Through the magic of Google Image, I found the image here. One can click Back and Next to see other pages (it's a report on a computer program that detects music genre), but I still see no indication of what the frequency scale really is. This info had to exist to make the images and write the report, but somehow it didn't make it in. I can only suspect the frequency scale is arbitrary units, and perhaps it's a log scale:
OpenStax CNX
And like Ken you seem to have failed to read the linked paper. This is ALL covered there, in an easy to digest form.
However if you have a classic 3-way, with (say) 95dB/W drivers and you have a lossless crossover a single tone at 95dB@1m at 100Hz will need 1W from the amplifier. Two tones, both of 95dB@1m one at 100Hz and one at 1KHz will need 4W. As the (unread) article states 3dB gain is the theoretical maximum and 1-2dB is more usual, but once you add 3dB of BSC active looks a lot more interesting, esp if you love low power single ended triodes.
This should be something to be discussed, rather than flamed.
Help me out on the maths, what if I have 1000's of tones say a white noise, or music, instead of just two!
Do I need 100000's of watts to reproduce 95dB? does that sound right?
Please do not refer me back to your bible 'the article'.
Eventhough through the magic of mathematics (post #159) it was shown my argument was correct, somehow our friend bill has chosen to remain silent about it.
Thank you but there was no need, I had supplied the link in post #135 along with two more graphs and links to their origins.benb said:
I suppose all three were misused/misquoted/irrelevant.
BTW it is not logarithmic, it is linear and the frequency scale is obvious to anyone who has any experience with audio graphs.
At any rate there are two more graphs and links if you are not convinced.
You were not 'proved right' based on your infamous 'wrong' post. It was just shown (as the first table in the elliot article also shows) that there is more power in lower frequencies. This was never in disagreement. As for your question I still can't work out if you are trolling for lolz or actually interested in the discussion. I fear the former based on you hanging onto to one point like a terrier to a stick.
For the record, I do not view the biamping article as a bible, but I repeat that IF YOU BOTHERED TO ACTUALLY READ AND UNDERSTAND IT you wouldn't be the major noise source on this thread.
You have been as noisy as me, NO?
My question was a genuine one, but it seems to me, you have accepted the article in question as gospel and you keep referring me back to it (as you just did again).
I do not wish to read that article as the first few chapters I read were flawed, and I would like to move from it.
It seems you are claiming certain 'facts' that do not concur in real life. like your example that for 95dB on one tone you need 1W for two you need 4W - it follows that for 1000's of tones you should need a million watts! but in real life a 50W amp can reach 100dB or more with complex musical passages consisting of 100's of (if not 1000's) of tones, just shove 5 seconds of Beethoven 5th into a an FFT mincing machine, see what you get.
Your arguments don't make even common sense.
The basis of the whole discussion arose from a statement I made, that an active system with two 100W amps is not going to be as loud or powerful as a 400W passive system, when it comes to reproduction of music ( tones are for lab), anyone who has ever had an active system would witness to that.
Active speakers are not designed to cheat the loudness game, they are designed to be accurate.
To prove my assertion, I stated that the majority of the power needed to create music is in sub 200Hz range (bass region) - the math confirms this. Furthermore, I offered graphs of 'music power distribution' to show that in addition to laws of physics, music somewhat contains more bass energy than the rest of the bandwidth. Even the BSC phenomena tells us 'bass region' is the tricky one.
So I suggested that a system's power performance is almost limited to the amount of power available for this crucial region.
Have you heard a 400W amp into a pair of passive speakers capable of handling it?
it'll blow the roof off.
Then go and listen to a pair Meridian (or ATC) active speakers with about 200W under the hood .
Sure they'll be beautiful and loud, but there would be no comparison.
Last edited:
Hi,
KT why don't you give it a rest. You don't know stuff.
Your graph is total b*llsh*t and verifies nothing. I
don't know where you got it from, but what it shows
in your terms is totally unrealistic, in any real world.
Of course a bi-amped 100W + 100W system cannot
better a 400W system, but the point is it can generally
do better than a 200W system with wideband material,
and in theory equal voltage swing levels of a 400W.
rgds, sreten.
KT why don't you give it a rest. You don't know stuff.
Your graph is total b*llsh*t and verifies nothing. I
don't know where you got it from, but what it shows
in your terms is totally unrealistic, in any real world.
Of course a bi-amped 100W + 100W system cannot
better a 400W system, but the point is it can generally
do better than a 200W system with wideband material,
and in theory equal voltage swing levels of a 400W.
rgds, sreten.
Last edited:
Yet you cannot actually produce any evidence that he is wrong backed up with your maths. It seems you 'believe' he is wrong, despite his being a decades experienced engineer. His assumptions are there. Tell me what is wrong. The article is once again not 'gospel' but explains things in an approachable manner with workings and assumptions and sources given. It is therefore worth the effort to read and understand
You are confused about how power works and possibly how SPL meters work. FWIW I do often look at the FFT of real music. If the peak meter shows 0dBFS then you won't have any of the FFT bins anywhere near that level
There is a 50/50 split around 350Hz. so no, the majority is NOT below 200Hz.
If you need to go louder than the ATCs you are either deaf or running a disco. The SCM100A is rated for 115dB/1m continuous. They still have headroom for peaks at that level. the SCM100A has 350W under the hood. Some research before posting just maybe? And I have heard then and they are sublime, if too insensitive for this thread.
Hi,
Whilst ~ true in practice the crest factor below 350Hz is much lower
than the crest factor above 350Hz. Most power is in the bass region.
rgds, sreten.
I won't argue with that, and it was covered in the article he didn't bother reading. However the discussion was on average levels as pink noise spectra were hauled out 🙂
Just doing some analysis on bruckner 9. Interesting energy balance on the recording I have. I need to dig more and redownload audacity after my recent hard drive rebuild. Not sure how much DG munge the low end on recordings these days.
Attachments
Hi everyone! Happy New Year!
It has been quiet here, but not without progress. I placed 12CXT coax on 2x2 ft foam board, so its open baffle for now. I will reinforce the baffle with truck bed liner as recommended in another thread.
I am tweaking simple crossover now, have 1 mH coil to the midbass, and 4.5 uF for now to the tweeter with L-pad. Sounds great on SE tube amp, with lows missing offcourse. Not only the cancelation, but 12CTX itself does not go too deep. So next step would be 2x15" woofers to cover the bottom end, biamp it after active crossover at 160 Hz.
Should be fun.
Well, I measured the fr response of the 12cxt open baffles and found that while on axis the response is relatively flat, off axis past 30 degree there is beaming in 1-2 kHz range. This is most likely due to 12" mid starting to beam, so I decreased the crossover fr by increasing the cap to tweeter. 6 uF was almost ok, 8 uF looks perfect. I had to dial down the L-pad to tweeter a little, as it raised the tweeters level. Now the L-pad is at little less than half. Fr response is flat on axis and off to the 30 degrees no problem.
There is just small up and down ripple in above 10 kHz, most likely due to less then perfect transition from tweeter horn to cone. B&C made quite a mess here with big voice coil lip around. I removed the dust cap and trimmed the lip with nail clippers as best as I could. Anyway, nothing is perfect, even Tannoy has some issue here.
Bottom line, these speakers sound good, with very simple crossover now. With 98 dB sensitivity, any amp will do.
Although they roll of below 100 Hz, some may not need sub. I do. Thats next project.
There is just small up and down ripple in above 10 kHz, most likely due to less then perfect transition from tweeter horn to cone. B&C made quite a mess here with big voice coil lip around. I removed the dust cap and trimmed the lip with nail clippers as best as I could. Anyway, nothing is perfect, even Tannoy has some issue here.
Bottom line, these speakers sound good, with very simple crossover now. With 98 dB sensitivity, any amp will do.
Although they roll of below 100 Hz, some may not need sub. I do. Thats next project.
Last edited:
That's right!
I encourage anyone who has speakers with 93+ dB/w/m to post some pics and short description.
Where in the world do you get flamed from? I read and understand the article, and am just saying that the power advantage is not quite as dramatic as the extreme example. Pretty mild stuff.
OK, here's my work in (very slow) progress. B&C DE250 on an Autotech Iwata 600, over a Faital Pro 12PR300 in a repurposed box. Xover by MiniDSP 4 x 10HD, BW4 @ 1040 on the woofer, BW2 @ 1380 on the tweeter. Amp Camp Amp driving the horns, Chinese built TPA3116 on the woofers. There are two NHT 10" subs for the bottom end. You can just see one peeking out from behind the speaker, the other is centered on the opposite wall.
The S&L Woofer Tester 2 measured the woofers at 96.3663 dB @2.83Vrms/1m. The responses shown were taken of the individual drivers on the tweeter axis with auto-detected gating, and summed in Holm.
I'm hoping to finally make some 'real' boxes and permanent horn mounts this spring.
Bill
Attachments
- Status
- Not open for further replies.