@Dyson - I literally had no idea what voice banking was until seeing this thread a few minutes ago, so I'd not be surprised if you've already found this, but just in case you haven't:
https://www.reddit.com/r/MachineLearning/comments/er3ng8/d_how_to_save_my_fathers_voice/
There are a lot of good suggestions in there on how (and what) to start recording immediately, so that you have the best possible source data available for when the technology develops sufficiently.
In terms of what laptop to use, recording sound is 'cheap' in terms of hardware generally (even 'lossless' codecs - obviously no digital recording is truly lossless, but are imperceptibly close) so would assume you won't need much, also you could start now with a basic laptop and get something better if required once you can find out more about requirements. Anything with 4 threads and 8GB of RAM will be more than enough, if it's comparable to what you'd need to record a single mic/instrument with no effects.
WAV files are generally the highest bitrate and are understood by everything, you can record in this format using windows inbuilt recorder, Audacity (open source) or a myriad of other software. Cakewalk is a good full featured recording environment that is now free, or for a mac, Garageband is free/cheap with logic being the higher end offering.
Most of these will also support almost any other format with the right codecs installed (mp3, flac, aac) - some info on different codecs here:
https://www.audiophileon.com/news/best-music-codec
I would also ensure you have multiple redundant copies of the recordings to prevent losing anything from a hard drive malfunction, just a few external hard drives or large USB sticks would be fine for this (the key is mutiple physical devices, so not just a bunch of partitions/folders
), just copy the recordings to them after each session (you could probably write a simple batch/python script to copy it all over so they would just have to double click an icon), or could look at a RAID solution of some sort. Either that or use a cloud based (google/onedrive) backup.
I assume the processing of the audio in to a voice model will be done on the service provider's servers - have you/they decided which of the listed services (on your linked site) will be used?
Once the model is compiled you may need something with decent CPU power to be able to build sentences from the model!
This post got a bit long and waffly, but I hope it helps you towards your goal.
Again, I would 100% recommend that they start making recordings (of basically anything!) ASAP so you have as much quality data as possible to work with. If you haven't already I'd also contact the service vendors directly as they will be able to offer you specific guidance (one of them posted in the reddit thread I linked!)
