Edgar

Don't you know? We have sister states up here in New England: "Varmint" and "New Hamster".

This discussion seems to have degenerated into a squabble about whose definition of "accuracy" is more ... accurate. For a long time I have maintained that the mark of truly great loudspeakers is that people either love them or hate them. And the particular loudspeaker that I have used as the representative example of this is none other than the mighty Klipschorn. Now, I have never heard the Jubilee, but it appears from this and other discussions that it takes over where the Klipschorn left off, with particular regard to the love/hate perception. That's high praise in my book. With regard to "accuracy", I will never claim that the Klipschorn is the most accurate loudspeaker that I've heard. It does not have the flattest response, or the broadest bandwidth, or the smoothest, most coherent sound. But it does something better than any other loudspeaker I've ever heard; it captures the dynamic, effortless nature of live sound. If the Jubilee is at least as good as the Klipschorn in that regard, and better than the Klipschorn in others, then it must be a truly phenomenal loudspeaker -- for people to whom dynamics and effortlessness are important. For others, there may still be better choices, and there's no need for anyone to be affonted by others' preferences. Greg

As with all things, it's in the execution. Part of this is related to the concept of gain structure -- cut followed by boost can raise the noise floor and reduce dynamic range. Boost can also easily be abused, as in attempts to use narrow PEQ filters to "fill in" room nulls (it is impossible). Cut is perhaps less easily abused, but ultimately there is nothing intrinsically wrong with either cut or boost, properly applied. Greg

I'm in general agreement, but I do have a couple of comments: First, the Jubilee is a BIG loudspeaker. Putting a BIG speaker into a small room can cause problems that EQ might not be able to solve. Many of the comments I've read in this thread about how the K402 sounds look to me like they can be explained by the fact that the listener was in the near field of the horn. Move away from the horn, into its far field, and all will be well. Room is not large enough to do that? Then the room is too small for the speaker. Second, if you're going to use an active crossover, BUY THE BEST THAT YOU CAN AFFORD. People think nothing of spending thousands of dollars on a preamp, then cheap-out on an active crossover. The crossover's job is MUCH harder than the preamp's job! Don't skimp here! Some of the modern pro audio digital speaker processers are simply outstanding, but you have to pay the price. Greg

For transient response comparable to that of 6 dB crossovers, but with much steeper rolloff rates, try this crossover topology. Copies of the paper are available free from me; just send a PM if interested. Greg

The noise is there on the disk regardless of how it's played. By summing to mono, you are cancelling common-mode noise and increasing differential-mode noise. Of course, the way that mono records were cut causes noise to be predominantly common-mode. But with processing, worst-case noise reduction is at least as good as best case with sum/difference (because sum/difference is still an option with post-processing), so it makes sense to me to record both channels. You are, of course, free to do what works best for you. Greg

With signal processing there is so much more that can be done to remove noise than just L+R or L-R, particularly if one has multiple copies of the same signal. And with modern digital signal processors getting more powerful in every generation, things can now be done in real time that were unimaginable not very long ago. Greg

I kinda like 'em. I think they'd disappear into the corners. I wonder if the same could be done with the Jubilee.

I recommend that you record in stereo, if possible. There are emerging DSP restoration methods that compare the two supposedly identical channels and use differences between them as the basis for noise reduction. Even if you don't use one of those methods right now, you'll still have the original two channels available if you decide to in the future. In the mean time, you can sum the recorded channels during playback on your PC and get the same results as if you had used a mono cartridge or a mono switch on your preamp. Greg

Joining this thread late, so apologies if someone else has already posted this link: http://ldsg.snippets.org/HORNS/images/roundbend/ Greg

Why not start with a Belle bottom (no pun intended)? It's 30.125" wide vs. 23.75 for the LaScala. There's at lease one set of Belle plans published in the forum messages. The K402 is almost 40" wide. Greg

K31 is a pretty sensitive driver to start out with. Running two of them adds 3 dB. Running in halfPI space adds another 9 dB. So 107 dB SPL is not outrageous. The horn is so short that it's probably not doing much at 40 Hz. Greg

I lurk at Audio Asylum, but I missed that one. It is not at all uncommon to find David McBean there. Greg

If you look at Figure 20.2 here, Fta is the angle between the y-axis and the line extending from the y-axis to the horn contour (the lines in the figure with a big dot on each end). For a given horn length, if Fta is 90° then you have a full horn for that given throat and mouth size. If Fta is less than 90°, then you have only part of what would be a longer horn if you could see the whole thing. Longer horn = lower cutoff frequency, but the more you shorten the horn (relative to its full length), the more mouth reflections become a problem. And if you shorten it to something below about 1/4 the wavelength of the design cutoff frequency, you are basically wasting your time. Greg

I think that the Excel spreadsheet automatically reduces the mouth area by 1/2 in each of the different radiating angles. That is to say, it gives a full-sized mouth for free (4 PI) response, half-sized mouth for floor (2 PI) response, quarter-sized mouth for double corner (PI) response, and eighth-sized mouth for triple corner (half PI) response. Hornresp does NOT do this. Hornresp uses exactly the same horn regardless of the radiation angle, but it computes the response of that full-sized horn radiating into the smaller angles, as appropriate. With Hornresp you can specify the tangent angle at the mouth of the horn, the maximum for a full mouth being 90°. I haven't figured out what the program uses for a default value; it doesn't seem to be 90°. Greg

You're welcome. Try opening up the Tractrix Flare calculation box by double-clicking on either S1 or S2, clicking on the "F12 Calculate" radio button, and entering not only the length but the Fta (tangent angle at the mouth). Greg

At 200 mph, you travel 1/4 mile in 4.5 seconds. The Top Fuel record is 4.428 seconds. It's close, but the dragster wins by about 21 feet.

And here's another way to think of THAT: I measured a dollar bill, and it's about 0.0035" thick. Forty-one trillion of them, laid flat, would create a stack of cash 2,264,835 miles high. That's enough to reach from the earth to the moon nearly 9½ times. Greg

I just noticed that I forgot to reset the Atc and Vtc parameters. If you set Atc to the throat area and Vtc to zero, the high frequency hump disappears. There is essentially no effect below 1000 Hz, though. Greg

Here's what I get for a pair of JBL 2226J in the horn that you describe. Modeled as an exponential contour. Somebody check my numbers, please. Greg

I took another look at my Jubilee HORNRESP model, and found some mistakes. I think that this one is better. The frequency response differences between new and old are very small, essentially identical below 100 Hz and above 500 Hz, with a maximum difference of about 2 dB at 200 Hz. Greg

Not really an indoor antenna, but could be mounted indoors: Gray-Hoverman.

Roy knows how to warp time. Seriously, though, don't forget that there's some horn length attributable to the depth of the compression driver. Greg

For what it's worth, this has been my experience, as well. Recently I've been designing a corner horn about 1.9 meters long. No matter what I do to the contour, the response cuts-off somewhere in the region of 45 Hz. The contour only affects the smoothness of the response. Basically it seems that a horn ain't a horn until it's at least 1/4 wavelength long. Greg

Roy can answer for himself, but I'll take the liberty of offering my opinion. There is more to performance than frequency response. For effortless, natural, undistorted, dynamic bass performance, I've never heard a DR that could approach a horn. Vented boxes are especially poor at this, in my experience. Standard disclaimers apply: IMHO, YMMV, etc., etc. Greg