But a number of posts have said it helps. Maybe I should just try it out and see for myself.

I can see the rising response in the whole series of 10" drivers, and the 12's and 15's as well. Part of the reason is the increasing DI with increasing frequency. I think the drivers are designed with constant power response as the aim, so on-axis response has to peak as the DI goes up to keep the power response flat. But that should be happening at a higher frequency; up around 1k5Hz for a 10". I have a TAD article that discusses the issue concerning their 15" drivers. They suggest a passive single pole LPF to drop the on-axis response. I don't know the reason for the peak at 800Hz on 10" drivers, but I agree that it should be notched out.

Most of my problems are in the range of 1k to 8kHz. A peaky, harsh sound with pink noise. It still sounds OK with music. But now that I know the harshness is there, I'm beginning to hear it in the music as well. Having said that, I still love listening to these speakers with well recoreded material. But listening to poor recordings can be frustrating; they uncover every blemish. My crossover point is currently set to 1k5Hz to match the DI of the 10" at the crossover frequency with the constant DI of the 2344. But I've experimented with the crossover frequency to little avail as the problems still persist higher up. I think that with new diaphrams I'll be able to tackle the other problems more effectively.

Cheers,

Ralph.

Robh3606
11-28-2004, 12:23 PM
Hello Ralph

"That system of yours must really kick in the upper bass region! Where do you cross over between the 145's and the 2123's? Also, did you find much difference after adding a polarising voltage to the HF capacitors? From a scientific point of view I can't see the logic in it. But a number of posts have said it helps. Maybe I should just try it out and see for myself."

They sound very quick through the upper bass the crossover is at 300Hz to the 2123. With the caps I am running a mix of Mylar and Polpropylene and it sounded smoother to me with the bias. Easy enough to try with a single cap in series give it a whirl. I think that harshness could be the diaphrams. I am running new titaniums in the 2416's and they sound fine.

Rob:)