FWIW, I have heard the Lansche, albeit in an unfamiliar room and system. It seemed excellent- or at least it wasn't a limiting factor to the system sounding excellent. Very clean sounding at what seemed to be quite a high volume. The RAAL 70-10D is the best tweeter I've heard under familiar circimstances- I own a pair- but they need to be crossed over relatively high, IME. As noted above, a lot depends on where the hf driver needs to be crossed over in the context of the whole system.
Thank you for the feedback, it gives an idea.
I'm very familiar with all RAAL and in my opinion the 70-10D is the least impressive of them all (but still a good tweeter compared to non-RAAL tweeters)
on the cheap and effective side, I think a K-tube and modest 1" format compression driver makes a nice tweeter (won't go "low")
I'd recommend testing the Scanspeak D3004/604010 Beryllium small format dome tweeter. I want to see how the best small format dome tweeter hangs with the big boys.
FWIW I've found only one THD test of the RAAL 70-10D which was not published by someone who sells them: Robsan-DIY | RAAL 70-10D
I am rather underwhelmed.
I am rather underwhelmed.
Tweeter in isolation (say, above 2kHz) does not work as a listening test, except for incredibly low listening levels. Otherwise the ear's distortion will mask everthing (tons of IMD). Even the best tweeters sound like sh1t when listened to with highpassed music signals.
Agreed, that is why it will be matched with a woofer/midrange.
woofer/mid will be fixed, the same for all tweeters tested and obviously all level-matched to said woofer/mid.
Ok, then you're actually testing a 2-way. The hard part will be to get the crossover right. For all setups you'll need to have the exact same acoustical transfer function (amplitude *and* phase) down to -30dB at least, otherwise you'll test mainly the crossover, not the actual tweeter.
It can be done of course, but it will be a lot of work -- even with expert tools like LSPcad pro etc.
It can be done of course, but it will be a lot of work -- even with expert tools like LSPcad pro etc.
If testing this way, use of a dsp-based xover is probably mandatory, because the acoustic centre of the drivers might be different and has to be compensated for.
For a better understanding, please refer to the midrange blind testing on the link below. We'll use the same room and set-up for the tweeters. Except a midrange will be installed underneath the Cube.
miniDSP 48db/oct is used.
Now it's plateamps, but soon it will be nanoDigi so we can use some high quality DAC.
http://www.diyaudio.com/forums/mult...idrange-blind-testing-need-your-help-100.html
From a purely scientific point of view, the tweeters blind test is more flawed than the midrange blind test... Yes, because of the 2-way thing.
While the midranges are tested alone (360hz-7.2khz) and somewhat listenable in that limited bandwith, there is no way to do the same with tweeters, obviously any 2khz-20khz listened alone would be intolerable.
That is why we must match a midrange in the test set-up. Hopefully, that will be a midrange that emerge somehow from... the midrange blind test, and therefore will be acceptable as a reference.
While the midranges are tested alone (360hz-7.2khz) and somewhat listenable in that limited bandwith, there is no way to do the same with tweeters, obviously any 2khz-20khz listened alone would be intolerable.
That is why we must match a midrange in the test set-up. Hopefully, that will be a midrange that emerge somehow from... the midrange blind test, and therefore will be acceptable as a reference.
Ribbons with low distortion are rare. Usually, rather high frequency crossover is required.
You mean something like a LR8 XO (electrical) between tweeter and woofer? That will not solve the issues we mentioned (though better than any shallower slope XO).
8th order is pretty much too steep in my book and will promote directivity jump issues.
Last edited:
Also note all tweeters will be EQd and level-matched, same as midranges are.
Educated-guess here, based on what i observed so far with the midrange test:
There will be more noticeable differences between tweeters than there is between midranges (once EQd and level-matched, of course).
I might be wrong, but that's my feeling.
Educated-guess here, based on what i observed so far with the midrange test:
There will be more noticeable differences between tweeters than there is between midranges (once EQd and level-matched, of course).
I might be wrong, but that's my feeling.
The first part of this test (and the midrange as well) will be an identification one.
Before any appreciation given, we must be sure we can identity.
The directivity issues, theoratically, should help identify drivers from each others. But preliminary tests (midrange) shows that participants are having difficulties to identify drivers which are extremely differents.
I won't put the same ressources on the Tweeters blind test as i'm doing on the Midranges test, because of the reasons mentionned above, but also because many ''top'' tweeters seems very difficult to obtain and/or very expensive.
However, i will test what i'll find.
However, i will test what i'll find.
For me the question still is : are you going to use precisely defined acoustical targets for both woofer and tweeter or not? What is the actual acoustic XO function you plan to use?
Not sure to understand your question, but i plan the use the steepest slope available (48db/oct in that case) at the lowest frequency possible based on the limits of the drivers. Target is 3 octaves, but it could go to 2½ or 3½.
If we want to compare a big AMT or dome with a small ribbon or AMT, we won't have the choice but to cross higher.
- Status
- Not open for further replies.