A Better way to Measure Panel Resonance... Maybe
ฝัง
- เผยแพร่เมื่อ 20 ต.ค. 2024
- If you don't understand what I mean when I say "offset", you need to watch the video I made before this that shows the problem and why there is an offset:
• How Much Noise does yo...
I think this is a lot more accurate than the method I used in that first video. Using the 100hz tone as a control means I can more precisely calculate the offset between the accelerometer and the microphone. But there are other variables that will make a difference, like the distance the mic is from the speaker. So that offset only applies to this test, and will almost certainly be different for another speaker.
Also, the accelerometer is measuring how much the entire box is moving, along with how much the panel it's attached to is vibrating. So that's another variable that will change from speaker to speaker.
One obvious conclusion I reached while doing this is that resonances up higher in frequency will be much easier to hear. Our hearing is also more sensitive in the midrange frequencies, so you'd want to do what you can to try to prevent the panels from vibrating excessively. Using the right thickness plywood or MDF, bracing and damping is all that's needed to drive down the audibility of these panel vibrations. No need to go to extreme lengths or use exotic materials to build the cabinet, when there like won't be audible gains over a normally well built box.
The speaker in this video is made from cheap 3/8" spruce sheathing plywood. It has big unbraced panels and very little damping. So it's probably an example of the worst case.
Still, thousands of speakers have been made in a similar way by big speaker manufacturers over the last 40 years. Most people would enjoy the heck out of the ones I have here and not hear any problems with the way they sound.
Also an important point that I show in the other video is how much the wall beside the speaker vibrates - it was about the same as the panels in the Elac speaker. So unless you are going to do the same thing with the walls, floor and ceiling of the room you listen in as you do with the speakers, there's no getting away from some amount of panel vibration.
You can help support the work I do in making these videos:
Project plans for sale: ibuildit.ca/pl...
Join the ibuildit community on Loacals: ibuildit.local...
Support this channel on Patreon:
www.patreon.co...
#diyspeakers
#johnheisz
#audio
My "Scrap bin" channel:
/ ibuilditscrapbin
My main channel:
/ jpheisz
Website: ibuildit.ca/
Facebook: / i-build-it-25804801424...
Instagram: / i_build_it.ca
If you don't understand what I mean when I say "offset", you need to watch the video I made before this that shows the problem and why there is an offset:
th-cam.com/video/xHY45RmAbDc/w-d-xo.html
I think this is a lot more accurate than the method I used in that first video. Using the 100hz tone as a control means I can more precisely calculate the offset between the accelerometer and the microphone. But there are other variables that will make a difference, like the distance the mic is from the speaker. So that offset only applies to this test, and will almost certainly be different for another speaker.
Also, the accelerometer is measuring how much the entire box is moving, along with how much the panel it's attached to is vibrating. So that's another variable that will change from speaker to speaker.
One obvious conclusion I reached while doing this is that resonances up higher in frequency will be much easier to hear. Our hearing is also more sensitive in the midrange frequencies, so you'd want to do what you can to try to prevent the panels from vibrating excessively. Using the right thickness plywood or MDF, bracing and damping is all that's needed to drive down the audibility of these panel vibrations. No need to go to extreme lengths or use exotic materials to build the cabinet, when there like won't be audible gains over a normally well built box.
The speaker in this video is made from cheap 3/8" spruce sheathing plywood. It has big unbraced panels and very little damping. So it's probably an example of the worst case.
Still, thousands of speakers have been made in a similar way by big speaker manufacturers over the last 40 years. Most people would enjoy the heck out of the ones I have here and not hear any problems with the way they sound.
Also an important point that I show in the other video is how much the wall beside the speaker vibrates - it was about the same as the panels in the Elac speaker. So unless you are going to do the same thing with the walls, floor and ceiling of the room you listen in as you do with the speakers, there's no getting away from some amount of panel vibration.
I get what you are saying about vibrations in the room, but assuming that the room resonances are sufficiently far away from the speakers wouldn't they be much less bothersome than resonances coming directly from the speakers, particularly for higher frequencies? If our auditory system can localize/isolate those room sounds wouldn't it be able to filter them out from the signal proper, thus allowing various illusions to be generated more cleanly? (And wouldn't it be far more difficult to filter/separate out cabinet resonances?) For example, if someone plays an instrument in a room and it causes room vibrations, as long as they aren't extreme they don't significantly affect our judgment of the sound coming from the instrument/player-we get used to the room and for the most part focus on the musician. That is, our auditory system separates out signal from noise, agents from environment, pre-consciously. I imagine that things like soundstage and imaging would be particularly sensitive to such differences-and speaker manufactures certainly to try to sell this idea-but that's just a guess.
Great vid. As a speaker designer I can advise that most commercial designs are checked to ensure that panel resonances are at least -45dB on program to be acceptable by reputable companies at least. Your -28dB is pretty high so demonstrates why panel damping or thicker panels OR additional bracing would be needed. You have to consider that stiffness changes the frequency but not necessarily reduces output depending on panel thickness. The rule of thumb is to make bracing spacing non uniform so that sub panels resinate at different frequencies hence sum of frequencies is way less in total amplitude than equidistant ones.
Cab thickness to brace relationship: doubling thickness of panels is equivalent to adding 5 internal braces in stiffness and amplitude of resonance.
The accelerometer is way more relevant and useful than the panel exciter. It is useful for testing a full frequency response of individual panels but the accelerometer measures the output describing panel behaviour and time of resonance better than a mic.
Keep up the good work!
Interesting problem and approach. I'm hoping you do more along this line.
Hi John. I recommend you do what I have done. Measuring panel resonance is best done when when the drivers are NOT playing. Get an adjustable frequency vibration shelf, and use the shelf vibrating between 20hz and 20k. Then you are ONLY measuring the cabinet itself. My 24" x 18" shelf cost $3900 used however. This way was taught to me by Paul Barton of PSB, about 25 years ago. Works perfectly
The point was to find out how much the panels vibrate while the speaker is playing, and how much sound energy they are putting out.
I can easily find out what frequencies a panel vibrates at by striking it with a rubber mallet and recording the sound that produces, but that's not what I'm looking for.
@@IBuildIt What makes you think that the panel vibration of a speaker contributes anything to speaker output at all? I've never heard of cabinet or panel resonance being an issue with any speaker.
Okay, back in the 70's there was an "invisible" speaker driver. It was essentially the magnetic of the speaker connected to a bolt and that bolt was screwed to any surface - like the back of a table top, door, inside the dry wall of a wall, and most of the time - it was just a gimmick. It sort of work - especially on hollow core doors. I think one of those would be a good direct driver for this sort of testing.
Yep, nut screwed to the panel and screw for additional change.
I’m loving this set of experiments you’ve been doing. I was thinking that using periodic pink noise bursts would be a good way to determine resonances since it would quickly burst all frequencies and give time for decay. I think REW can do that, but I forget.
I'd wonder about trying to give it an impulse instead. Sure, perfect impulses are impossible and only get more so down the signal chain but doesn't that same distortion affect noise bursts as well?
@@hubbsllc I was thinking that a burst would allow it to receive the full bandwidth of sound then have a decay. I’m not aware of what you’re meaning by noise bursts having more distortion, but if that’s true that could be a problem.
The sweep REW puts out does everything that's needed. It excites all frequencies in the range it's set to.
I am struggling to understand Bluetooth speakers and how they compare in audio quality to the old hardwired speakers. Can a Bluetooth speaker attain the same range, and quality of a hard wired speaker system? I would like to have a audiophile quality Bluetooth speaker -- is there such a thing?
There's probably a number of quite good or even excellent Bluetooth speakers on the market. Class D power supply, DSP (for active crossover + eq), DAC, Amplifier, Drivers all in the same box. Does that help?
Do you think it is accurate to describe this as the "relative noise floor" of the speaker? That 47db difference is a pretty big difference, so it probably is not noticeable to most people. It would matter in a critical environment, but people who do critical work already have silly expensive speakers that control resonances much better than your shop speakers.
The 47db difference was between the speaker output and the tone I injected in, but the higher frequency panel resonances were quite a bit higher than that tone.
It's been my contention that when you build a reasonably well built box you won't be able to hear the panel resonances - they'll be too far down and the output from the speaker will swamp it. The question too many have is what constitutes a "well built box".
@@IBuildIt What do you think about an infinite baffle arrangement to test this? Easiest way would be a plywood door in an existing doorway, and cut an opening in the door the size of the speaker cabinet. Then mount the speaker in that opening. With this setup you can get a simultaneous measurement of the sound output from the front compared to sides and back. Both would be the same type of measurement, with no adjustments or offsets to worry about.
It would have to be an extremely well built door to block all of the sound from the front of the speaker, but it is an interesting idea.
@@IBuildIt Just make sure the cabinet resonant frequency is considerably higher than the playable range for the woofer.
@@IBuildIt I would try making the door/ panel from something cheap and easy, like 1/8" plywood glued to both sides of a piece of 1" foam board. Of course you would need to take a reference measurement with the speaker playing in the other room and the panel opening blocked. Some sound pases through the walls regardless, so you need to know how much. With a good reference measurement, you can take the difference between the reference and the measurement of the cabinet when installed in the panel/door.
If panel resonance is an issue, one solution may be to use something very rigid. How about cast iron? But what about going the other way? Build the box out of foam, something squishy, just stiff enough to hold a speaker.
The point is that it is not that big of an issue. But even still you can buy cast aluminum speakers. But I'll leave the cast iron speakers to you!
@@fredygump5578 oh I know, being silly.
@@EF-69 It is hardly silly but, rather, essential to achieving accurate conclusions to have both positive and negative controls to direct measurements of any sort.
A box made of damped panels work better than a box made of heavily braced, stiff panels. Ex. Jean Maurer, sand filled Sandwich panels. A Cast iron 'cabinet' rings like a bell at certain frequencies :)
Lead, concrete, asphalt, whatever.
What if cabinet vibration is a good thing? Isn't that what flat panel transducer speakers are?
I used to own a pair of Techniques 12" 3 way speakers. There was ZERO bracing in them. They were made from cheap, thin, lightweight particle board material. In fact, most modern consumer speakers lack bracing, and are made from the same material, and cm? thickness.
Only higher end speakers, tend to put bracing into them... and even most of those, dont have bracing that looks anything like what you have built here (typically inferior).
If anything is going to De-Form a speaker boxes shape... its going to be from the air pressure and vibrations created by large and powerful woofers being fired, at high volume levels. And that De-formation, could alter all of the drivers soundwaves... as they are attached to that vibrating and distorting cabinet.
Now, Im not an Engineer... so maybe Im getting two things confused. I understand, for example... that if there was a loose speaker wire in the cabinet... that one might hear it rattle. But outside of that, Im not sure I see much reason to even care about minimal midrange "leakage".
Some speakers have the mids and tweeters sealed off from the main air chamber. Either by making the drivers sealed, or by a separate chamber. If these were not separate (as many poor quality speakers do not) the air pressure could effect how they perform... as the bass pressure could hamper their ability to drive without distortions. The amount of interference, could also be a factor, depending on how powerful the bass drivers are... and if they are sealed vs ported.
You don't need to be an engineer, just an audiophile. Some simple speakers sound great, but in hifi we're not only chasing pleasure, but also accuracy. We want the *real thing*. Speakers have dozens of nasty gnomes in them. Each on their own is tiny and maybe inaudible, but if you fix 3, then suddenly you made a perceivable change. For most people this is a waste of time and money, but in reality you don't know until you try. We do it for science and OCD. Curiosity maybe didn't kill a cat, but wasted it's time. :-) The cat had some fun though.
@@rhalfik I became one, on my first Sennheiser audiophile line headphone. Even more so, when hearing... then later purchasing a pair of 1970s era EPI 100v speakers.
I simply meant that I do not understand why he is testing some mild leakage... at low volume levels... in an empty cabinet...
When in fact, it appears that the worst effects of a poor cabinet, would be in its flex and deformation. And that would occur most, when its pumping out high volume level output from the enclosed woofer. It would cause all of the drivers to potentially shift in different directions, as the pressure and vibrations effect them...
And the effect would be greatly multiplied... if the cabinet flexes... as the drivers would theoretically shift in different vectors to each other.. depending on where the flex is.. how much... etc.
It seems very different from him trying to get a reading of a tone from inside of the cabinet.... to outside of it.
Adding to that... Especially with a cabinet that already has a pretty insane level of Bracing. You rarely see a cross brace right behind the woofer. Its probably one of the more optimal places to stiffen a cabinet. However... that said... such a placement is likely to cause more distortions to the driver... due to its close proximity to the projected air-flow / air-pressure. (which is why most cabinet probably dont brace there)
(it might be better, to use a thin metal bar, that connects the two sides. That would allow the high pressure air to easily go above or below it, without crashing into it... which btw... if its being hit heavily.. it could theoretically pull the cabs sides inwards)
Hi, if resonance is the problem, based on the fact the speaker is in a box. You may try to make the box big enough to accept foami made for sound proofing. It has one side flat and the other is a multitude of pyramids which purpose is to disrupt sound vibration. Please let me know if my theory is stupid and eventually if you try it what is the result.
I'm not John, but I think it is indeed silly. :)
Now I'm going to refer to a whitepaper by KEF about researching their LS50 as a successor to BBC LS3/5a, where they explore solutions to the problem. The paper is brief, widely accessible and easy to read, so there is no excuse not to familiarise yourself with it. :) Just search: "kef ls50 white paper".
Aha, the foam you mentioned is used as a choice for aesthetics rather than sound absorption. It's meant to be placed on top of actual absorptive installations to make them look pretty. It's contribution to sound is tiny compared to the forces we're talking about. So the correct methods of damping here are either vibroisolation or CLD, but only assumed a proper bracing in the first place, which thie box in the video is certainly lacking.
@@rhalfik Interesting information thanks, I gona have a deeper look into.
5:54 Am I the only one, who don't hear anything :D
unclear