Hi everyone,
So a long time ago I spent a lot of time here talking about making a 2 way using a 6.5" ScanSpeak Revalator. One of the controversial points was that as a short cut I did NOT use quasi-anechoic woofer measurements. In fact I deliberately measured with the cabinets in the room where I expected them to play.
I measured with OmniMic at 1M, using gated measurements through the crossover region.
I've moved about 3 times since then, bought a home and finally got down to measuring and the results in the bass were unexpectedly good. Given my experience I would encourage anyone building a 2-way to use a similar method. Am I being naive? Is there a part of the process that could have gone better which I got randomly lucky with?
I did use DSP to clip some peaks, and here is the average measurement from 3 points on my listening couch:
After these measurements were taken I did add a 4 dB boost shelf at 100Hz. There's a lot more detail in my blog post:
https://speakermakersjourney.blogspot.com/2021/12/room-speakers-eq.html
So my learned friends, given these results, and the reduction in complexity in attempting close mic measurements here, what's wrong with this short cut? Am I ruining the next generation of 2-way speaker builders by recommending this? Is this how you do it?
Best,
Erik
So a long time ago I spent a lot of time here talking about making a 2 way using a 6.5" ScanSpeak Revalator. One of the controversial points was that as a short cut I did NOT use quasi-anechoic woofer measurements. In fact I deliberately measured with the cabinets in the room where I expected them to play.
I measured with OmniMic at 1M, using gated measurements through the crossover region.
I've moved about 3 times since then, bought a home and finally got down to measuring and the results in the bass were unexpectedly good. Given my experience I would encourage anyone building a 2-way to use a similar method. Am I being naive? Is there a part of the process that could have gone better which I got randomly lucky with?
I did use DSP to clip some peaks, and here is the average measurement from 3 points on my listening couch:
After these measurements were taken I did add a 4 dB boost shelf at 100Hz. There's a lot more detail in my blog post:
https://speakermakersjourney.blogspot.com/2021/12/room-speakers-eq.html
So my learned friends, given these results, and the reduction in complexity in attempting close mic measurements here, what's wrong with this short cut? Am I ruining the next generation of 2-way speaker builders by recommending this? Is this how you do it?
Best,
Erik
The point of using less room in a measurement during design, is to hedge bets in your favor...it isn't a requirement. Designing for a specific room, leaves room for potential disappointment in some other configuration that is outside of the original design....that is...designing towards a FR, using less room, results in designing towards more headroom. Once again, not a requirement, more rather, a philosophy...
I think they say "too much headroom is just enough"?...something like that lol
I think they say "too much headroom is just enough"?...something like that lol
This is true, I had no way of knowing how this would play out 3 apartments later. In this case I got lucky.
But I did know more or less that I'd probably not be able to put the speakers far into the room, and in that sense my current placement mirrors that original situation. The room dimensions are totally different, but yes, they are pretty close to the rear wall still, but now have side walls in close proximity as well. I was lucky enough to have started with a number of acoustic panels and bass traps which I still use. Perhaps the acoustic panels being there from the start along with being at least close to the wall have kept my choices working for me.
But I did know more or less that I'd probably not be able to put the speakers far into the room, and in that sense my current placement mirrors that original situation. The room dimensions are totally different, but yes, they are pretty close to the rear wall still, but now have side walls in close proximity as well. I was lucky enough to have started with a number of acoustic panels and bass traps which I still use. Perhaps the acoustic panels being there from the start along with being at least close to the wall have kept my choices working for me.
Interesting. Anytime I see a rise in the response below the speakers cut-off its not real and caused by bad signal to noise ratio due to the gradually reduced amplitude, then I would improve the accuracy of the measurement. Listening to test tones starting at say 100hz and slowly decreasing, reveals the the speakers in room roll off. You are hearing a 30-20hz tone? Room modes would have to be strong to produce enough gain at the lower end of the roll off. I don't think in the blog you said it was eq'd below roll off. With enough bass traps room gain will be reduced.
I'll have to examine this further. Probably missing something. Wouldn't be the first time
I'll have to examine this further. Probably missing something. Wouldn't be the first time
I'm absolutely hearing / feeling 20 Hz, sinusoidal steady-state tones.
I should say that I tried re-measuring with both the OminiMic mic and a UMIK. At 20 Hz the two deviated betlow 25 Hz. Umik registered about 5 dB less, but by 30 Hz the two were in complete agreement.
I am absolutely not claiming these speakers will play bass loudly, BTW, but I don't listen to music that way. I've measured the bass limits and compression and published a chart with that near the end of my blog post.
I should say that I tried re-measuring with both the OminiMic mic and a UMIK. At 20 Hz the two deviated betlow 25 Hz. Umik registered about 5 dB less, but by 30 Hz the two were in complete agreement.
I am absolutely not claiming these speakers will play bass loudly, BTW, but I don't listen to music that way. I've measured the bass limits and compression and published a chart with that near the end of my blog post.
Last edited:
Erik - Your thread prompted me to go back and look at some measurements... When setting up my woofer cabinets and the active crossover from woofer to mid, I made free field non-gated measurements from 20 to about 800 Hz. ARTA calls this technique a dual gate measurement, and OmniMic calls it a blended measurement. I also made near field measurements which were adjusted to be 4-pi equivalent.
Here is something interesting: If I had used a free field non-gated technique for my low frequency assessment instead of the near field (quasi anechoic), I think I would have ended up with very close to the same DSP crossover result. So this confirms your finding... as long as the room is a good room, it can work.
j.
Here is something interesting: If I had used a free field non-gated technique for my low frequency assessment instead of the near field (quasi anechoic), I think I would have ended up with very close to the same DSP crossover result. So this confirms your finding... as long as the room is a good room, it can work.
j.
Gating isn't very useful for frequencies up to about 200Hz, of course depending on the gate you apply. Most of us don't get that beyond 10ms, so time/frequency/amplitude measurements of the lows are compromised anyway.
Next to that, LF systems behave quite like our theories, certainly with BR and CB systems. Measuring is mere confirmation (and a check on the execution).
Last but not least, with frequencies up to 300Hz, we are listening to the room essentially. Above that too, of course. But the lows are dominated by the uneven room response, which is in no way as easily to predict as our LF speaker system.
The dual gate ARTA (and others) offer is very useful when setting up a stereo system. In fact you only need the single gate measurement to get the speaker design itself right, because in that case you want to make sure you're not measuring the room.
Next to that, LF systems behave quite like our theories, certainly with BR and CB systems. Measuring is mere confirmation (and a check on the execution).
Last but not least, with frequencies up to 300Hz, we are listening to the room essentially. Above that too, of course. But the lows are dominated by the uneven room response, which is in no way as easily to predict as our LF speaker system.
The dual gate ARTA (and others) offer is very useful when setting up a stereo system. In fact you only need the single gate measurement to get the speaker design itself right, because in that case you want to make sure you're not measuring the room.
I did want to measure the room, actually, when I started. There were big room modes I had to mentally ignore when deciding on exactly whhow much I wanted to trade bass and efficiency in the low pass filter. When I say I gave up on quasi-anechoic, that's what I mean. And for me with a 2 way, in my rooms it's worked.
I don't really know what I'd do if I had to make a 3 way though, with a low pass filter between 100 and 200 Hz I could be in trouble.
I don't really know what I'd do if I had to make a 3 way though, with a low pass filter between 100 and 200 Hz I could be in trouble.
http://www.xlrtechs.com/dbkeele.com/PDF/Keele (1974-04 AES Published) - Nearfield Paper.pdfThis is a paper on nearfield measuring.
Hsu ships out a test CD, or they used to, which was from the Boston Audio Society. The first track has an organ and let me tell you, I absolutely could not hear or feel some of those notes, I could only tell they were happening because I could see the woofer's pumping.
So clearly this was well below the limit of audibility here.
So clearly this was well below the limit of audibility here.
Well I agree with Erik,
When designing a 2 way speaker, with a typical crossover response between 1-3Khz; what are the advantages of doing the near field?
You can get away with a single mic position, in fair field, at a distance equal to the narrowest dimension of the baffle eg. twice the baffle width; you don’t even need to a meter.
For the low end? well that fate was sealed by the bass tuning.
you can do a outdoor ground plane measurement to confirm, or
you can model it in your baffle diffraction simulator, or
You can measure in-room; which is probably more important, to your ears, than theory.
All 3 above will get be quite close.
how far away you put the speaker from the front wall it the main determinant of how much baffle step compensation you needed.
Or how much extra bass your ears like.
Whether it’s good enough will depend on your room and tastes.
howver, I would not say that it’s good practice. And certainly not say it to a newbie. You know these thread goes- it gets wild quickly.
I think it’s good practice to do it properly, if for no reason other than you taking your speaker for a demo to a friends place who doesn’t live in an apartment; or group listening demo room in a large space well away from walls, well that could be embarrassing…
When designing a 2 way speaker, with a typical crossover response between 1-3Khz; what are the advantages of doing the near field?
You can get away with a single mic position, in fair field, at a distance equal to the narrowest dimension of the baffle eg. twice the baffle width; you don’t even need to a meter.
For the low end? well that fate was sealed by the bass tuning.
you can do a outdoor ground plane measurement to confirm, or
you can model it in your baffle diffraction simulator, or
You can measure in-room; which is probably more important, to your ears, than theory.
All 3 above will get be quite close.
how far away you put the speaker from the front wall it the main determinant of how much baffle step compensation you needed.
Or how much extra bass your ears like.
Whether it’s good enough will depend on your room and tastes.
howver, I would not say that it’s good practice. And certainly not say it to a newbie. You know these thread goes- it gets wild quickly.
I think it’s good practice to do it properly, if for no reason other than you taking your speaker for a demo to a friends place who doesn’t live in an apartment; or group listening demo room in a large space well away from walls, well that could be embarrassing…
Last edited: