...But weren´t able to do that - so If you could give me some hints here, would be appreciated.
<tutorial>
First, the goal (stated clearly): finding early reflections in the (REW) measurements, and correlating how strong they are and where they are in-room.
Next, the most-used plot to find early reflections: in REW, the energy-time curve (ETC), which is found under the "Filtered IR" (Impulse Response) tab.
Below you will see "ETC" overlay curves for a couple of measurements of my center loudspeaker (tri-amped CP25 tweeter, K-510/K-69-A midrange horn/driver, Belle bass bin), using two different techniques:
Sort of a mess to look at. But essentially the top orange trace was taken at the listening position (LP--about 3m away in-room, centered on the tweeter axis) without any temporary absorption on the floor between the microphone and the loudspeaker. The bottom trace is taken at 1m with 2-3 heavy blankets laid out transversely on the floor between the microphone and loudspeaker. Why? The human hearing system is used to hearing floor bounce, but REW doesn't know what to do with it, so removing it temporarily for acoustic measurements is necessary.
We are looking for reflections in-room from 0 ms (the shifted time to peak of impulse response) to about 6 milliseconds-ms). These are the most troublesome in terms of disturbing the soundstage/imaging of loudspeakers. Measuring one loudspeaker at a time (only--
not two loudspeakers at once).
We see three peaks in the orange trace at 2.5 ms and 3.65ms within the first 5 ms. How far are these reflections from the tweeter's exit? The speed of sound in air at room temperature is ~345 m/s. Multiply the speed of sound by the delay seen, 0.002 x 345 = 0.69m and 0.00365 x 345 = 1.26m (that is, ~27 inches and ~49.5 inches for US-based readers).
What was at this distance in-room? The floor (in this case, bare areas of a tile floor on either side of a dog bed between the loudspeaker and microphone). That's why you don't see those same three reflections in the blue trace, because they were suppressed by the blankets.
What does this tell you? It's difficult to see. Why not try another view type to see what's happening? Spectrogram view. First the blue trace (suppressed floor bounce with microphone at 1m):
Now the measurement with floor bounce taken at the listening position (LP) of 3m:
Now we have something that is much easier to pick out. Time on the horizontal axis (just like the ETC curve), but now we have frequency (including the acoustic size of the reflected object) vertically, and acoustic (SPL) intensity in color.
We can see the two floor bounces at 2.5 and 3.65 ms from just above 2 kHz up to 10-11 kHz. Is that information that you can use? Yes. It tells you the thickness of the absorption on the floor that you need to absorb those floor bounces. Thin is okay here.
What else can you see in the spectrogram plot? (A lot...) It shows you a lot of spread out energy from 600 Hz to 1 kHz. Also below 300 Hz.
Now let's look back at the 1m spectrogram with absorption on the floor to control floor bounce: those extra areas of acoustic energy are not there. Same loudspeaker, different microphone distance, with and without floor absorption.
Are you beginning to see the pattern between comparing these two spectrograms? It tells you a lot. It also tells you the approximate size of the acoustically reflecting object (i.e., a half wavelength fits across the dimensions of that reflector corresponding to frequency). Then you can go on a Sherlock Holmes expedition to find the objects in-room that cause early reflections.
You can also do this with a coffee table between you and the loudspeaker, or a rack between two stereo loudspeakers (with and without temporary absorption on over the rack to absorb early reflections). You can do this with large TV screens (with and without a blanket hung over the screen area). The key here is to also listen carefully to each case and correlate what you hear to those conditions. Some absorption will be preferred, others, not to much.
</tutorial>
Chris