Assuming the phase plug is ~13543/pi/6000 = 0.72" dia., then here's your dip between it and the whizzer.
Last night and this morning I played with the KEF LS50 Meta SP1793 coax drivers (OEM $150/pr) and it was a learning experience. The alloy cone/waveguide contributed a lot more HF energy than the central tweeter, whenever off-axis. Sure the cone breakup will be suppressed by LPF and effectively a notch filter, but the cone/waveguide was firing/bouncing strong 8-12khz directly off-axis even without tweeter playing. My point is don't ignore the cone as a widefield reflector.
What are the distances of A - center of speaker to the top, B - center to side, and C - center to the middle of suspension rubber as in https://www.diyaudio.com/community/threads/6-5-full-range-dip-at-6k.415909/post-7753871 ?
Straight shortest lines, without bends
Last edited by a moderator:
No to all of this.
This is measurement error created by the phone.
It has multiple microphones. And you dont know which one is being used.
There is a microphone on the back by the camera lens.
There is a microphone on the top of the screen
There is microphones on the bottom of the phone next to speaker holes.
We have no idea what one is being used, we have no idea if he is on axis or off axis.
we have no idea if house curve removes typical notch filters for phones
Some or most phone mics for voice are mounted 20 to 30 degrees off axis inside the phone.
Some or most phone mics have a Notch filter ironically usually at 5k to 8k
usually close to -10 dB
Depending if you hold the phone using the voice call mics
this would be bottom of the phone where speaker holes are at.
Even " on" axis they will likely be 30 degrees off axis inside the phone
As with any and ALL speakers if you are off axis to a 6.5 inch speaker
you will have a huge notch at 5 k
This will increase even more with a ground plane and one side wall.
Any wild guessing from wide baffle or dayton nonsense blah blah
goes straight out the window.....
Move along to the usual = Phones are not audio measurement devices
and bash away iphone 14+
because you samsung android yada yada yada blah blah does the same thing.
This is measurement error created by the phone.
It has multiple microphones. And you dont know which one is being used.
There is a microphone on the back by the camera lens.
There is a microphone on the top of the screen
There is microphones on the bottom of the phone next to speaker holes.
We have no idea what one is being used, we have no idea if he is on axis or off axis.
we have no idea if house curve removes typical notch filters for phones
Some or most phone mics for voice are mounted 20 to 30 degrees off axis inside the phone.
Some or most phone mics have a Notch filter ironically usually at 5k to 8k
usually close to -10 dB
Depending if you hold the phone using the voice call mics
this would be bottom of the phone where speaker holes are at.
Even " on" axis they will likely be 30 degrees off axis inside the phone
As with any and ALL speakers if you are off axis to a 6.5 inch speaker
you will have a huge notch at 5 k
This will increase even more with a ground plane and one side wall.
Any wild guessing from wide baffle or dayton nonsense blah blah
goes straight out the window.....
Move along to the usual = Phones are not audio measurement devices
and bash away iphone 14+
because you samsung android yada yada yada blah blah does the same thing.
i use spectroid for measurement and it reproduces well what I know from professional measurement system
upper frequency limit was correct with spectroid with another app not.
upper frequency limit was correct with spectroid with another app not.
You cannot mix all the problems and solve with one single solution, you need to separate them:
1. Personal listening experience: can you hear that dip as in your first post - at ~5.9kHz? It is a pretty narrow dip and pretty far from the voice range, heavily into harmonics territory and even from some instruments territory also. Such narrow and deep dips can be heard only after you know that they are here. In normal listening - very likely you will not pay attention to it at all... Play a sweep of 20Hz to 20kHz from youtube, and I am almost sure, that it will be not the biggest problem you will hear in your system. Solve bigger problems first.
2. Presentation of the measurement software - it looks for me, that the software you are using does not have option to choose graph smoothing range. Some of the pretty good sounding systems will show big dips/peaks depending on the simple smoothing option. REW software, which is totally free we mostly use here, has an option to choose smoothing option to show in graphs and also many more, which does impact how the SPL graph visually will look like
Here is the same measurement graph of the speaker while using 1/48 octave smoothing and 1/3 octave smoothing. Can you see the dip you are interested in in 1/48 grpaph? 1/24? 1/12? In my screenshots It is a really bad speaker design in totally unsuitable enclosure, bad measurement conditions, but you get the idea.
3. The problems in the ~500Hz range can be several:
a) Room related. Room influence starts from ~1kHz as very minor down to hearing range of ~20Hz as major. You can see if it is a room related by moving, for example, speaker further or closer to the wall, also moving the same amount measurement microphone to leave measurement geometry the same. Most likely you cannot change the listening room, but you may change listening position and placement of the speakers a little. Special room treatment is another topic of headache. If you have natural room with more-or-less unobstructed path from speaker to listening position, soft furniture, as sofa, carpet, some bookshelves on the walls and so on - you are OK. Probably you can smooth some room modes with little EQ. 0,7m wavelength of 500Hz is also very much into the multiples of the floor, rear, sidewall, some random coffee table which is in front of the speaker bounces...
b) 500Hz range can be influenced by the spider resonance and basket on the opposite side of the speaker. This is not easy to solve, also this is not the thing to focus first - you have plenty of more important things.
c) Also this may be because of sound reflection from the back or side or top/bottom panel inside your cabinet. Using volume filling can solve this - you probably already done that well. This is another very widely discussed topic.
d) 500Hz is ~0.7m wavelength, so this also very related to your external cabinet dimensions... This can be the price to pay for having wide front baffle.
4. Prioritize quality over SPL smoothness: Do not focus blindly on the SPL graph dips. You will not achieve flat graphs with single driver full range system in normal living room. Focus on the parasitic vibrations. Your cabinet panels are of ~1cm thick plywood, what is considered as very sturdy and well built by the commercial speaker standards, but if you listen at 3+W sound levels - it needs internal cross bracing and probably additional layer of any cheap wood panel glued and bolted to inside front panel. SPL graphs will stay similar, but the unwanted panel vibrations in the low Hz up to ~5kHz will be much lower. You will hear the difference and can measure difference in energy decay measurements. Internal bracing does not have to be overly complicated or with expensive material, it just has to be there and make its job. Diminishing returns rule here. Many topics on this forum.
5. Panel damping with rubber like substances. You already done good job with silicone (?). Do not search for miracle substances, anything will do. The good option is self adhesive butyl rubber sheets for car audio - 5-10 USD worth of them is more than enough for couple of cabinets. Glue inside cabinet 5% of the of front panel area and call it a day, the leftovers can be glued to metal of inside your car trunk somewhere near wheel arches. This is less important than bracing of your cabinets, rule of very diminishing returns. This influences low-mid Hz range, not ~6kHz
6. As planet10 mentioned, you may have too big cabinet. Bracing will take at least 1 Liter already. You can reduce internal volume by placing inside some wood slab or similar thing, to see if you can hear the difference. Just not obstruct vent. Be generous - will be hard to hear the difference with anything less than 3-5 Liters of reduced volume in your case. From my experience I never liked the sound of the exact software recommended boxes, but they still were pretty close to it. It may be also the matter of personal taste and influenced by the things which software cannot calculate.
7. How much filling you are using? In your cabinet I would start from 400g, that is 1 pound and move to heavier side. The right amount can make miracles. It is important to have some behind the driver also making sure it is not touching spider or membrane.
8. Speaker driver related 5.8kHz dip. As already mentioned yourself - the whizzer cone may be and probably is the cause of it because the dip is sharp and narrow, wavelength also shows, that the source of the problem is in 6cm range from driver center. It may be additionally influenced by the bad luck of having the multiples of that exact wavelength to the sides of the cabinet, but I doubt that. Reflection and diffraction related dips/peaks are wider and more in ~5dB range, not ~15dB as you have.
9. Angle of listening. Any listening angle, which sounds good is normal. Some manufacturers even make drivers to sound good not on 0 degrees. SEAS has some examples if my memory is right. This is often the case for coaxial and FR drivers. 30 degrees is nothing out of extraordinary. Your measurements on that angle looks better to me and the dip of 5.8kHz is not there anymore. Is it correct?
That are my thoughts what I would do in your place.
You already done a very good job with your cabinets, with little additional work you will get much closer to audiophile Nirvana
p.s. Also White Dragon said what I haven't thought about - your initial measurement data can be rubbish. External microphone and PC is the way to measure.
Last edited:
Wow!!!
Thanks again for the very detailed assessments.
Even though I am under verbal restrictions to not build another….I’ll be adding a calibrated microphone to my next order….and utilizing REW….will update.
I’m easy to please (excessive decibels exposure for forty years….lots of headphone and IEM use ….DJ at university with small club large PA exposure). Thus, I don’t hear a problem….but I was drawn to the event horizon of objective speaker evaluation after decades of just building for fun.
Very appreciative of the expertise and time taken to help me. Thank you.
Thanks again for the very detailed assessments.
Even though I am under verbal restrictions to not build another….I’ll be adding a calibrated microphone to my next order….and utilizing REW….will update.
I’m easy to please (excessive decibels exposure for forty years….lots of headphone and IEM use ….DJ at university with small club large PA exposure). Thus, I don’t hear a problem….but I was drawn to the event horizon of objective speaker evaluation after decades of just building for fun.
Very appreciative of the expertise and time taken to help me. Thank you.
It will be a good investment. Expensive, but will pay itself in saved money after several audio projects.
My personal opinion and experience is to look at hearing as to muscle, or brain: with some "training" you can make it better. It works for brains, for muscles, even for eyes - I do not see, why it cannot work for ears. My hearing is good, but with this attitude and some training I have little better vision than some years ago.
SPL measurements are baseline - your ears may have different taste, it is the same as amount of pepper on steak: it will not impact if the food is less or more healthy, but the preferable amount differs for everyone.
Good luck!
A. 22cm
B. 21cm
C. 7.5cm
You are correct, svp; at listening position 30 degrees off horizontal, the 5K dip isn’t there.
Will be interested to see the calibrated microphone + REW results (might be in a couple weeks….still mentally building the next one…it’ll be a hemisphere semi OB (will calibrate REW measure as a comparison)
Last edited:
Baffle step boost from the width is up to ~800Hz.
Distance to the sides and top is also almost exact, so... It may cause bigger peaks and on ~800Hz and lesser peaks on multiples of it - 1600Hz, 3200Hz and some on 400Hz to the ones you already have. They like additional 1-3dB crests to the driver related ones you have - in total 6-7dB and even more. Edge! shows a little different graph, but you get the idea.
Why? Because your driver plays also as a mid and tweeter, and tweeter (and upper-mid) placement is VERY important on the baffle and baffle shape is also. The diffraction occurs in the driver cone itself, as it is a pretty big driver for high frequencies, but some also on the outside, depending on the wavelength. In your case you are multiplying diffraction by additional 50%
With other words: A being the same as B is probably one of the worst things you can do. Play with Edge! by moving the speaker.
Also diameter of 15cm and 21/22cm is very close to exact 2/3 ratio - that is not the best thing also.
In any wide baffle you will get low end boost, which is considered as positive thing, but also diffraction because of it which is IMO far worse than any of the boost you get. In your case also too many very nice dividing dimensions. Yes, you can EQ them, but only on single listening position.
The fix could be as narrow baffle as possible, or... Very wide:
The remedy for smooth high freq is round or stepped baffle and... practice with measurements. VituixCAD is better option to simulate rounded baffle, and it is pretty easy to use if you are designing just for 1 speaker driver.
Btw, baffle is not the cause of ~6kHz peaks or dips in your case and you will not get rid of the diffraction - your task is to get them into more manageable 3-4dB range.
And yet - IMO it is still LESS important than making the cabinet vibration free.
Last edited:
I purchased a DA iMM6 and downloaded then applied the calibration file to REW. Of course, my dumb Windows computer will only recognize the USB microphone as a speaker…thus unable to set the USM calibrated microphone as microphone in REW (can set it as the speaker only)
But, I was able to import the calibration file into House Curve then used the iMM6 as my iPhone’s microphone.
I’ll bring my laptop from work later this week and try the USB microphone with it.
Right then Left speaker on axis at 7mm
And then right left listening position
But, I was able to import the calibration file into House Curve then used the iMM6 as my iPhone’s microphone.
I’ll bring my laptop from work later this week and try the USB microphone with it.
Right then Left speaker on axis at 7mm
And then right left listening position