I made a test that might show the problem pretty well (make sure you only tap once). The measurements are usually not considered accurate though so take it with a grain of salt.
wackytoaster.at/parachute/audioLag.c3p
For explanation, this takes three timing measurements:
Once the tap is registered, once the audio starts playing and once the audio stops playing.
And then shows timings based on:
Duration between registering tap and playing audio, duration from audio start to audio stop and discrepancy between the measured audio duration and the actual audio duration.
On the Samsung Browser: No audio lag, low duration from tap to audio playing, discrepancy is in the low ms range, 10ms at most or even in the negative. I´d say an expected range.
On Chrome: Audio lag, low duration from tap to audio playing, discrepancy is often arund 50-60ms.
As it seems the tap is indeed registered asap, but the audio lags behind? However, 50ms isn´t all that much (3-4 frames) so considering how long the lag is... I dunno if it´s that or only that. Another thing is that the discrepancy sometimes even shows high negative values (-40ms) but there is still a very noticeable lag.