Just going back through some of the comments and questions above...
Regarding the effects of grilles and lamellae: Anything put in from of a loudspeaker (or any sound source, for that matter) that has (as a rule of thumb) a size of more than 1/4 of a wavelength of the frequency content will have an effect on the directivity of the source.
If you have multiple things that are causing this effect, then you may/will have a subsequent effect on the magnitude response (the "magnitude response" is what people mean when they say "frequency response") at any one location in a free field (a "free field" is a space that is free of reflections - an infinitely big room, for example...).
If you're building a loudspeaker that has such things (say, an aluminium grille or the lamellae that has an acoustical impact) then these effects are measurable as a function of level vs. frequency vs. angle. (a fancy way of saying "the frequency responses in different directions").
If the multiple things in front of the source are regularly spaced, then the angle-dependent magnitude responses are more likely to exhibit a regular pattern vs. angle when measured in a free field.
So, we then have to consider what the measured effect is:
- In the very simplest case, you have no effect. Or, the measured effect is so small that it's less than the difference caused by a cat walking into the room.
- In a simplest case, you have an impact on the magnitude response that is the same in all directions. This is unlikely, but possible. (For example, in a case where the grill is so restrictive that it creates a resonant cavity behind it (and in front of the driver) and that resonance is measurable everywhere.)
- In the more likely (and most complicated) case, you have an impact that is different at different angles. Then the question is "how big is the impact, and how much does it change as a function of angle?"
After those evaluations are done, then the first question to ask is "although this is a measurable effect, in what cases could it be audible?" If the effect can only be audible in an anechoic chamber while listening to anechoic xylophone music (yes... I have a recording of anechoic xylophone music - on one of the few CD's that B&O has produced...) then there's not much to worry about.
Another question is: "is this something that can be taken care of "upstream" in the signal processing?" As a simple example: if the effect of a grille is to create a single minimum-phase resonance at one frequency with a fairly low Q, then this is something that can be easily "undone" in the signal processing. You just put a filter in the system that has the opposite phase characteristic, and it will cancel the effect of the resonance in the time domain (and also, therefore, in the frequency domain).
Another question to ask is "does the effect matter, given the expectations and use cases of this loudspeaker?" For example: the grille on a portable loudspeaker may cause diffraction that results in a wider directivity (or "diffusion" as some people call it) - but this is exactly what you might want for such a product. (The acoustic lens generally falls under this heading - it's a thing in front of a loudspeaker driver that has an effect on its directivity - but it's on purpose...) This could also be considered to be an advantage for a centre loudspeaker, since the purpose of a centre loudspeaker is to make the centre image more easily locatable for a large group of people. If you had one chair and no friends, you would not need a centre loudspeaker - assuming your chair was in the right place... So, a highly directional centre loudspeaker makes no sense.
So, to take this into the real world: When we make a loudspeaker with a grille, we do the initial measurements without the grille, and then repeat the measurements with the grille - and assess whether we need to react. Maybe the effect is negligible. Maybe the effect is measurable, but not audible in most cases. Maybe the effect is audible, but we can undo it using signal processing. Maybe the effect is audible - and desirable. Or maybe it's really bad and there's no way to fix it. In that last case, we'll have a meeting to start thinking about what we can do about the grille. At least one of those things will happen (probably more than one - since we're talking about different effects at different frequencies...)
Jumping to the end of the development: the final assessment and tuning of the loudspeaker (here "tuning" is done both with measurements and by listening to the loudspeaker in different rooms and situations) is done on the complete final product. So, if you listen to the loudspeaker with the grille/lamellae/whatever OFF, then you are not hearing what we heard when we decided that this is the way the loudspeaker should sound.
However, it could be that, due to your preferences, or your room, or your cat, a loudspeaker without its grille on sounds better to you with the song that you're currently listening to, an the volume you're listening to it at, in the place where you're sitting, today. (As a simple, but highly unlikely example due to the extreme numbers: If a loudspeaker fabric drops the level by 1 dB at 20 kHz, and therefore we have put in a boost of 1 dB at 20 kHz to compensate for this, but you have a slight hearing loss of 1 dB at 20 kHz, then by taking the fabric off the loudspeaker, you'll get a 1 dB boost that will sound better to you...)
A long answer for a simple question.... sorry.
-g
