Even it is another topic, I think it is the most interesting one this point, where we can make highly smooth waveguides that can do almost anything we want regarding the DI.
It is probably widely accepted that 1) the reflected sound should be spectrally "similiar" to the direct sound and 2) that the power response should not be flat, i.e. should be falling slightly. What I see as still not conclusive is whether the listening axis response should be also falling (in case of flat DI) or flat as possible (in case of slightly rising DI). This is really the same question as bmc0 asks, i.e. how much "similar" the reflected sound needs to be and if it's good to sacrifice the flatness of direct sound for this. Or is there a particular reason for the direct sound response in itself not to be flat?
So my question would be: Is the downward slope of the listening axis response with flat DI really the best combination?
I somehow feel that a flat direct sound response with a slightly rising DI could be even preferable but that's what would need to be found out.
It is probably widely accepted that 1) the reflected sound should be spectrally "similiar" to the direct sound and 2) that the power response should not be flat, i.e. should be falling slightly. What I see as still not conclusive is whether the listening axis response should be also falling (in case of flat DI) or flat as possible (in case of slightly rising DI). This is really the same question as bmc0 asks, i.e. how much "similar" the reflected sound needs to be and if it's good to sacrifice the flatness of direct sound for this. Or is there a particular reason for the direct sound response in itself not to be flat?
So my question would be: Is the downward slope of the listening axis response with flat DI really the best combination?
I somehow feel that a flat direct sound response with a slightly rising DI could be even preferable but that's what would need to be found out.
Last edited:
The answer can probably be found by reviewing the directivity index of the Revel Salon 2, and the fact that most listeners preferred it to the JBL M2. Even though the M2 is probably more dynamic.
I think there were so many variables in that test that it's not safe to make this conclusion.
For me this would take several different waveguides (maybe free standing) in the same setup, with the only difference being the DI and listening axis response slope(s). Which could be done, I think.
For me this would take several different waveguides (maybe free standing) in the same setup, with the only difference being the DI and listening axis response slope(s). Which could be done, I think.
Last edited:
I'm not certain that this point is widely accepted. I usually run my speakers with a flat listening axis response, which of course also means that the power response is flat from 1kHz to 15kHz.
Last edited:
Which must sound very bright unless you have a highly dampened room. 😱
- Maybe instead of talking of the power response of the speakers alone we should consider the averaged room response at the listening location. Is that still flat in your case?
- Maybe instead of talking of the power response of the speakers alone we should consider the averaged room response at the listening location. Is that still flat in your case?
Last edited:
It's nearly flat from 1k-10k (falls about 1dB). Below 1k it's obviously a bit more sloped. The room has pretty even decay (vs frequency) and is not particularly dead. I point the microphone at the ceiling for in-room measurements. Measurements done with the microphone pointed at the speakers will have more HF falloff.
Personally, I don't find the system to sound excessively bright.
Personally, I don't find the system to sound excessively bright.
Last edited:
Well if that's the right tonal balance for you, a different waveguide would hardly made an improvement.
Do you remember how was the M2 set up regarding the EQ? If I'm not mistaken it leaves the factory "flat" and to be fine-adjusted later in place. Is that right?
Regardless, I think there's probably a perceptually significant difference between tilting the power response and tilting the direct sound. As far as I'm aware, with the exception of very early reflections, the human auditory system can largely separate direct and reflected energy above the modal region of the room. If that statement is true, it seems likely there'd be a perceived difference between the two aforementioned cases.
As Power response is the total sound power radiated into whole space vs. frequency, the statement
A speaker may very well be level straight on axis but may lack energy in certain direction due to phase problems (see e.g. MTM config).
//
seems strange?
A speaker may very well be level straight on axis but may lack energy in certain direction due to phase problems (see e.g. MTM config).
//
My speakers have a flat DI (±0.5dB) from 1kHz to 15kHz (see my post here). The quoted statement obviously isn't true for all speakers, but it is true for the speakers in question.
Last edited:
I thought about the error function criteria for the optimization algorithm a bit. Maybe I would define a target DI curve and evaluate the error function as the residual sum of squares between the target and an average DI curve calculated for a selected listening window (not necessarily containing the on-axis response). Maybe this all weighted with frequency somehow.
Last edited:
Sorry but the inference is still not true. Your power response can still be flat - sure.
//
I think it is true. That's the definiton of DI.
Flat on-axis and flat DI means flat power response.
Flat on-axis and flat DI means flat power response.
Last edited:
Late arriving sound will still affect the perception of timbre even if the ear can separate the signals. And "very early reflections" are not something that we can just write off as they are crucial to imaging.
After having a very similar discussion many years ago here on DIY, we concluded that as the DI rises the listening axis level must go up to sound "neutral". With a more constant DI the listening axis level needs to be lower to account for the added HF signals arriving from the room due to the reverberant energy. This not only seems logical - to me - but is exactly what I have found in practice.
The problem with the higher/rising DI is that the listening space over which the sound is neutral is much smaller than that with constant directivity - a not insignificant advantage.
Last edited:
You mean the target? That could be anything but most probably just a flat or tilted (poly)line. Generally anything as a set of (f, DI) points. That's not really the point.
Last edited:
I didn't say that late-arriving energy would have no impact on timbre, but that changing the spectral content of the late-arriving energy would be perceived somewhat differently than changing the spectral content of the direct sound.
Regarding VER, I wasn't writing them off. My point was that they are perceptually "fused" with the direct sound, which explains why high-level VER affect imaging (shift and spread) and timbre so much.
Sure, that's true, but it doesn't say anything about preference or even neutral timbre wrt DI. But it seems to me that if the two - early and late arrivals - have similar signatures (spectral content and temporal) then I would suspect the later to be less "noticeable".
Agreed.
What I fail to see is any rational for increasing DI(f) other than "it sounds good to me." I'm not a big fan of that kind of evidence.
I think no such claims were made. There's really no strong evidence either way and that's the point. For one thing, as the results show, somewhat increasing DI(f) typically comes with (or allows for) a better overall smoothness - and according to Olive's findings this is what counts the most. It could be interesting to find out more. I agree that in a situation where there's the demand to cover considerably more listening seats than one, the best constant directivity is really the most desirable feature.
Last edited: