panomaniac said:
It can only sound different if you can hear it above the excitation.
To understand this problem it is essential to understand the concept of radiation efficiency. Take a dipole versus a monopole with identical cone excursion. The differences in output are enormous. Amplitude alone is a very poor indicator of sound radiation.
gedlee said:
The paper contradicts your 1st point. If we extend the "dowel" cross-section to the point where is extends across the whole cabinet then the panel is stiffer yet. If we extend it in such a way that the aspect rations of the 2 subpanels increase then it gets stiffer than if the aspect ratio shrinks.
Further unequal subpanels is better than equal subpanels (not necessarily in stiffness, but in terms of distribution of enery required to excite panel resonances).
I have been working very hard towards an elegant solution to box bracing that is simple and effective. My latest creations are getting pretty good (matter of fact, you came to my aid in one thread where that structure was being discussed)
dave
panomaniac said:
Agreed on most all points. I used the iPhone as a comparison to a loudspeaker (in an enclosure) moving with similar energy and noted in a later reply that the folly in the comparison includes the isolation of the backwave from the listener, and the fact that a driver is generally intended to act as a piston.
thadman said:
This is a much better articulation of what I was trying to convey when I used "sound at the listener" or some variation thereof.
panomaniac said:Exactly my point!It's rather hard to figure out just how much the walls are radiating. For the reasons posted above, as well as others.
But radiate they do. As we can see in the graphs that Rudolph linked to, bracing makes a difference for sure. I did a little overlay of the graphs of those 2 plots just for my own amusement. Will post later, if Dave doesn't beat me to it. Although the resonance point went up with bracing, the amplitude is lower and it decays much faster. That ought to sound different. Maybe less noticable.
There are going to be a number of factors at play in determining what someone in a reasonable listening position (ear not against a cabinet) would hear.
I'm inclined to believe that if a reasonable amount of effort is put into appropriately eliminating panel resonance via typical construction techniques (bracing, cld, etc.), the incremental gain from going further will be overshadowed by the resonances of all the other surfaces aside from the cabinets that are present in any given room.
I know I can get kitchen cabinet doors to start resonating in the next room at higher volumes. The objects in the room with the loudspeakers have to be doing something and it would seem to me that the summation of those items is more likely to be an issue than the relatively minor cabinet vibrations that remain after good cabinet construction.
Now I'm off to go perform a "laying on of hands" on the various objects in my listening space.
I've posted this before, and I know, without comparison to others, or measurements of the actual sound radiated from the box, it's of little help in such a discussion, but it was informative to me. The above image is from an accelerometer reading on the side of a 5 cubic foot ported subwoofer box. The walls are roughly 2" thick CLD, with 3" thick CLD front baffle. Unfortunately the builder didn't listen to my instructions and used nails to hold the entire thing together while the adhesives set, so each panel isn't necessarily free to move like you would want. None the less, I asked for a flexible adhesive be used to laminate MDF together. The inside has interlocked window pane braces. The walls for the left and right side also have panels attached with numerous 4" and 6" holes cut into it. Don't know what to call it, I stole the idea from Focal. Finally, once I received the finished box, I lined all interior surfaces with a spray on dampening compound, followed up by a multi-layer dampening foam, like black hole, totaling 2"s thick. The box was filled with wool batting, and the port lined with the same foam Dr. Geddes uses for his horns. This was my attempt at an all out assault on minimizing vibrations.
I think it's worth noting that, it's just not that dead. My windowing and range makes it look worse than it really is, but frankly, I expected to measure next to nothing, and that's not what I got. Also keep in mind that, this box, raw materials alone, was almost 300 dollars.
This is my build of Dr. Geddes box for the Abbey, measured in the same way, just different software. Worth noting here is that the accelerometer can not be calibrated to eliminate it's own response issues, such as that peak at 20khz, nor does it have a good noise floor. With it attached to nothing, and no sound, the noise floor is, though uneven, around 50-60db's. Also keep in mind, these are not calibrated accelerometer measurements. I have no way of isolating the output of the panels from that of the box to calibrate the device, so I can only use it in relative terms. The output of the box is not 80db's, I have no idea what it really is. Also note, the output voltage was 5.6 volts for the Abbey, and 15.2 for the subwoofer (into an 8 ohm load).
Probably most important, how does any of this relate to how the speakers sounded. What do those measurements, which seem to be pretty industry standard for manufacturers, correlate to some impact on sound. I have no way of of interpreting these. I took them, sat down, studied them, and still am left with no good way to interpret them. With the Abbey box, I can say there are no resonant modes apparent. The vibrations from the box hit the noise floor around 1khz or possibly lower. Adding additional bracing and dampening to the Abbey had almost no effect, as can be seen when comparing the blue and red lines, blue being with added dampening and bracing.
I see two issues though. First, let's say that it goes from being -80db's to -95 db's, does this really matter. it's already so low in level, I can't imagine the change matters.
The second problem is that it's not really decibels. Maybe someone more knowledgeable than I can better discuss what it does or does not mean. We have a device designed to convert vibrations picked up on solid surfaces and convert them to voltages. It isn't a microphone, though I'm sure it's fundamentally similar. However, I believe it's a piezo type device and is optimized for lower frequency vibrations. While the program, which is designed to convert voltage to decibels, decided that it was a 6-14 db difference, it's not really db's, and so I don't really know what it means. I might be all wet here, it may be that this really is basically just a microphone, and that, if I could set a reference level for it, that is accurately showing a reduction of 6-14 db's. What the software developer told me was that, it's not an accurate corrected device, you can't take measurements with it and use the db numbers to mean anything as compared with sound output. It's sole purpose is for relative comparisons. If you see a peak, there is a resonance, if not, there isn't one. If you add bracing and dampening and things change, then you have imparted change relative to before, but no comparison to other methods can accurately be made.
The software designer also told me not to use data below 100hz with this device, as it's either unimportant or unreliable (I'm not sure what he meant, he is German, and his English is only ok).
Then there is that issue that has been harped on over and over, does a 10db reduction between 80 and 120 hz have any impact on sound. If I saw this in the midrange to lower treble, I would be far more likely to agree it might matter, but in this upper bass range, I'm not so convinced. The room becomes so dominant at those frequencies, that I'm just not so sure it matters.
For those interested, the thing that created that difference was spraying the entire inside with dampening compound, then using a flexible adhesive to adhere dampening material with a metal foil later. Followed by 2" thick multi layer black hole 5 like material. Total cost for materials was in excess of 150 dollars for all three speakers. The added bracing was Batons, but I took measurements with just them and they were almost 100% identical to without.
The second problem is that it's not really decibels. Maybe someone more knowledgeable than I can better discuss what it does or does not mean. We have a device designed to convert vibrations picked up on solid surfaces and convert them to voltages. It isn't a microphone, though I'm sure it's fundamentally similar. However, I believe it's a piezo type device and is optimized for lower frequency vibrations. While the program, which is designed to convert voltage to decibels, decided that it was a 6-14 db difference, it's not really db's, and so I don't really know what it means. I might be all wet here, it may be that this really is basically just a microphone, and that, if I could set a reference level for it, that is accurately showing a reduction of 6-14 db's. What the software developer told me was that, it's not an accurate corrected device, you can't take measurements with it and use the db numbers to mean anything as compared with sound output. It's sole purpose is for relative comparisons. If you see a peak, there is a resonance, if not, there isn't one. If you add bracing and dampening and things change, then you have imparted change relative to before, but no comparison to other methods can accurately be made.
The software designer also told me not to use data below 100hz with this device, as it's either unimportant or unreliable (I'm not sure what he meant, he is German, and his English is only ok).
Then there is that issue that has been harped on over and over, does a 10db reduction between 80 and 120 hz have any impact on sound. If I saw this in the midrange to lower treble, I would be far more likely to agree it might matter, but in this upper bass range, I'm not so convinced. The room becomes so dominant at those frequencies, that I'm just not so sure it matters.
For those interested, the thing that created that difference was spraying the entire inside with dampening compound, then using a flexible adhesive to adhere dampening material with a metal foil later. Followed by 2" thick multi layer black hole 5 like material. Total cost for materials was in excess of 150 dollars for all three speakers. The added bracing was Batons, but I took measurements with just them and they were almost 100% identical to without.
pjpoes said:
You wouldn't have a bigger version would you -- and one with the time axis in periods.
It looks like you have been pretty successful in moving the resonances high enuff that a filtered sub will not excite them.
Some of the described treatment (ie spray on damping) will counter the positive benefits of some of the other stuff you did.
dave
planet10 said:
Hi dave,
You are quite confident regarding Your attitude against more relaxed enclosure building. What would be the use in rearrangeing a graph related to the topic? You should already know for long what is going on.
If You are going for an orientation, that You want to gain from others experiments, would it be more promising to let all prejudices behind for a while? All You are talking about is "could" and "should". Anything depends on everything. Some kind of alchemy, isn't it? In the end, what counts is whether or not You've made gold from - You know ... straw. What counts is not only the color but weight also, ductability etc. I can' stand to await Your very own measuremnts on that topic any longer. We've seen so much of comprehensive results that count against Your attitude, You have to have found some really great stuff and methodology, if You get even better yield. I must see Your concept of maximizing the effort working. Let the soundfield speak now! What would (not) be heard with Your recipes applied?
Thank You very much, Tausend Dank!
Matt
Because of the way an accelerometer works it is basically like a microphone (but for acceleration not pressure) and dB in the software is valid. There is a lot of uncertainty at LFs because the acceleration is getting very small. In other words the transducers noise floor is rising at the low end.
The 6-14 dB would be significant, IN acceleration, IF you had spatially averaged data, but as with any measurement the one measurement point is going to be quite different from another and the only really valid measure would be an average.
It is completely unknown if any of this reduction is audible or not. The lower the frequency of the reduction the less audible I would expect it to be because the ears sensitivity drops significantly at LFs.
Because of the way an accelerometer works it is basically like a microphone (but for acceleration not pressure) and dB in the software is valid. There is a lot of uncertainty at LFs because the acceleration is getting very small. In other words the transducers noise floor is rising at the low end.
The 6-14 dB would be significant, IN acceleration, IF you had spatially averaged data, but as with any measurement the one measurement point is going to be quite different from another and the only really valid measure would be an average.
It is completely unknown if any of this reduction is audible or not. The lower the frequency of the reduction the less audible I would expect it to be because the ears sensitivity drops significantly at LFs.
xpert said:
All you ever learn as you go along is how little you know. More data is always helpful.
Rearranging the graph to periods allows direct comparison of the audiability and severity of a resonance, With a scale in msec a resonance of less severity than one at HF may still appear worse if one doesn't mentally adjust.
It saves from a lot of misinterpretation.
dave
planet10 said:
Wrong. It's accelerometer data. Not soundfield data. You always mix that up! Servere mistake. If not to say crucial.
planet10 said:
Exacting! There's a lot of misinterpration around. You told us above about acoustically transparent enclosures on resonance. What does that mean to the regular guy? If my ears were eyes I could look through? Does it harm? Honestly, I don't know what could be meant with Your note. Does it affect the soundfield? How much which way?
Thank You
xpert
Hey, lighten up! You make mistakes and have not offered up much data on anything that I can recall. You criticise a lot but offer up very little.
Dave was correct about "acoustic transparency". There exists a situation in a panel whereby the sound field on one side of the panel can go through the panel unattanuated to the other side - and it is said to be "transparent" - just like in optics. The effect is called coincidence and "transparent" is a valid word to use in this case. When applied to panels of finite extent and with fixed edge sthe effect is far less pronounced, but similar effects do occur.
So a little more "understanding" and less criticism would be appreciated by all.
Hey, lighten up! You make mistakes and have not offered up much data on anything that I can recall. You criticise a lot but offer up very little.
Dave was correct about "acoustic transparency". There exists a situation in a panel whereby the sound field on one side of the panel can go through the panel unattanuated to the other side - and it is said to be "transparent" - just like in optics. The effect is called coincidence and "transparent" is a valid word to use in this case. When applied to panels of finite extent and with fixed edge sthe effect is far less pronounced, but similar effects do occur.
So a little more "understanding" and less criticism would be appreciated by all.
xpert said:
It is data.
The audibility refers back to the earlier paper linked too. If the time scale is in periods you can directly compare the Q of the resonance. A higher Q resonance means that there is likely to be sufficient energy in a musical signal to excite the resonance (something i knew), and, as i learned from the paper linked above, the higher the Q then the higher the audiability threshold (ie higher Q is less likely to be detected).
dave
Here it is larger, but alas, I don't have it in periods. I once again have had a catastrophic computer failure, and have no way of getting to that data. Not such a big deal for my audio hobby, kind of a big deal for my thesis, for which the most updated copy was on that laptop. Who would have thought that RAID drives would be useful for word documents, but here is a good example. While typing some updates into my lit review section, the hard drive failed as a result of a part failure on the laptop motherboard causing catastrophic and permanent damage. While I update my backup regularly, I don't update my backup while in the middle of typing. I type quickly, and as a result, had probably changed close to 500 words or more since I last backed it up. At the same time that this happened the IRB emailed to let me know I needed to sit on my project until they had time to re review the project and approve a larger sample we had collected, putting my work on ice for three days. It was just a great day all around.
Anyway, once I get my new laptop and get all the software installed, I will try and take measurements as suggested.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.