Well, I couln't stand it any longer, so I finally got an SPL meter (Ratshack digital), and before I post the results into my project thread, I need some advice. Yesterday, I moved my X-baffles with the Fostex FE108EZ sigmas into my basement office and took a bunch of near-field measurements. I measured both the horn mouth and the driver. I got some surprising and not so surprising results. Overall, this project looks very promising. There are no huge holes or spikes. The changes are quite smooth over the entire spectrum...BUT:
First the good news. It appears the horn works spectacularly well. It's only down 4 dB at 70Hz and has a rounded peak of +6 dB around 175Hz, then rolling off gradually past 500Hz.
The bad news: The midrange rises steadily from reference at 1000Hz to a peak of +16 db around 7000Hz. I knew these were a bit bright, but not that bright, and now I know why I prefer them shoved up against the wall. Again, the rise is very gradual and consistent but I feel just a bit too much overall.
My question is whether it is possible to tame this midrange rise a bit without using a notch filter or BSC? What are my 'natural' options. Would changing the compression chamber help, or perhaps the throat opening? What about playing with the line stuffing? Damping the basket? Was my testing methodology sound? I really need some experienced minds working with me here. I have a lot of unanswered questions around this initial test.
I don't necessarily want to eliminate this rise, as that is what makes this driver sound so special. I just want to tame it a bit.
Methodology: Fast C-weighting. Reference dB set at 1000Hz. 1/12 octave samples from 50Hz to 500Hz on the horn mouth and 125Hz to 17000Hz on the driver (outside those ranges, it was always down more than 10 dB, so I stopped). NCH Tone Generator software using sine wave. Meter position was 10 cm from the driver and about 1.5 inches from the center of the horn mouth, and I used a tripod. The room is pretty bare, low ceiling, sheetrock walls, and hardwood floor. About 120 sq. feet, shaped somewhat triangular. Lots of slap echos. The speaker was placed at least three feet from any wall.
Right now, all I have are columns of numbers. I hope to get them into a spreadsheet soon.
Thanks,
Doug
First the good news. It appears the horn works spectacularly well. It's only down 4 dB at 70Hz and has a rounded peak of +6 dB around 175Hz, then rolling off gradually past 500Hz.
The bad news: The midrange rises steadily from reference at 1000Hz to a peak of +16 db around 7000Hz. I knew these were a bit bright, but not that bright, and now I know why I prefer them shoved up against the wall. Again, the rise is very gradual and consistent but I feel just a bit too much overall.
My question is whether it is possible to tame this midrange rise a bit without using a notch filter or BSC? What are my 'natural' options. Would changing the compression chamber help, or perhaps the throat opening? What about playing with the line stuffing? Damping the basket? Was my testing methodology sound? I really need some experienced minds working with me here. I have a lot of unanswered questions around this initial test.
I don't necessarily want to eliminate this rise, as that is what makes this driver sound so special. I just want to tame it a bit.
Methodology: Fast C-weighting. Reference dB set at 1000Hz. 1/12 octave samples from 50Hz to 500Hz on the horn mouth and 125Hz to 17000Hz on the driver (outside those ranges, it was always down more than 10 dB, so I stopped). NCH Tone Generator software using sine wave. Meter position was 10 cm from the driver and about 1.5 inches from the center of the horn mouth, and I used a tripod. The room is pretty bare, low ceiling, sheetrock walls, and hardwood floor. About 120 sq. feet, shaped somewhat triangular. Lots of slap echos. The speaker was placed at least three feet from any wall.
Right now, all I have are columns of numbers. I hope to get them into a spreadsheet soon.
Thanks,
Doug
Hi Doug. Just like you I'm not a great fan of "notch filter" Here's some ideas that might help.
1- Try to take the same measures you did but at half way to your listening position. (at 10cm it's almost useless)
2- Try the speakers close to rear wall (it will reinforce bass/mid-bass and should give better balance)
3- Play with the stuffing to get more bass/mid-bass.
4- Try a felt ring around the drivers (I've used it many times (with good result) to tame bright tweeters.
Perhaps it's a good idea to implement only one step at a time and evaluate the result.
Good luck and give us news on the result !!!
1- Try to take the same measures you did but at half way to your listening position. (at 10cm it's almost useless)
2- Try the speakers close to rear wall (it will reinforce bass/mid-bass and should give better balance)
3- Play with the stuffing to get more bass/mid-bass.
4- Try a felt ring around the drivers (I've used it many times (with good result) to tame bright tweeters.
Perhaps it's a good idea to implement only one step at a time and evaluate the result.
Good luck and give us news on the result !!!
Thank you nikita,
I'm trying really hard to get a handle on testing speakers. I read so much yesterday on the net that I gave myself a headache. I see now that near-field testing is only good to 500Hz. I was trying to eliminate room effects. At least the horn mouth results are probably accurate (which I think are great, by the way).
Okay, so first I have to retest.
Any advice on eliminating room effects? I guess I will test at 1 meter. I most want to measure the speaker itself right now, and I really don't trust anything I hear in the room the computer is located. I could move the speaker outside with about 20 feet of speaker wire. I understand outdoors is a form of anechoic chamber. Would this give me more accurate results?
Doug
I'm trying really hard to get a handle on testing speakers. I read so much yesterday on the net that I gave myself a headache. I see now that near-field testing is only good to 500Hz. I was trying to eliminate room effects. At least the horn mouth results are probably accurate (which I think are great, by the way).
Okay, so first I have to retest.
Any advice on eliminating room effects? I guess I will test at 1 meter. I most want to measure the speaker itself right now, and I really don't trust anything I hear in the room the computer is located. I could move the speaker outside with about 20 feet of speaker wire. I understand outdoors is a form of anechoic chamber. Would this give me more accurate results?
Doug
Taperwood said:
I too am starting to learn about speaker measurements and have had some difficulties making good measurements.
A couple of notes. With my Radio Shack 33-2050 SPL meter I use a C weighting and SLOW response. Also, the Radio Shack meters are not very linear. Are you using any corrections?
What I struggle the most with, is the distance from the driver. Here is a little excerpt from Zaph | Audio (http://www.zaphaudio.com/setup.html)
So 10 cm seems a little close. I suspect that a 4" full range may have to be measured in two steps and this is the problem I have also been seeing in my measurements.
Good luck and keep us posted.
Regards,
Gio.
Hi Doug, a felt ring is a ring of felt place around the driver's frame outside of the box, glued on the baffle. Like I said before, sometimes very effective with bright tweeters but I never used it on single fullrange!!!
Your drivers do not measure so bad, with a little rise from 2.2K and a more severe peak at 14K,
Don't loose your time in measures outside or on the moon, it's not where you're listening your speakers! I heard many speakers measuring very flat in anechoic chamber but sound awfull in my house! (including my present speakers (Magneplanar) they are very room dependant)!
Maybe your room is too lively (many windows, hard walls/floor, hard furniture) if you can put some absorbing material in your room (carpet, heavy drapes, cork panels, whatever absorbing...) and see what happens.
BTW what is your amplification system?
Good luck!
Your drivers do not measure so bad, with a little rise from 2.2K and a more severe peak at 14K,
Don't loose your time in measures outside or on the moon, it's not where you're listening your speakers! I heard many speakers measuring very flat in anechoic chamber but sound awfull in my house! (including my present speakers (Magneplanar) they are very room dependant)!
Maybe your room is too lively (many windows, hard walls/floor, hard furniture) if you can put some absorbing material in your room (carpet, heavy drapes, cork panels, whatever absorbing...) and see what happens.
BTW what is your amplification system?
Good luck!
I'll jump in real quick. Thanks everyone so far.
John, the big problem is that I measured near-field, which I have found out is almost useless above 500Hz. Until I know if there is a problem, I can't address the cause (we can only trust my ears so much
). So, I'm now back to square one. Apparently, the horn measurements will be okay, though.
How about this everyone: If I scrounged every blanket and sleeping bag in the house and spread them on the floor and walls, etc., and then placed the speaker in the middle of the room, at least three feet from any walls, and measured at 1 meter, would I stand a reasonable chance of getting accurate measurements without much room interference?
Doug
John, the big problem is that I measured near-field, which I have found out is almost useless above 500Hz. Until I know if there is a problem, I can't address the cause (we can only trust my ears so much
). So, I'm now back to square one. Apparently, the horn measurements will be okay, though.How about this everyone: If I scrounged every blanket and sleeping bag in the house and spread them on the floor and walls, etc., and then placed the speaker in the middle of the room, at least three feet from any walls, and measured at 1 meter, would I stand a reasonable chance of getting accurate measurements without much room interference?
Doug
nikita said:
Yes, the room I'm measuring in is very lively. The living room is more 'normal.' Unfortunately, measuring in the living room would require moving a computer with VERY deep roots.
My amp is a Sonic Impact Super-T. Got it about two months ago. Quick assessment is that it's a dryer sound but doesn't add any brightness. I like it.
Doug
Re: Re: Taming rising frequency response 'naturally'...
You really can't hear a peak other than it can sometimes sound a bit "hissy." But I still need to figure out the best and fairest way to measure because my initial readings are wrong.
I hope it doesn't come to using a circuit, but until I get some accurate measurements, I really won't know. It's frustrating that there are no standards for measuring speakers.
Doug
lndm said:
You really can't hear a peak other than it can sometimes sound a bit "hissy." But I still need to figure out the best and fairest way to measure because my initial readings are wrong.
I hope it doesn't come to using a circuit, but until I get some accurate measurements, I really won't know. It's frustrating that there are no standards for measuring speakers.
Doug
I wouldn't say there are no standards. I think your results will just improve as you understand what you're up against (mine do). Measuring too close to the baffle won't register the effect of the baffle, too far and the room comes into play. BUT, the results will still be somewhat predictable with a little experience.
There are standing waves and reflections. Standing waves usually mess up 200Hz and down. Sometimes they are obvious and you can draw a line straight through them (using your best estimate and intuition).
Reflections are often an issue at higher frequencies. You can pad your walls. You can do things like: measure the distance from the bass driver to the floor and back up to the mike (as a reflection), then subtract the direct distance. Where the difference is a wavelength you will measure a hump, half wavelength you will measure a dip etc.
The MLS method uses a predetermined noise burst that tries to play all frequencies in a limited time. The measuring software will notice when the signal gets repeated that it is a reflection and ignore that.
There are standing waves and reflections. Standing waves usually mess up 200Hz and down. Sometimes they are obvious and you can draw a line straight through them (using your best estimate and intuition).
Reflections are often an issue at higher frequencies. You can pad your walls. You can do things like: measure the distance from the bass driver to the floor and back up to the mike (as a reflection), then subtract the direct distance. Where the difference is a wavelength you will measure a hump, half wavelength you will measure a dip etc.
The MLS method uses a predetermined noise burst that tries to play all frequencies in a limited time. The measuring software will notice when the signal gets repeated that it is a reflection and ignore that.
Indm, Thank you. Your post helped a lot in helping me focus on this issue.
Martin, you would ask the hardest question. I guess it's because it is my design, and until I can squeeze every last drop of performance out of it, I don't want to introduce anything external.
I didn't say I would not use circuitry, but I'm not ready to throw in the towel yet. I trust my ears. When I am done, everyone who hears them will at least say "yea, they sound pretty darn good." Of that I am confident. Whether they say "WOW", however, is up to the individual's taste.
I hope this answers your question. Besides, I'm having a ton of fun learning all this stuff.
Doug
MJK said:
Martin, you would ask the hardest question
. I guess it's because it is my design, and until I can squeeze every last drop of performance out of it, I don't want to introduce anything external.I didn't say I would not use circuitry, but I'm not ready to throw in the towel yet. I trust my ears. When I am done, everyone who hears them will at least say "yea, they sound pretty darn good." Of that I am confident. Whether they say "WOW", however, is up to the individual's taste.
I hope this answers your question. Besides, I'm having a ton of fun learning all this stuff.
Doug
Here are some things Ted Jordan has written on this topic (emphasis mine):
A further issue to be addressed is the sound distribution pattern throughout the listening room. The wide random sound spread favoured by convention cannot support stable stereophonic imagery. JORDAN drivers have a slightly directional and gently rising high frequency response and should be placed with their axes crossing well in front of the listener. This approach results in a substantial improvement in image stability and reaches its full potential in our 'Linear Array' designs. Again, despite convention, there is strong argument in favour of placing the loudspeakers as close to the wall as possible. This reduces the time lag between the direct sound and the reflection from the 'virtual images' and substantially improves coherence and spatiality. (The ideal, but usually impractical, solution, is to mount the drivers in the wall where there would be no virtual image).
http://www.ejjordan.co.uk/basics.html
A further issue to be addressed is the sound distribution pattern throughout the listening room. The wide random sound spread favoured by convention cannot support stable stereophonic imagery. JORDAN drivers have a slightly directional and gently rising high frequency response and should be placed with their axes crossing well in front of the listener. This approach results in a substantial improvement in image stability and reaches its full potential in our 'Linear Array' designs. Again, despite convention, there is strong argument in favour of placing the loudspeakers as close to the wall as possible. This reduces the time lag between the direct sound and the reflection from the 'virtual images' and substantially improves coherence and spatiality. (The ideal, but usually impractical, solution, is to mount the drivers in the wall where there would be no virtual image).
http://www.ejjordan.co.uk/basics.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.