Just an observation about voltage swing in actively crossed-over tweeters.

[Phloodpants]

I will stick it on the scope today.

Drivers are:
Tweeter - Hiquophon OW-1 - 87db
Mid - Focal 5K4211 - 88.6db
Woof - Focal 10K6411 - 92.1db
Sub - TC Sounds 15" - Not sure of sensitivity

System is balanced with an approximately 7db "tilt" from bass to treble.

 

[Phloodpants]

Here's the same track with 300Hz and 3K 4th order crossover points.

As you can see, raising the crossover point statistically reduces the number of full-scale peaks, but does not eliminate them.

Once we get to a 10KHz crossover point, the full-scale peaks are finally eliminated.

At 300 and 3K crossover points, the peaks are only occasional, and if you had an amp that could not swing that far, it's unlikely you would actually hear the clipping. But I don't want clipping.

If you're running a super-tweeter with your full-rangers at a high crossover point, then a small amp will do...

 

[sreten]

Hi,

I guess you can appreciate the bandwidth of a full scale peak, and its
fundamental varies statistically. Why I'm a big fan of doing active as
equal power amps for bass and mid/treble with an active x/o bass
to mid, and a passive x/o mid to treble with active EQ for any BSC.

Typical recordings can't use the dynamic possibilities of 300Hz/3KHz
3-way fully active, as its so alien to having the simple single amplifier.

rgds, sreten.

 

[Phloodpants]

OK, watching the scope at a moderate volume. (It's all relative)

I don't have a peak-hold scope, so I am running at 0.5sec/div and watching the phosphors glow.

Observed peaks: Lunch w/Gina
Sub - 2.5v (Summed L+R so this should really be 1.7v)
Bass - 3.7v
Mid - 3.8v
Tweet - 3.8v

Pretty much as predicted by the waveforms. My tweet and mid have similar sensitivity.

 

[Bigun]

One of the tidbits people have posted about which may be a source of confusion is the analysis of power content of typical music - I think Rod Elliot says somewhere on his website (IIRC) that half the power is below 300Hz and half above it - for typical music.

 

[Phloodpants]

Here's an extreme-bass example. Dieter Moebius - Womit.

Top is the bass low-passed 4th order @ 70Hz.
Bottom is midrange, bandpassed 4th order 300-3K

Here we have bass peaks as high as the mids, and obviously more RMS power, but we still need an amp with the same swing as the sub!

 

[Phloodpants]

I should note, regarding the above peak observations on the subwoofer. I use a 2db boost at 16Hz combined with a 2nd-order highpass at the same frequency. (Q of the filter is higher than 0.707.)

And my sub is in a 14cuFt enclosure, so is more efficient than most. Of course, a sub in a tiny sealed box with an extreme EQ is going to need more power.

 

[sreten]

Hi,

Which all goes to prove that going active seems pretty pointless
with common recordings that routinely hit the end stops of level.

Only with much higher dynamic range purist recordings does
the dynamic aspect of going active really affect max levels.

rgds, sreten.

 

[Phloodpants]

Interesting point, but with bass playing out the same amplifier, the rails can sag and cause clipping. Unless you have an amp with a regulated power supply.

 

[Phloodpants]

But as you mention, splitting the bass off and doing mid/tweeter passive, that would effectively keep the amp rails high for the mid/tweet amp.

 

[sreten]

But as you mention, splitting the bass off and doing mid/tweeter passive,
that would effectively keep the amp rails high for the mid/tweet amp.

Hi,

Yes. I'm pretty sure its the best way of going active, for a 3-way.

rgds, sreten.

 

[infinia]

we know the average power levels taper off with frequency but we don't know that the instantaneous peaks vs frequency do. so that only gives clues towards heat sinking and PS hold up times.
sooo other than that, I reckon you can only size each amp by the relative driver impedances and sensitivities.

 

[Phloodpants]

Another factor occurs to me... Pop music producers typically use side chain filtering to apply more compression to the bass than the treble. It's the key to a tight drum-kit sound and a steady, even bass level in the mix. But you don't want your treble sounds pumping up and down based with the bass levels!

Also, a lot of stereo imaging information exists at high frequencies, at low amplitude, and would be scrambled in amplitude if it were intermodulated with the bass compression.

 

[cbdb]

Simple explanation. Bass is slow, peaks are fast.

 

[cbdb]

I think you mean multi band compression. And it's usually used on one instrument at a time , not the whole mix.

 

[cbdb]

Side chain uses a different signal to cause compression. Like using the kick drum track to compress the bass track so the kick pops more. Or taking a vocal track Ewing the siblance louder and then feeding that into the side chain of the same vocal to reduce the siblance. You get the picture. Multiband splits the signal like a crossover and let's you apply different compression to each band.

 

[Phloodpants]

I hate to be pedantic, but I'm pretty sure multi-band compression is a type of side-chain compression. The different signal causing compression is an equalized copy of the original. User "Rif" on this discussion thread seems to have a good explanation...

Without a sidechain filter, the compressor's detector (the thing that determines how many compression should be applied) is controlled by the complete frequency range of the signal. A sidechain filter can manipulate the signal that's going to the detector, thus the compressor reacts only to the filtered signal. The processed (compressed) audio is not affected by the side chain filtering at all.