The best articles for this are [1] (google translate it) and [2]. In [1], the author uses .asc files while in [2], .img files are used. You can treat these files as basically the same as .hgt but remember to reproject the data as different DEM sources uses slightly different projections.
Not to be confused with citizenlab.ca, a research lab of University of Toronto that deals with high profile civil right violations using OSINT and tech analysis.
This is not correct. The security page only shows that they use Google for Speech to Text.
Instead, Text to Speech is done using technology developed by Lyrebird.ai, which Descript has bought over. Descript rebranded it as Overdub. Note that style transfer learning of voices is a hard problem and Overdub seems to have nailed it perfectly. I speculate that the underlying technology of Overdub is based on sv2tts (https://arxiv.org/abs/1806.04558).