In a recent Japanese movie, the leading actor is acting as Uncle Vanya on a stage, saying 「おまえはぼくの人生を踏みにじった」, in which the 「じ」sounds to me very similar to 「し」.
I wonder how this sounds to a native speaker and what's happening here:
【Edit】
Recently I heard two more cases, where the 「じ」 in 「人生」 sounds to me very similar to 「し」. I wonder if all the three cases can be uniformly explained:
I'm a native Japanese speaker, and it sounds like じ to me. It does sound slightly "weaker" or "less explosive" than the first じ (in 人生) if I listen to them carefully, but the second じ (in 踏みにじった) is voiced enough.
If you heard a clear difference, it may be because じ is pronounced differently depending on its position in a word (i.e., allophone). じ is typically pronounced as [d͡ʑi] at the beginning of a word, and [ʑi] in the middle of a word. As a native Japanese speaker, [d͡ʑi] and [ʑi] sound almost identical to me, but these sound clearly different from unvoiced し ([ɕi]).
Related (the last one is your question):
Your ears are picking up on a difference, but strictly speaking it is less about the degree of voicing and more about the degree of sibilance.
I found another sample of a more run-of-the-mill ふみにじった and have created a audio file comparing it with your sample:
Using this audio, I generated a spectrogram with Audacity and manually labeled it here:
From this we can see that your sample is indeed different, in that it has much louder sibilance: the higher frequencies normally associated with し, generated by air moving quickly between a tight passage between your tongue and roof of your mouth.
But, it has nearly equivalent amounts of voicing: the lower frequencies generated from your vocal chords vibrating.
(One note is that the voicing in yours does seem to end a tiny bit earlier in preparation for the っ, which also likely contributes to what you are hearing.)
Either way, the presence of voicing throughout the majority of the consonant is enough for it to sound like じ as opposed to し to a native speaker in this context.
