Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Gareth James]

Can anyone tell me any real figures for a typical transient response (not sure if my wording is right here) or the time for the emitted sound from a basshorn or other bass speaker to decay by for instance 6 or 10dB after receiving a given impulse.

I realise this is quite a complex question to answer properly but if anyone has any general ideas they would be greatly appreciated!

[Brad Litz]

Here is a measurement I took of two LABsub cabinets.

[Michael 'Bink' Knowles]

Brad Litz wrote on Sat, 20 August 2005 15:33

Here is a measurement I took of two LABsub cabinets.

That's really cool, Brad. I've never seen a graph like that -- I don't do impulse testing in my live sound work.

Can you explain what it is we are looking at? For instance, why are there little low-SPL hillocks near 30, 40 and 50 Hz that are stretched out in time? Do those hillocks show a leading sound or a trailing sound? Where is time = zero? Close to the viewer or far from the viewer at the dB plane?

-Bink

[Brad Litz]

The chart is a cumulative spectral decay plot, commonly called a waterfall plot. Time zero is at the "back wall" where the level is the highest. The "low-SPL hillocks" you mentioned are indicating that there is some reduced contribution to the response at those frequencies for a longer period of time.

I took the measurement outdoors as you can see in the photo. Perhaps the rather flimsy garage door baffle contributed some sound of its own.



Here is a link with some explanation.

http://www.libinst.com/wattlar.htm

[Gareth James]

Wow, thanks brad thats exactly the kind of info i was looking for...

So if im reading this right, thats about 40ms average decay time below 100hz, with peaks up to 100ms or more at the resonant frequencies of horn (and drivers?)

I think you're right, maybe one of those peaks is your garage doors...  

still some awesome data.... i assume a vented box would have a greater average decay time, maybe approaching double the 40ms figure...

[Too Tall (Curtis H. List)]

Hey Brad, clever setup.

All I have is a fiberglass door, though it might work after all.

Here is a Praxis waterfall taken from an Impulse Response of a single LAB sub outdoors. This is a version 1 design.
I believe the mic is near field at the mouth of the horn.

The plot is 100ms deep with each line 2ms.
Freq range from 20Hz to 100Hz.

My next post shows the same sub taken at 20feet (about 6 meters).

[Too Tall (Curtis H. List)]

The plot is 100ms deep with each line 2ms.
Freq range from 20Hz to 100Hz.

This shows the same sub taken at 20feet (about 6 meters).

[Too Tall (Curtis H. List)]

[quote title=Bink wrote on Sat, 20 August 2005 18:44]

Brad Litz wrote on Sat, 20 August 2005 15:33

Here is a measurement I took of two LABsub cabinets. That's really cool, Brad. I've never seen a graph like that -- I don't do impulse testing in my live sound work. Can you explain what it is we are looking at? For instance, why are there little low-SPL hillocks near 30, 40 and 50 Hz that are stretched out in time? Do those hillocks show a leading sound or a trailing sound? Where is time = zero? Close to the viewer or far from the viewer at the dB plane? -Bink

BTW Bink, if you want to do Waterfalls (called CSD in Praxis, Cumulative Spectral Decay) or do FFT with gates, ETC, Group delay, add/subtract time files, multiply/divide freq. response and about a dozen other things all you need to do is start saving your IR from Smaart as ASCII files and import them into the Praxis demo.

If you have any questions about settings let me know.

[Brad Litz]

Too Tall wrote on Sun, 21 August 2005 16:56

All I have is a fiberglass door, though it might work after all.

I think the the fiberglass door may be better than the sheet metal one I have. The roll up design of mine keeps the panel size small and fairly stiff.

Since the LABsub horn design is intended for a group of four on an open ground plane, two "in the garage door" should be close to the same.

[Peter Morris]

The LAB horn is about 10 feet long, so the resonance associated with its length will be at about 30Hz.  Thats what you are seeing, resonance, the cone or anything else for that matter including the box, keeps vibrating after the signal has stopped.  What you see as time goes on (the Z axis in this case) is an ever decreasing amount of energy, mainly at 30 Hz and some at 60Hz, which is exactly what you would expect in this case.  There is some at 40, this maybe the box, the wall, don’t know.

As you put more of the LABs in a stack and the mouth becomes bigger it will begin to behave more like a horn and not a resonant length of pipe at these low frequencies.

Having said that – that’s a very good plot in any case, but they are loaded in 1 / 4 space.

Peter

[Too Tall (Curtis H. List)]

Peter wrote on Sun, 21 August 2005 20:01

The LAB horn is about 10 feet long, so the resonance associated with its length will be at about 30Hz.  Thats what you are seeing, resonance, the cone or anything else for that matter including the box, keeps vibrating after the signal has stopped.  What you see as time goes on (the Z axis in this case) is an ever decreasing amount of energy, mainly at 30 Hz and some at 60Hz, which is exactly what you would expect in this case.  There is some at 40, this maybe the box, the wall, don’t know. As you put more of the LABs in a stack and the mouth becomes bigger it will begin to behave more like a horn and not a resonant length of pipe at these low frequencies. Having said that – that’s a very good plot in any case, but they are loaded in 1 / 4 space. Peter

Ooops!

The descriptions of the setup on those graphs are incorrect.

That particular batch of data was a failed attempt to measure the change in freq response and impedance as you make bigger stacks. The project was cursed from the start and right through to the end.

It started out with a hum when I hooked the laptop up to the power amp rack. From there it went to an argument with my help on stacking. I wanted to measure a box by itself. My help insisted that it was good enough to stack all four subs in a block at the beginning and just hookup first one, then two, etc.

We know this is wrong, but the help would not be convinced.

The final nail in the coffin was that the NL4 input panel on the back of the amp rack had bad solder connections so I don’t know just how many speakers were actually working at any one time.

So that graph WAS just one LAB working, but it was part of a stack of four where the other three were not driven.

I found another set of data that I believe was from the very first LAB that was built here. It is a ground plane at 1 meter outdoors.

Sorry about the confusion. As you can see the graphs are fairly similar.

[Michael_Elliston¶]

Gareth James wrote on Wed, 17 August 2005 05:33

Can anyone tell me any real figures for a typical transient response (not sure if my wording is right here) or the time for the emitted sound from a basshorn or other bass speaker to decay by for instance 6 or 10dB after receiving a given impulse. I realise this is quite a complex question to answer properly but if anyone has any general ideas they would be greatly appreciated!

http://homepages.paradise.net.nz/sbk1/imp.htm

[Peter Morris]

There is one point to note.  When you use one LAB its mouth is not big enough to produce 30 Hz properly and it behaves a bit like a resonant 10 ft pipe. What you see on Brads plots is the resonance decay at about 30Hz and the next harmonic 60 Hz associated with its length.

I have not measured it but when you stack about 6 LABs together in half space the mouth is big enough and I would expect things to improve.

Peter