When I placed the [bɑ] of [ɑbɑ] in front of the [ɑ] of [ɑdɑ] I got the result sounding like [ɑbɑ] to me. As the task was formulated the way that made clear it was not the result we were expected to receive, I tried to listen to the new sound file more attentively. I should admit there is a slight difference in the consonant sound in [ɑbɑ] compared to the initial file. It seems like place of its articulation is not entirely labial but switched a little backwards, closer to alveolar. In the second situation when I took [ɑb] from [ɑbɑ] and placed it before the final [ɑ] from [ɑdɑ] I did not again the pure [ɑbɑ] again. I did not sound completely like [ɑdɑ] but definitely was some dental-alveolar shade in the bilabial [b]. Changing the duration of the stop did not help a lot in both situations and I did not get the pure bilabial [b] when I doubled its length. It might mean that the vocal context plays a crucial role in perception of stops. .
When produced not in isolation burst that make up the consonant-stop sound somehow affects the acoustic structure of the preceding and following vowels, slightly changes the position of the vocal tract. As a result, sounds [ɑ] in [ɑbɑ] and [ɑdɑ] are not the same sounds, but allophones of one phoneme sounding differently in different surroundings. Listeners pay attention to vowel sound as well while perceiving stops. It is hard to splice in Praat because of the nature of that group of sounds. When we pronounce stops we release the burst, the noise (not sure, if I can refer to that noise as to friction here) that merges with surrounding vowels. It is very difficult to find the transition point between the vowel and the stop. .
2. When I switched [k] and [g] in [kɹim] and [gɹin], the result did sound a little unnatural to me. It was very surprising, because I did not get the expected [gɹim] and [kɹin] , but the resulting words were slightly different from initial ones, even there was almost no difference that my ears could catch.