All,
I want to build my own 3 way closed box, active speakers. Is it worth me buying an SPL meter, or should I get some speaker measurement software?
I have a scope and signal generator already. The SPL meter I have in mind measures 35-130 dB, over the frequency range 31.5Hz to 8KHz and costs 135 euro. Any advice appreciated.
Mark
I want to build my own 3 way closed box, active speakers. Is it worth me buying an SPL meter, or should I get some speaker measurement software?
I have a scope and signal generator already. The SPL meter I have in mind measures 35-130 dB, over the frequency range 31.5Hz to 8KHz and costs 135 euro. Any advice appreciated.
Mark
I have used the Radio Shack SPL meter and a signal generator and that worked pretty well...but in a room with reflections, you don't get a very true frequency response on the meter. But owning the Radio Shack SPL meter makes it easy to use one of the speaker programs, such as ADUA's Speaker Workshop (free on the internet) which will measure speaker response. The Radio Shack SPL meter doubles as a preamplified Mike which you plug right in to your computer. Also, Old Colony has the Winair program ($50) which measures frequency response and you would use the Radio Shack meter plugged into your sound card here as well.
all,
thank's for your replies. i looked at the spec on the realistic unit and could not find a frequecy reponse. This concerns me as there's a model in the Maplin catalogue in the UK, which has a response of 300 Hz to 8KHz, which is not low enogh for my needs.
Mark
thank's for your replies. i looked at the spec on the realistic unit and could not find a frequecy reponse. This concerns me as there's a model in the Maplin catalogue in the UK, which has a response of 300 Hz to 8KHz, which is not low enogh for my needs.
Mark
ive been using the dazyweb tms1 from http://www.dazyweblabs.com/tms1/index.html
seems to do what is says on the tin and best of all its free
seems to do what is says on the tin and best of all its free
There is no question that if you want to design your own crossovers you need the ability to make quasi anechoic measurements, eg. MLS, TDS. You cannot do this with an SPL meter--you end up measureing the room, not the speaker. You also can't measure phase with an SPL meter. Phase is critically important for crossovr design. Get yourself a good measurement program and sound card and a panasonic WM60 mic capsule (cheap). It's not even absolutely critical that you callibrate the capsule as the mic is much flatter than you will ever get your speakers (unless you're doing DSP like I am).
John
John
All,
Thank's again for your replies.
John,
The object of this project is to investigate Rod Elliot of ESP's excellent bi-amping article. I also wish to measure how much Q, i prefer in a speakers bass response.
i will be using the ESP p9 electronic L-R crossover 24dB/oct which is phase coherent, so i was under the impression that i could build a pair of speakers using that, just setting them up in the frequency domain. Am i mistaken?
I was going to use peerless or audax drivers and compare the results to my conventional crossover speakers, which use volt and scanspeak drivers. I think DSP and digital amplifiers for the bass units is probably the way ahead in the long run, but i think i will start simple first.
themedhippy,
thanks for the dazyweb link, looks like a good starting point, when i git the mic input on the soundcard working.
Regards all
Mark
Thank's again for your replies.
John,
The object of this project is to investigate Rod Elliot of ESP's excellent bi-amping article. I also wish to measure how much Q, i prefer in a speakers bass response.
i will be using the ESP p9 electronic L-R crossover 24dB/oct which is phase coherent, so i was under the impression that i could build a pair of speakers using that, just setting them up in the frequency domain. Am i mistaken?
I was going to use peerless or audax drivers and compare the results to my conventional crossover speakers, which use volt and scanspeak drivers. I think DSP and digital amplifiers for the bass units is probably the way ahead in the long run, but i think i will start simple first.
themedhippy,
thanks for the dazyweb link, looks like a good starting point, when i git the mic input on the soundcard working.
Regards all
Mark
Hi Mark,
You wrote:
"i will be using the ESP p9 electronic L-R crossover 24dB/oct which is phase coherent, so i was under the impression that i could build a pair of speakers using that, just setting them up in the frequency domain. Am i mistaken?"
I'm not familiar with that particular crossover, but in general, a crossover that sums flat electrically will not sum flat acoustically. You need to offset the response of the drivers themselves as well as compensate for any physical offset. You could do that by trial and error until you get a flat frequency response, but unless you get lucky that would take billions and billions of years. A much less painful way is to measure the frequency and phase response of the individual drivers and then design a crossover that from the start will give you a good response. If you're going to do any more than a little playing around in this hobby, I really recommend investing in a decent measurement setup. You will be glad you did.
Regarding DSP, it's easier than you think. Assuming you've got a PC and a sound card with enough outputs (I use the M-Audio Delta 1010--highly recommended), all you need is a program called wavewarp (www.soundslogical.com). I could help you out designing some filters, assuming you can find a way to measure your speakers. I will say with absolute confidence that a digital crossover with the Delta 1010 will absolutely blow away any analog crossover--ruler flat magnitude and phase response on axis and much better off-axis response.
John
You wrote:
"i will be using the ESP p9 electronic L-R crossover 24dB/oct which is phase coherent, so i was under the impression that i could build a pair of speakers using that, just setting them up in the frequency domain. Am i mistaken?"
I'm not familiar with that particular crossover, but in general, a crossover that sums flat electrically will not sum flat acoustically. You need to offset the response of the drivers themselves as well as compensate for any physical offset. You could do that by trial and error until you get a flat frequency response, but unless you get lucky that would take billions and billions of years. A much less painful way is to measure the frequency and phase response of the individual drivers and then design a crossover that from the start will give you a good response. If you're going to do any more than a little playing around in this hobby, I really recommend investing in a decent measurement setup. You will be glad you did.
Regarding DSP, it's easier than you think. Assuming you've got a PC and a sound card with enough outputs (I use the M-Audio Delta 1010--highly recommended), all you need is a program called wavewarp (www.soundslogical.com). I could help you out designing some filters, assuming you can find a way to measure your speakers. I will say with absolute confidence that a digital crossover with the Delta 1010 will absolutely blow away any analog crossover--ruler flat magnitude and phase response on axis and much better off-axis response.
John
Hi John,
OK I’m convinced; I need to think about phase as well as magnitude, so now I have lots of questions for you!
You’ve recommended investing in a “decent measurement setup”, would you like to recommend a particular decent setup and give a ballpark cost if you know it?
If a can now measure phase error, how would I correct for it in my existing setup? Obviously I could alter the physical position of the separate drivers, but how do I correct errors within part of a drivers frequency range.
Also, you’ve mentioned that with a DSP solution, you get a “much better off-axis response”. I was under the impression that the off-axis response was a property off the drivers themselves, not the X-over. Could you please explain what I’ve missed? Sorry for having so many questions!
Regards
Mark
OK I’m convinced; I need to think about phase as well as magnitude, so now I have lots of questions for you!
You’ve recommended investing in a “decent measurement setup”, would you like to recommend a particular decent setup and give a ballpark cost if you know it?
If a can now measure phase error, how would I correct for it in my existing setup? Obviously I could alter the physical position of the separate drivers, but how do I correct errors within part of a drivers frequency range.
Also, you’ve mentioned that with a DSP solution, you get a “much better off-axis response”. I was under the impression that the off-axis response was a property off the drivers themselves, not the X-over. Could you please explain what I’ve missed? Sorry for having so many questions!
Regards
Mark
"You’ve recommended investing in a “decent measurement setup”, would you like to recommend a particular decent setup and give a ballpark cost if you know it? "
I'm probably not the best guy for that. I use Matlab and a Panasonic mic capsule, but I'm checking ebay weekly until a 1/4" measurement condenser mic setup shows up. I know there are some free programs out there, but I couldn't tell you where. The program from Liberty Instruments is really nice, but costs around $700. I am sure there are some programs in between these two extremes that would do nicely. For me this is something worth spending a few bucks on if you're going to be in this hobby for a while.
"If a can now measure phase error, how would I correct for it in my existing setup? Obviously I could alter the physical position of the separate drivers, but how do I correct errors within part of a drivers frequency range."
It's not so much a matter of correcting for it as taking it into account when you design the crossover. You adjust for it by adjusting the slope of the croosover. There are programs (I can't remember the names right now) that will design passive or active analog filters for you that do this automatically. You just have to input the measurement data. If you're designing a digitl filter, this adjustment is done through the use of an inverse filter which corrects any magnitude/phase response issues of the drivers and then you don't have to worry about them when designing the crossover--it's absolutely trivial to do once you know the math and I can help you out or anyone else here who is adventurous enough to give it a try.
"Also, you’ve mentioned that with a DSP solution, you get a “much better off-axis response”. I was under the impression that the off-axis response was a property off the drivers themselves, not the X-over. Could you please explain what I’ve missed? Sorry for having so many questions!"
In the crossover region, you get lobing effects off axis because of the physical offset of the drivers. This lobing effect is minimized if the two drivers are in phase on axis throughout the crossover region. The only practical way to do this is with DSP.
John
I'm probably not the best guy for that. I use Matlab and a Panasonic mic capsule, but I'm checking ebay weekly until a 1/4" measurement condenser mic setup shows up. I know there are some free programs out there, but I couldn't tell you where. The program from Liberty Instruments is really nice, but costs around $700. I am sure there are some programs in between these two extremes that would do nicely. For me this is something worth spending a few bucks on if you're going to be in this hobby for a while.
"If a can now measure phase error, how would I correct for it in my existing setup? Obviously I could alter the physical position of the separate drivers, but how do I correct errors within part of a drivers frequency range."
It's not so much a matter of correcting for it as taking it into account when you design the crossover. You adjust for it by adjusting the slope of the croosover. There are programs (I can't remember the names right now) that will design passive or active analog filters for you that do this automatically. You just have to input the measurement data. If you're designing a digitl filter, this adjustment is done through the use of an inverse filter which corrects any magnitude/phase response issues of the drivers and then you don't have to worry about them when designing the crossover--it's absolutely trivial to do once you know the math and I can help you out or anyone else here who is adventurous enough to give it a try.
"Also, you’ve mentioned that with a DSP solution, you get a “much better off-axis response”. I was under the impression that the off-axis response was a property off the drivers themselves, not the X-over. Could you please explain what I’ve missed? Sorry for having so many questions!"
In the crossover region, you get lobing effects off axis because of the physical offset of the drivers. This lobing effect is minimized if the two drivers are in phase on axis throughout the crossover region. The only practical way to do this is with DSP.
John
Unless the demand is for Bruel &Kjaer for some specific reason, Behringer makes a rather decent measurement mike, the ECM 8000, for appx than USD 100.
http://www.behringer.com/02_products/prodindex.cfm?id=ECM8000&lang=ger
The Panasonic capsules are not bad though...
check Linkwitzlab
http://www.linkwitzlab.com/sys_test.htm#Mic
or at Speakerbuildning
http://www.speakerbuilder.net/web_files/Articles/diymic/diymicmain.htm
http://www.behringer.com/02_products/prodindex.cfm?id=ECM8000&lang=ger
The Panasonic capsules are not bad though...
check Linkwitzlab
http://www.linkwitzlab.com/sys_test.htm#Mic
or at Speakerbuildning
http://www.speakerbuilder.net/web_files/Articles/diymic/diymicmain.htm
Do a search for software called TrueRTA. You can get a fully usable version for free by download. You can pay more for more functions but the free version is a great start.
Vic
Vic
mics
"Unless the demand is for Bruel &Kjaer for some specific reason, Behringer makes a rather decent measurement mike, the ECM 8000, for appx than USD 100."
It is indeed value for money. The advantage of the 1/4" B&K is it goes higher and has lower noise. I want to be able to do accurate distortion measurements, so this is important. They are just so dang expensive, though.
John
"Unless the demand is for Bruel &Kjaer for some specific reason, Behringer makes a rather decent measurement mike, the ECM 8000, for appx than USD 100."
It is indeed value for money. The advantage of the 1/4" B&K is it goes higher and has lower noise. I want to be able to do accurate distortion measurements, so this is important. They are just so dang expensive, though.
John
I definately see your point vs the B&K...
As a college student in the mid 70-s , I more or less grew up with the full blown B&K package of the time....
Still use B&K every now and then for SPL and vibration measurements, on a rented basis though.... Just cannot let go.....
BTW,- I wonder if someone out there has any distortion figures for the modified Panasonic capsules ( a la Linkwitz and others....??
As a college student in the mid 70-s , I more or less grew up with the full blown B&K package of the time....
Still use B&K every now and then for SPL and vibration measurements, on a rented basis though.... Just cannot let go.....
BTW,- I wonder if someone out there has any distortion figures for the modified Panasonic capsules ( a la Linkwitz and others....??
- Status
- Not open for further replies.