So, I am not your follower, don't have a hearing impairment, and I stumbled on this post because I follow the assistive technology tag. But, I just took a whole semester class on assistive technology, and I'm actually using speech to text software right now to write this post. And I feel a little bad that this post only has 6 notes. (Because I have often had a horrible time finding help when I was looking for assistive technology myself…)
Anyway, those disclaimers aside, I can tell you a few things.
First, the word "tablet" can mean a lot of different things. Anything from a $100 android tablet that can barely were run a web browser to a $5000 Windows tablet computer capable of doing video processing would count as a tablet. But there is a big divide down the middle of the tablet selection, and what you choose does make a big difference for your software options. Basically, you have to choose if the tablet runs a mobile operating system (iOS or Android) or a desktop operating system like Windows. (I don't think Apple makes any tablets that run the full OS X, but I could be wrong about that.) If you have the budget for it, going with a full desktop OS will generally give you more processing power (which is important!) and more software options.
On Windows, Dragon NaturallySpeaking is the current best speech to text software on the market. The problem with that is that it is speaker dependent recognition. That is, it creates a separate profile for each person using the software. That's not a big deal if your grandpa is sitting down and having a conversation with one person at a time. He would just make a profile for each person he regularly talks with, and load up that profile at the beginning of the conversation. But it's going to be really hard (if not impossible) to use in a large group of people.
(In fact, I suspect you won't find anything that works in a large group of people. The problem of how to just identify multiple voices in a recording is really hard, and computer science hasn't really solved that one yet :-( Microphone placement would also be a challenge in a large group of people.)
Dragon is available in multiple languages, but you would have to buy a separate copy of the software for each language, which could get expensive depending on how many languages you want to support.
Windows Vista and later also have some built-in speech recognition tools. I've never found the dictation recognition (that is, when you're just dictating a document) to be good enough to be very useful for text entry. Depending on how much hearing your grandpa has, he might be able to get enough information from the recognition to be helpful, but I suspect it would still be tricky. I would really recommend going with Dragon over the built-in tools. The accuracy is better, and it's going to be less frustrating for everyone involved.
One other thing you can play with on desktop is Google's speech recognition tools. This works as long as you have a reliable Internet connection. Just open up Google docs in Chrome, make a new document, and find the voice typing option in the tools menu. This has the advantage of being speaker-independent, which would make it easier to use with many people. It's still not going to do a good job with multiple people talking to it at once, though.
On mobile, I can only really talk about Android. Google's speech recognition tools are built in to various parts of the android operating system. They generally require you to either tap a button to begin listening to your speech or to say a call phrase like "okay Google." I've never gotten as good of recognition accuracy on my android device, but I do wonder about how microphone quality effects things. Regardless of what tablet you use, I would definitely look for a good quality external microphone. If a conversation partner is willing to wear a headset, that's ideal. If your tablet has a USB port, there are loads of USB headset options. Bluetooth should be an option on just about any device (more expensive, but no wires).
Applications on android can use Google's speech recognition framework, so you might find an app designed to facilitate conversation for people with hearing impairment.
Dragon has a subscription-based app for android. I've never messed with it, but their recognition accuracy is generally good, so it might be something to look into.
So, to summarize, the two technologies I would look at first would be Dragon NaturallySpeaking and Google's speech recognition framework. I'm undecided about which one I would rank first versus second. Probably it depends on the exact use case. If the goal is to facilitate short interactions with a large set of people, then I would lean towards Google. If the goal is to help with longer conversations with close family or friends, then I would probably look at Dragon first. But, like I said, I don't have hearing impairment, so take this all with more than one grain of salt. (I use speech to text software because of my physical disability.)
Also, just a couple other thoughts coming from the assistive technology course I took last semester.
1. If your grandpa is interested in medical intervention, he could talk to his doctor about a cochlear implant. People who lose their hearing later in life (and, in particular, grow up with spoken language) often have very good results with modern cochlear implants.
2. If he has any visual difficulties (also super common with aging), be sure to look in the accessibility settings of whatever device you choose and adjust text sizes/colors, mouse pointer size/shape/color, and any other relevant settings for his comfort.
3. Depending on the exact use case, some sort of a stand for the tablet might be super useful. Holding something like that up can get tiring really quickly. A case with a shoulder or neck strap to prevent accidental drops might also be a good idea (think of the strap on a camera). With any assistive technology device, positioning and durability are things to consider.
Once again, just my super geeky thoughts. Take them with a grain of salt.
