Freq response charts vs real world experience

[billfitzmaurice]

Since I think this is a great topic, I'll offer a few questions to see how close we can come to reaching a preliminary agreement as to what should be measured, and how the measuring and reporting should be conducted.

How and where should microphones be positioned?

Should more than one set of microphone positions be used? If so, should they be reported separately or averaged?

How and where should speakers be positioned?

In an effort to establish a basis for comparison, should an effort be made to measure and disclose the listening/testing environment separate and apart from the speakers themselves?

What kinds of signals, signal paths, and signal generating equipment shoud be used?

At how many different SPLs should measurements be taken?

Which frequency range(s) should be measured?

Which frequencies should be measured?

What kinds of deviations from "flat" should be considered remarkable?

Which measurements should be used to ascertain sensitivity?

All of these issues were settled by the Audio Engineering Society half a century ago. The electric instrument speaker industry is just a wee bit slow on the uptake.

Hey Bill, you publish other speaker measurements at you site to show a comparison to your designs. How about doing more?

Bring me the cabs. I'll measure them.

 

[DeepBass9]

The only important use of specs (IMHO) I haven't seen mentioned, is in the matching of different cabs.

Putting together a multiple cab set up that can be mixed in different configs or used all together would be very easy indeed, with accurate information upon which to base one's decisions.

But, too, I guess the tendency is to use multiple cabs from the same mfg'er....

 

[joegeezer]

The real problem is having all the cabs a guy might want to try, available to you, TO try. I live so far out of a city that I have to buy a cab, to try it, then sell it, if I don't like it. I have done just that pretty much. I ended up with Avatar cabs , as my choice because of dollars for performance. I'm a Ten's guy, I've found. I also love the sound of Eden cabs, but only use the single ten's, the CX110 Eden cabs. I have 4 of those , but have the Avatars, for my two four ten cabs.I found Eden's 410 cabs too heavy for me to deal with. Eden's cabs and others are said to 'color' the sound, if so, I like the color very much. I have owned Bergantino EX112 cabs, and liked them, but can't justify the sound per dollar for them, over an Avatar 112 cab. I even built two of Bill F's cabs. An Omni 210 cab and a T24 sub cab. Biggest waste of time and money I spent in a long time. Sounded like crap to me. Sold them to a guy in Florida, who posted on Bill's site how great they sounded. It's all in the ear of the owner. I like 8-10's slammin it at me. And , I like to play loud. Everybody is different. But it would help somewhat to have the charts, but it's best, to just fire one up, and see what YOUR ears says. Thats all that counts anyway.

 

[greenboy]

In the pro audio world, specs are not only an aid to determine what will suit one's needs and to make easy comparisons, but can be a real aid in building/configuring systems that best work in a chosen environment. But you have to know stuff to do this.

 

[A9X]

The only important use of specs (IMHO) I haven't seen mentioned, is in the matching of different cabs.

Putting together a multiple cab set up that can be mixed in different configs or used all together would be very easy indeed, with accurate information upon which to base one's decisions.

But, too, I guess the tendency is to use multiple cabs from the same mfg'er....

Yes. If you used say two Brand X 4x10's you'd know they were the same sensitivity across the spectrum and would match. However I see many people use two different brands and types of cabs with differing sensitivity and can see why they get weird results.

 

[Crockettnj]

Being the devils advocate for a moment, maybe it is possible that the specs posted by many mfgrs are not necessarily "pure BS". Perhaps they measured things with the cab in a corner of a small room, thereby taking advantage of a heck of a lot of (room gain?) extra SPL's from corner loading.

It would be nice to see the TB community agree on a standard, easy to implement test procedure, share cabs, and build a database.

Specifically what type of software and hardware would I need to do this? Or, is there a link to a post where this has been discussed?

thanks

 

[greenboy]

Being the devils advocate for a moment, maybe it is possible that the specs posted by many mfgrs are not necessarily "pure BS". Perhaps they measured things with the cab in a corner of a small room, thereby taking advantage of a heck of a lot of (room gain?) extra SPL's from corner loading.

This too comes up in a lot of these threads. And this too makes any specs derived from such a method totally useless since "room gain" is not an evenly frequency-distributed phenomenom - balances changed radically by source postioning, listener/mic positioning, room size and characteristics etc. Even just one company couldn't reliably supply continuity for comparisions within their own brand name using such a method - not ones that could be relied on anyway.

So that makes such a method no more preferable than just making up "wishful stinking" numbers... because as Bill said, the electric bass industry is just a little slow on the uptake - methodology/standards that give meaningful results have been around for decades.

 

[DeepBass9]

Yes. If you used say two Brand X 4x10's you'd know they were the same sensitivity across the spectrum and would match. However I see many people use two different brands and types of cabs with differing sensitivity and can see why they get weird results.

And, as you could probably tell, I'm in the differing brands category. I'm fairly happy, considering I educated myself AFTER I got my 2 of my 3 cabs. I guess I should change happy to lucky. But, having said that, if I didn't have a 2 channel poweramp, I'd feel screwed. As it is, I can match perceptible volume by cranking the lower side up. But the obvious byproduct = lack of head room. So that makes me very . The good part is, that between my 3 I've got the freq. spectrum covered in the manner I like..so....if I can find myself a replacement for my odd man out, and it's got the same freq. response, just a higher sensitivity, I'd be right where I want to be.

Which leaves me pining for accurate specs, as well as accurate tonal coloration description of cabs (which seem to be easier to get, if you trust perception), to help narrow my field.

Oh, and a bag of free cash would help!

 

[A9X]

Being the devils advocate for a moment, maybe it is possible that the specs posted by many mfgrs are not necessarily "pure BS". Perhaps they measured things with the cab in a corner of a small room, thereby taking advantage of a heck of a lot of (room gain?) extra SPL's from corner loading.

That's a bit like saying my Trabant did 128mph yesterday. Without mentioning I'd pushed it out the back of a plane.

 

[A9X]

And, as you could probably tell, I'm in the differing brands category. I'm fairly happy, considering I educated myself AFTER I got my 2 of my 3 cabs. I guess I should change happy to lucky. But, having said that, if I didn't have a 2 channel poweramp, I'd feel screwed. As it is, I can match perceptible volume by cranking the lower side up. But the obvious byproduct = lack of head room.

But that's not how most people do it. Usually it's just a number of parallel wired boxes connected to the output of the same amp channel.

 

[greenboy]

Before I knew the little I know I was able to coax somewhat pleasing sounds out of a lot of gear {at least sometimes ; }, and was able to make some compelling purchases. Nonetheless, I am greatful for honest specs and the honing of my understanding and hearing. I learned a bit more.

 

[DeepBass9]

But that's not how most people do it. Usually it's just a number of parallel wired boxes connected to the output of the same amp channel.

Well, that's how I'm doing it, too, for my two 8 ohms. It's just that my they have higher sens, so they're louder at the same volume, than my one 4 ohm cab (which I run on the 2nd channel to keep the load evenly distribute, impedance-wise). So, I crank the 4 ohm up higher. Problem is, it taps out well below the highest output of my two 8'ers. Thus, loss of headroom. Speaking for my 3 cab setup, of course. When I run a two cab setup, I use the 8's and parallel them to 4 on one channel. Similar sens at the same impedance = similar volume levels.

 

[Petebass]

A couple of points:-

* 2 cabs can have the exact same frequency response chart, but still sound different. If their frequency response AND Phase graphs are identical, which incidentally is a near impossibility, then they will sound the same. That certainly puts a dent in the implication that frequency response graphs can replace your ears, doesn't it! Now if that ain't a reason to go back and read joelb79's post (Post Number 29), nothing is.

* Frequency response charts require testing under conditions which are simply unrealistic. OK we all do the occasional outdoor gig, but when was the last time you did a gig in an anechoic chamber? Having said that, the variables for indoor gigs are endless in terms of room acoustics. What's the solution? Learn to identify "Your Sound", then learn how to dial it in! (Go read Joelb79's post a third time). I have this novel idea, which I'm hoping will catch on, whereby you "Listen" to the cab before buying it. No I'm not too old to accept Internet shopping is a part of life now, I just refuse to accept that you can buy EVERYTHING on the internet without exception, and speaker cabs fall into that category of things you need to audition first.

* AX9 I'm happy to assist/participate in a Sydney cab-shootout, but can we include a series of real world gig tests and reviews to accompany the frequency response charts? After all, if it sounds good at a gig, it sounds good where you need it to! I'm a working, gigging bass player, and my opinions reflect that......

 

[A9X]

Since I think this is a great topic, I'll offer a few questions to see how close we can come to reaching a preliminary agreement as to what should be measured, and how the measuring and reporting should be conducted.

How and where should microphones be positioned?

Should more than one set of microphone positions be used? If so, should they be reported separately or averaged?

How and where should speakers be positioned?

In an effort to establish a basis for comparison, should an effort be made to measure and disclose the listening/testing environment separate and apart from the speakers themselves?

What kinds of signals, signal paths, and signal generating equipment shoud be used?

At how many different SPLs should measurements be taken?

Which frequency range(s) should be measured?

Which frequencies should be measured?

What kinds of deviations from "flat" should be considered remarkable?

Which measurements should be used to ascertain sensitivity?

Worth doing? Yes. Challenging? Certainly?

Your questions overlap a bit so I'll answer how I'd do it rather than your Q's individually.

Test equiptment; PC, soundcard (in my case M-Audio Audiophile 24/192), adjustable buffer, Quest QA3004 poweramp. Received by a mic that's been calibrated or there is a reference file for, mic preamp and soundcard. Software could be anything from Rightmark to Praxis. As I mentioned earlier I'll be getting Soundeasy and my friend has TrueRTA which is easy to use and not that expensive for the full version.

I have a Behringer ECM8000 and will likely also buy an IBF-Akustik EMM-8 microphone as well.

First test I'd run would be an impedance sweep to ascertain the port tuning and the true impedance of the speaker. This will be useful so as not to overdrive the speaker below it's tuning in higher power tests.

LF measurements are easiest outdoors in an open relatively flat area such as a carpark. Measurement mic can be placed on the ground with a windsock to reduce noise pickup at a fixed distance of say 1m. LF measurements could be several say 20-300Hz sweeps averaged to reduce noise pickup. Polar patterns aren't much use this low. If the box is rear ported, it should be turned around and the measurements taken again and added to the front measurements.

Then I'd do a slow sine sweep 20-20kHz at the reference level as defined below. As there are many configurations of cab layout, the mic could be placed at a few different levels, eg cone centre on a single driver cab, centre of a 4x10 or midrange on a multiway box. Then I'd do repeat sine sweeps (above box tuning) in 3dB power increments until there's no further increase which would show power compression and max output.

Sensitivity should be measured with a 2.83V signal at 400Hz at the speaker terminals, and that used as a baseline reference for each box.

I'd also like to do some distortion tests at a number of spot frequencies at the 2.83V reference for that box and the 3dB power increments.

Horizontal polar patterns via a sine sweep at 2.83V using the same height as earlier (and noted so ground reflections / cancellation can be determined) at 1m distance, with the angle determined from the front vertical centreline of each box. Angle increments, maybe 10*, from on-axis to 50* sound OK?

I'd also like to take the boxes indoors, and do a gated measurement to get cumulative spectral displays or waterfall charts as they reveal a lot of time domain information that's easily missed in FR graphs.

Once the test rig was set up, with a couple of people to do the moving of boxes and mics, each set of measurements shouldn't take too long.

 

[greenboy]

That certainly puts a dent in the implication that frequency response graphs can replace your ears, doesn't it!

Did someone imply this? I didn't know.

* Frequency response charts require testing under conditions which are simply unrealistic. OK we all do the occasional outdoor gig, but when was the last time you did a gig in an anechoic chamber?

Cute. But the purpose of graphs and industry-standard testing among other things is to supplement the ear's findings - not to replace them. As such, they need to be conducted in such a way as to get local-environment variables out of the way since no two local environments are the same. I've met a great many people, by the way, who can really tell a lot about a system by running through honest comprehensive qualified specs, and really maximize the performance of the gear by referring to them. They tend to be people with sound reinforcement and studio backgrounds however. But there's room for more than one approach.

That said, I don't expect that a majority of the bass/MI portion of the industry could ever be coaxed to spend the time and the money on supplying honest qualified specs - and why would they? - the audience that could actually make good use of them is limited anyway.

 

[A9X]

* AX9 I'm happy to assist/participate in a Sydney cab-shootout, but can we include a series of real world gig tests and reviews to accompany the frequency response charts? After all, if it sounds good at a gig, it sounds good where you need it to! I'm a working, gigging bass player, and my opinions reflect that......

Sure, I'm all for subjective opinions too, but I like the consistency of a decent set of measurements (not just FR) so that gear that's not available locally for audition and testing can be compared and some decisions made on whether to travel for auditioning is worth it or not. Handy for people who live away from major cities or are thinking of of importing an item not availbable in that country.

I also think it'd be educational for some people about what to listen for.

Something I also thought might be interesting, would be to record each of these cabs at a couple of different levels with a reference bass track that's taken off CD. The files could be sent all over the world and used similar to a test tone as a consistent reference.

 

[A9X]

PS: I'd also like to do this with preamps, heads and poweramps. Get the baseline FR, EQ sweeps, output levels, noise and THD spectra. Power levels could be measured then backed off 1dB and pulsed to 3dB above max to look at how they behave on transient overload. This could make interesting reading especially with the tube/SS debate over headroom.

 

[billfitzmaurice]

.

* Frequency response charts require testing under conditions which are simply unrealistic.

Not at all. They only require an echo free environment, so that you measure the cab, not the room. Most manufacturers use anechoic chambers because their factories tend to be located too close to other buildings and/or major roadways that would contaminate the readings. OTOH, Danley Sound Labs uses their parking lot; I use my backyard.

 

[fdeck]

A couple of points:-

* 2 cabs can have the exact same frequency response chart, but still sound different. If their frequency response AND Phase graphs are identical, which incidentally is a near impossibility, then they will sound the same. That certainly puts a dent in the implication that frequency response graphs can replace your ears, doesn't it! Now if that ain't a reason to go back and read joelb79's post (Post Number 29), nothing is.

Perhaps at high frequencies. But in the frequency range below a few hundred Hz, where the basic electromagnetic theory is quite accurate, two mainstream speakers can't have the same frequency response but different phase response. The proof is here:

Invalid Link Removed

I suspect that at high frequencies, phase response of speakers is a red herring. But I could be proven wrong.

* Frequency response charts require testing under conditions which are simply unrealistic. OK we all do the occasional outdoor gig, but when was the last time you did a gig in an anechoic chamber? Having said that, the variables for indoor gigs are endless in terms of room acoustics. What's the solution? Learn to identify "Your Sound", then learn how to dial it in! (Go read Joelb79's post a third time). I have this novel idea, which I'm hoping will catch on, whereby you "Listen" to the cab before buying it. No I'm not too old to accept Internet shopping is a part of life now, I just refuse to accept that you can buy EVERYTHING on the internet without exception, and speaker cabs fall into that category of things you need to audition first.

Horror of horrors, I don't even bother with an anechoic chamber. I pretty much proceed from a combination of theory for low frequencies, and the driver manufacturer's published curve for high frequencies, while making educated guesses about how the speaker performs in a real world environment. I work with driver makers whose specs are trusted to be accurate.

When I conduct actual measurements of "unknown" speakers, I use a near-field method, which provides me with the information that I consider to be most important.

Still, we don't all live in big cities with a bass speaker store on every corner. Ignoring giant speakers that are geared towards loud rock music, auditioning a speaker is a day trip for me. If I want to try the speaker at a gig, that's two day trips. Day trips involve negotiating with my spouse over managing the kids. Requiring me to go out of my way is not how to earn my business.

Response curves would allow me to pare down the list of speakers that are worth testing by ear. The listening test still occurs, and I still have to tinker a bit with EQ at gigs, but I don't waste my time buying speakers in a way that is intended solely for the convenience of speaker makers.

I am the customer, and the customer is king. I demand what I want. One speaker maker -- Eminence -- is willing to pamper me in order to earn my business. The only proviso is that I have to supply the plywood. That's still less time and effort than driving to Chicago.

* AX9 I'm happy to assist/participate in a Sydney cab-shootout, but can we include a series of real world gig tests and reviews to accompany the frequency response charts? After all, if it sounds good at a gig, it sounds good where you need it to! I'm a working, gigging bass player, and my opinions reflect that......

Real world gig tests would be great. Part of the "research" is figuring out what aspects of measured performance correlate with listening and playing experiences.

 

[seamonkey]

A couple of points:-

* 2 cabs can have the exact same frequency response chart, but still sound different. If their frequency response AND Phase graphs are identical, which incidentally is a near impossibility, then they will sound the same.

To me it seems that if two cabinets measure the same but sound different, then you've missed a measurement.

I recall reading somewhere that phase alignment between low and high frequency drivers was a red herring. It didn't matter to real hearing. Like in an orchestra playing a chord, it doesn't matter how far the piccolo is from the double bass. This does mean that two complex waveforms, can look different but sound the same, being that the complex waveform is the sum of all the frequencies together, including their phase.

In any case, it seems that many measurements could be taken, and people could listen to the cabinets and determine what's important to them. And then have a better idea what the next cabinet they buy needs to do.