Tower speakers on hardwood floors: spikes on discs/coins? Sorbothane/silicone pads? Springs?

[MAB]

Thanks. If you still have the files (spike and elastomer) I would be glad if you could share the files as well. As I’ve stated several times, there is usually nothing to find, in most cases. The overtones that radiates from floor and other objects are usually masked by the speaker distortion. In the example with the130 Hz fundametal the distortion is found in ≈ 5th-15th harmonics.

I forgot I closed REW and didn't save the Seas woofer data.

I took new measurements. This time with a 15" woofer, in a sealed Baltic birch cabinet, again in my living room with old oak floors. The 15" is slightly lighter than the Seas sub's overly heavy cabinet. I placed the woofer in the middle of a long wall. The room has several strong modes, and is not acoustically treated, it's a living room. I set the speaker on three elastomeric feet and compared to three spikes. The speaker's height off the floor is 2.5mm for both elastic feet and spikes. I swapped the feet for the spikes while being careful to keep the same position relative to the mic, much better than 1cm by my estimation. The mic and all settings stayed the same between runs.

This is very close to run to run variability. This is no change, as closely as I can control the position of the speaker. More on that in a minute.
Comparison of distortion is below:

There is no difference in distortion.
Here are the first few HD:

There is no difference in distortion between elastomeric feet and spikes.

The data is attached in the zip file below.

Regarding run to run variability. I took care to get the speaker in the same position so that even small changes in position wouldn't influence the results. The differences between spikes and elastic feet are smaller than small environmental changes, like the room warming up by a couple degrees. To simulate changes in room temperature, I put a space heater set on low 2 meters away from the speaker for 5 minutes. I compared the frequency response and distortion measured just before the heater and after 5 minutes of gentle heating.

The heater caused a larger change than swapping feet. Playing the speaker warms the driver too.

After I did the heater measurement, my wife came back home. Here is a screen grab of her footsteps on the hardwood floors as she walked in the other side of the house (and moments later laughed at me for being a supersonic nerd in the family room ).

If the floor were somehow vibrating significantly differently from the speaker, the mic would pick up that reradiated sound.

 

[Thomas_A]

Thanks for this. To me it seens there are some differences at e.g. 40 Hz and in the 100-200 Hz region. You had some 40 Hz difference also in the previous sub measurement which speaks for repeatability. While small they seem to be measurable and as I mentioned before, they may or may not be audible in some cases. I might look at it closer if the files are there. Traveling to London tomorrow so not sure when though.

 

[Purité Audio]

Years ago a laboratory isolation manufacturer measured here with a four plane accelerometer, with and without using their active isolation platforms ( not foo laboratory isolation for nuclear force microscopes ) they concluded there just wasn’t enough energy being transmitted to the floor to cause audible resonance.
And of course if there was structurally borne resonance you would still have to contend with the airborne.
It was interesting I did buy one of their platforms for under my turntable almost impossible to decide whether that made any difference to sq just because it was so difficult to compare once the platform had been fitted.
Keith

 

[Chromatischism]

We should mention the construction of the room. I'm in a 2nd floor room of a house made of wood. Your results will vary if you are on concrete, tile, etc.

 

[Purité Audio]

It is really down to the construction of the speaker, if you buy decent speakers there will not be an issue.
‘Accessories’ are targeted at audiophiles not too expensive and purported to bring an improvement.
Acoustic measurement with REW is free and will/may lead to tangible improvements.
Keith

 

[Esprit]

When I had passive speakers I used these (cheap) but I never measured...

index

 

[Thomas_A]

I forgot I closed REW and didn't save the Seas woofer data.

I took new measurements. This time with a 15" woofer, in a sealed Baltic birch cabinet, again in my living room with old oak floors. The 15" is slightly lighter than the Seas sub's overly heavy cabinet. I placed the woofer in the middle of a long wall. The room has several strong modes, and is not acoustically treated, it's a living room. I set the speaker on three elastomeric feet and compared to three spikes. The speaker's height off the floor is 2.5mm for both elastic feet and spikes. I swapped the feet for the spikes while being careful to keep the same position relative to the mic, much better than 1cm by my estimation. The mic and all settings stayed the same between runs.

View attachment 395781

This is very close to run to run variability. This is no change, as closely as I can control the position of the speaker. More on that in a minute.
Comparison of distortion is below:
View attachment 395791
There is no difference in distortion.
Here are the first few HD:
View attachment 395793
There is no difference in distortion between elastomeric feet and spikes.

The data is attached in the zip file below.

Regarding run to run variability. I took care to get the speaker in the same position so that even small changes in position wouldn't influence the results. The differences between spikes and elastic feet are smaller than small environmental changes, like the room warming up by a couple degrees. To simulate changes in room temperature, I put a space heater set on low 2 meters away from the speaker for 5 minutes. I compared the frequency response and distortion measured just before the heater and after 5 minutes of gentle heating.
View attachment 395796
The heater caused a larger change than swapping feet. Playing the speaker warms the driver too.

After I did the heater measurement, my wife came back home. Here is a screen grab of her footsteps on the hardwood floors as she walked in the other side of the house (and moments later laughed at me for being a supersonic nerd in the family room ).
View attachment 395804
If the floor were somehow vibrating significantly differently from the speaker, the mic would pick up that reradiated sound.

Could not load the file in my REW version... I might have an too old one? v5.20.13.

 

[Thomas_A]

Fixed it with an update. But there is no way to get the sweep out of REW? Only exporting impulse response?

 

[bothu]

My speaker stands are very practical. I can move my speakers to any new position in seconds. Extremely convenient when cleaning the floor.
Drawbacks in sound quality, no.
Try this and be free .

 

[MAB]

Thanks for this. To me it seens there are some differences at e.g. 40 Hz and in the 100-200 Hz region.

Yes, there are differences, and they are tiny.

The 56 Hz difference is also tiny.
Same for 100-200Hz, tiny:

Same for distortion:

These are absolutely tiny differences. Slightly larger than run to run variation.

The change between feet is smaller than gentle application of heat:

Also a tiny difference, but larger than the impact of the feet, and affects much broader frequency distribution. And I barely heated the driver.

You had some 40 Hz difference also in the previous sub measurement which speaks for repeatability. While small they seem to be measurable and as I mentioned before, they may or may not be audible in some cases. I might look at it closer if the files are there. Traveling to London tomorrow so not sure when though.

Fixed it with an update. But there is no way to get the sweep out of REW? Only exporting impulse response?

The entire measurement file is there. I was able to open it and see all of the data.
I attached the datafiles again, this time individual sets for each run. The data is the same.

I don't think this level of discussion or microscopic evaluation helps. People genuinely wonder what they are missing in spikes, elastomers, and other advanced approaches to vibration isolation. The industry is filled with some really bogus claims about speaker resonances, and sell some pretty premium products that ridiculous. There just isn't enough vibration to transfer as much energy as people think into the floor. Picking up a speaker by the cabinet while operating at high volume will give a hint as to how little vibrational energy there is. The data I posted shows the reality, the differences are so small that stability, adjustability, and flexibility are by far the most important choices. For instance, some people with carpets have better stability with spikes than any kind of feet, so spikes it is. Some people struggle with elaborate spike setups on hard floors, thinking they hear differences, I just showed the differences are too tiny to worry about, and not better or worse.

 

[Thomas_A]

Yes, there are differences, and they are tiny.
View attachment 395869
The 56 Hz difference is also tiny.
Same for 100-200Hz, tiny:
View attachment 395894
Same for distortion:
View attachment 395873
These are absolutely tiny differences. Slightly larger than run to run variation.

The change between feet is smaller than gentle application of heat:
View attachment 395874
Also a tiny difference, but larger than the impact of the feet, and affects much broader frequency distribution. And I barely heated the driver.



The entire measurement file is there. I was able to open it and see all of the data.
I attached the datafiles again, this time individual sets for each run. The data is the same.

I don't think this level of discussion or microscopic evaluation helps. People genuinely wonder what they are missing in spikes, elastomers, and other advanced approaches to vibration isolation. The industry is filled with some really bogus claims about speaker resonances, and sell some pretty premium products that ridiculous. There just isn't enough vibration to transfer as much energy as people think into the floor. Picking up a speaker by the cabinet while operating at high volume will give a hint as to how little vibrational energy there is. The data I posted shows the reality, the differences are so small that stability, adjustability, and flexibility are by far the most important choices. For instance, some people with carpets have better stability with spikes than any kind of feet, so spikes it is. Some people struggle with elaborate spike setups on hard floors, thinking they hear differences, I just showed the differences are too tiny to worry about, and not better or worse.

The reality is that people are using various supports for their speakers, flloors of different kind, desks and bookshelves. In worst cases vibration could be audible as distortion as I have shown and others did 40 years ago when spikes arived. In this regard coupling and spikes are the worst. Any soft foot that keeps the resonant frequency below the lowest frequency of the speaker output is better. It will ensure that this kind of problem will not occur irrespectively if there is an issue to start with or not.

(Regarding the file you sent, yes I can investigate in REW, but would like to export the sweep in wav. Just to look at harmonics up to 20 or so. It might just be me but I just get the impulse response exported from REW. )