Japanese similarly is primarily done with QWERTY and an IME. Other layouts do exist but they're not very common.
Chinese input is quite varied though. Taiwan uses Zhuyin (or bopomofo) which is a phonetic alternative that doesn't use Latin. Hong Kong, Cantonese, and/or misc. types of Chinese may use other methods like Cangjie (which does use a QWERTY layout). Cangjie is notable for not being phonetic at all -- likely similar to the typewriter input methods mentioned in the article. Words are typed as ordered sequences of pseudo-radicals (pieces of Chinese characters) that are memorized by their associations to Latin letters and color[0].
Although there's no real reason either Japanese or pinyin IMEs should have to use QWERTY -- in fact it's likely quite suboptimal as neither make use of all the keys, or at least not as a primary function (many IMEs are configurable and/or allow non-standard characters to map to more complex sequences of input). I personally use both through Dvorak (which, being optimized for English, is similarly suboptimal).
As for voice input, Chinese is easier than Japanese because of the larger phonetic inventory and use of tones to form minimal pairs. Although the different types of Chinese (regional dialects (as opposed to different Chinese languages)) and their variations in pronunciation complicate things. Japanese does have pitch accent, but it's not really consistent across Japan, so I don't know how helpful it would actually be for a voice recognition system.
Purely from personal experience, I can rarely get Japanese voice recognition to work if trying to search for a word in isolation. It's quite good with sentences though. This might just be an artifact of my poor pronunciation though.
I think typing in general, as it is now, is pretty primitive and has a lot of room to be improved. I've been tempted to try to make some new, and possibly more efficient, input methods, but haven't been able to justify giving any a try yet. I do think even an IME at the level of what currently exists in Chinese and Japanese would be wonderful to have for English though. Intelligent auto-complete, shortcuts, semantically aware insertion of unicode (Emoji are actually natural to use in IME-using languages; just type what you want and select the emoji that pops up).
[0]: https://en.wikipedia.org/wiki/Cangjie_input_method#/media/Fi...