kimmosto
Active Member
- Joined
- Sep 12, 2019
- Messages
- 215
- Likes
- 513
No. That was consumer's/your perspective. Designer, manufacture or audio salesman don't select speakers for you, build listening room acoustics, locate speakers and listening point and listen there so consumer/you should be able to study differences and optimize combination of speaker, listening setup and environment. Part of that is to understand or study empirically how different concepts/constructions interact with environment and how different spectrums such as acoustic resolution and soundstage depth can be controlled and balanced.You are asking as a speaker designer. This is the wrong perspective (not inherently wrong, wrong for the question I asked you). The correct perspective is that of assessment, when the design decisions have already been made and you are looking at a complete product, like the KH150. So how should we assess waveguides? By using what data?
This may apply your small room, but generally EDT is one the most important parameter especially if preference is high acoustical resolution / low noise due to flutter echo.
- Doesn't matter.
Very small rooms such as <15m^2 are special cases though octave transmission index is able to indicate acoustic noise and resolution also there. I've tested that analysis from tiny 6m^2 rooms to few hundreds of m^2, more than 50 speaker+room combinations within ca. two decades. It's not perfect and all inclusive for sure because it cannot "hear" for example effect of envelopment and 3D dimensions of sound stage, but it's easy and visual for revealing differences in acoustic resolution; level and especially balance. Note that I'm not talking about weighted total STI.
- STI is an irrelevant metric in small rooms. This is used in large buildings, classrooms, lecture halls.
That looks intentional misunderstanding. Please interpret so that individual responses and response groups such as on-axis, LW, ER, PIR, SP are kinda "ideal" i.e. linear, and tilt of those responses/groups variates so that we have few significantly different DI tilts to compare.
- DI is meaningless on its own since its based on aggregated data. You need access groups of curves to understand speakers.
Of course directivity variations such as too steep tilt down or clear step in power (and PIR and DI) are the most common errors I don't love. Many commercial applications have a bit too large and deep horn/wave guide for tweeter alone. That causes some power step and power dip at crossover range with phase-matched XO. Interaction with room splits into two different frequency bands. LF...MF has low acoustic resolution without very effective acoustic treatment, and tweeter range is very dry - mostly direct sound without natural sound distribution via room. Sound does not breathe in balanced way such as with natural instruments - especially to off-axis, and listener is kinda forced to hear details at HF. Perspective to music is narrowed because focus is at HF details.All I'm really asking for is some data or evidence which—separate from speaker directivity in general—shows that waveguides produce unnatural, not open, etc., sound. If all you are really talking about is directivity, then fine, all's clear and the conversation is over. But if you are saying that waveguides have some kind intrinsic weaknesses that produce a characteristic sound, then I want to know what that means.
Another common problem is unbalanced compression spectrum which turns sound bright and harsh when LF radiator heats up while HF compression driver with horn maintains pressure. So LF radiator and it's power amp should be quite much over-dimensioned to maintain balance. For example 10" woofer + 1" Ti driver in a horn could be a problem.
Horns can produce also higher IMD than separate direct radiators. This can be tested with (tractrix) horn with coaxial compression driver. Lower tone to mid and higher tone to tweeter range. IMD is not necessarily very bad, but it is there. I have few old measurements in archive, but usually I don't save this kind of data. Nothing you can do it so better to listen and try to forget.
Also NBD (near band deviation, familiar with preference rating) could be two...three times higher than with conventional hifi drivers. Some coloration may remain no matter how many PEQ bands we set to DSP to make ON and LW as flat as possible. Also this is highly case-dependent because there are better and worse equipment...
The last funny feature is quite emphasized/shooting sound of some transients and percussion. I've never tried to investigate actual reason, but it might be related to "unnatural" wavefront or/and flare. I don't have horns anymore at home so please don't continue to ask data.