If I understand this correctly, we are unable to hear reflections below -20dB... so in fact I would not hear any reflections in my room when using the absorption panels.
From:
https://users.aalto.fi/~ktlokki/Publs/mst_laukkanen.pdf
3.4.1 Toole et al.
Figure 9: Detection thresholds for a single lateral reflection. Different colors represent
results from different studies conducted with various test signals exhibiting
different degrees of temporal extension (continuity). All studies are conducted in
anechoic chamber. Drawn after Toole [56].
Toole et al. have contributed significantly to the research of sound perception in
small rooms [42,55,56]. They have striven to solve the relationship between objective
measurements and perceptual effects in small room acoustics. However, as Toole
states in [56] and [55], a lot is yet to be done to completely understand sound
perception in small rooms. Toole et al. have studied sound perception both from
the entertaining point of view, as well as from the perspective of critical listeners.
Thus, a large part of Toole’s research is relevant to this thesis.
The most interesting results from Toole’s experiments are the thresholds for
different perceptual cues for reflections coming from different directions in different
time delays. Large part of the listening tests have been done with speech, but some
studies also with music. Figure 9 illustrates the results of different studies in the
reflection perception. All studies were conducted in anechoic listening conditions,
with slight variations in the horizontal angle for the lateral test reflection. From Fig.
9 it can be picked that with pink noise and classical music, the perception threshold
for a single lateral reflection is approximately -20 dB related to the direct sound
during the first 50 ms. With clicks, threshold is as high as -10 dB within first 3 ms,
but drops dramatically after that.
Figure 10 illustrates results of Barron [4] drawn after Toole [56, p. 87], where
the perceptual effects of single lateral reflection arriving from 40 to the side was
studied. Listening tests were made with classical recordings (Mozart) [42]. Figure
10 illustrates that lateral reflections coming within the first 10 ms causes image shift
towards the direction of the reflection. In control room, this means widening of the
stereo image and thus enhancing the sweet spot size if the room is symmetric. The
amount of widening required for the stereo image is a matter of opinion. Compromise
has to be done between the stereo image accuracy and the widening of the stereo
image, e.g., size of the sweet spot. From Fig. 10 it can be also seen that lateral
reflections within 10 - 35 ms introduce tone coloration regardless of the level related
to the direct sound.
Other notable point in Toole’s experiments is that the sequence of several lowlevel
reflections and a large single reflection were observed almost equally loud [56, p.
91]. The message here is that it could be misleading to assume that if large reflecting
surfaces are broken on the basis of impulse response measurements, the audible
effects of reflection will be reduced. This kind of effect discloses the persistent problem
in a relation between measurements and psychoacoustics, as human perception
is usually nonlinear while measurements are linear.
Considering room reflections, Toole concludes that the reflections from front and
back do not have any positive effects on the listening preference. Instead, early
lateral reflections from the side seemed to contribute in a positive way at least in
entertaining purpose. It can be also concluded from the study of Imamura et al. [26],
that absorption of the first reflections on the side walls causes "the width of sound
image" to be narrower and "Envelopment" to be lower. Absorption of the first
reflections on the front wall and ceiling make "the width of sound image" narrower
and "Clarity" increase. Absorption of the first reflections on the back wall also
makes "the width of sound image" narrower and "Clarity" increase.
In [56, p. 177], Toole presents results of a massive listening test conducted
by Olive et al. [41], where three different loudspeakers were subjectively rated in
four different rooms. First experiment was conducted in a "live" manner, where
listeners evaluated all three loudspeakers in one room before moving to the next one.
Reproduction was recorded binaurally for each loudspeaker-room combination, and
the same test was conducted using headphones. Results showed that a loudspeaker
was highly significant and room was not a significant factor and that the results
of live and binaural tests were essentially the same. From this it can be deduced
that listeners adapted to a room and were able to judge the pure loudspeakers quite
accurate.
In a second test, using the same binaural recordings, room-loudspeaker combinations
were judged in different context. Now listeners rated the same loudspeakers
located in the same position between the four rooms. Thus, there were four comparisons
per one trial. In the second test room became the highest significant variable
whereas the effect of loudspeaker was not found significant, meaning that there were
highly significant differences in preference due to the room factor. These results show
that adaptation has a major effect when judging loudspeaker performance. However,
this is not saying that there were no interactions between individual loudspeakers
and individual rooms. There actually were, and especially at low-frequencies. It was
also noticed, that in multiple comparison tasks, listeners tend to make judgements
on a relative scale but they were less able to make consistent judgements on an
absolute scale. Thus, test showed that the context in which comparison were made,
influences listeners’ preference ratings significantly.
3.4.2 King et al.
Figure 10: Subjective effects of a single reflection arriving from 40 side, including
the effect of reflection level and delay compared with direct sound. Data is from
Barron [4] and picture is drawn after Toole [56, p. 87]. Experiments were done
with classical music. The lowest curve indicates the hearing threshold. Above this
at short delays (less than 10 ms), listeners reported an image shift in the direction
of reflection. At delays larger than 10 ms, listeners reported "spatial impression"
where the source appeared to broaden and the music started to gain body and
fullness. The spatial impression increased as the level of reflection increased, which
is illustrated in the figure by the increased shading density. The curve of an equal
spatial impression shows that for short delays, the reflections must be higher in level
to produce the same effect. At high levels and long delays, disturbing echoes were
heard, which is the upper right corner in the figure. At delays between 10 ms and
40 ms and at all levels, some tone coloration was heard (colored brush strokes in
the picture). The areas identified as exhibiting an image shift refer to impressions
that the principal image has been shifted toward the reflection image. At short
delays, this would sum up the localization of reflection to the leading loudspeaker.
At longer delays, the image would likely be perceived to be larger and less clear.
Finally, at yet longer delays and higher sound levels, a second image at the location
of the reflection will be perceived as an individual sound source. From the data of
Barron, it is not clear where exactly these divisions occur.
King, Leonard and Sikora have conducted several studies related to the effect of
room reflections in critical listening [32] [34]. In [32], King et al. explored sound
engineer performance when lateral first reflections were, minimized, maximized and
diffused. Sound engineers were asked to adjust the level of a female soprano to
the orchestral backing track in all these three conditions. The results shows that
there were slight differences in the time needed for completing the task depending
on the order in which different treatments were tested. However, after adaptation
to the test procedure, normal performance was achieved. No significant difference
was noticed in accuracy between different treatments. Conclusion was that due to
the adaptation, no prominent difference in sound engineer performance occurred
between three alternative side wall treatments. However, there is a limit to what we
can adapt to, and as Toole speculates in [57], adaptation very likely utilizes some
portion of our neural capacity and perhaps causes fatigue and stress. Thus, working
longer periods of time in conditions that requires adaptation to a certain deficiency
or a coloration could probably be unhealthy for mixing engineer. However, this kind
of effects are still waiting to be proven and topic needs further research by neural
scientists in collaboration with acousticians.
In [34], Leonard et al. studied the effect of a room in adjusting reverberation level
in a mix. The task was to add a reverberation to a fixed stereo mix, first in a standard
studio control room (averaged T60 of 0.2 s), and second in a highly reflective mix
room (averaged T60 of 0.4 s). Results showed significant differences in reverberation
levels set in each acoustical environment. The reverberation was mixed 1.32 dB
lower in the reflective environment on average. Thus, the conclusion was that the
room treatment had a significant effect to the reverberation level adjustment.
it is just different. 
