We will have to look at the games to see what audio format it uses.
In general on the DS then most video is usually silent with the audio coming from elsewhere, which is nice as it spares you having to remux video files if it is the case. Most games on the DS use the SDAT audio format which encompasses three types of audio
1) SSEQ. This is kind of like of like midi. Flanked by SBNK files that contain the sounds they play (if you are sticking with midi and similar terminology then this would be the sound font)
2) SWAV and sometimes seen as SWAR (swav archive) where potentially thousands are collected together. Usually used for things up to about 10 seconds of sound effects, voice clips and such like.
3) STRM. These are long audio streams sometimes minutes long (the first Tony Hawk game is a reasonable example)
If you just want to poke around inside the file to see what is there and maybe rip some things then
https://github.com/vgmtrans/vgmtrans is probably where most start (downloads a bit lower down on the page for various operating systems).
Samples and STRM files are often somewhat lower on the sample rate side of things, sometimes to get the space down a bit (audio files tend to be by far the bigger chunk of the game and getting it down makes a smaller cart and more profit so yeah) and sometimes because the system might not be able to handle decoding higher quality things and playing video/game/whatever at the same time. I will also note the little DS digital to analogue converter (DAC being what most use instead) is generally not the best thing in the world so there are limited gains to be had here. I am also not sure where you plan to find higher quality versions of some of those but I will assume there is a CD or a port to another device with better quality for this.
Injecting new things is possible but most of the time people alter things rather than seek to increase quality. To increase quality you will also have to alter the relevant portions of the file that note what the sample rate and possibly bit depth if you are changing that are changed to.
http://www.romhacking.net/documents/469 covers some of the SDAT format enough that you can probably get something done here.
You can make swav and STRM files from wave files quite happily (I don't know if tinke,
https://www.romhacking.net/utilities/817/ , is still what many look at, vgmtrans combined with some other stuff might even do it). There are technically things you can do to disassemble and reassemble STRM files but I find them unreliable in putting them back together so tend to go manual here.
There is also the secondary option of getting it to play back another external file
https://github.com/Dirbaio/ASMPatchTemplate being where many start with that.
If it is not SWAV and STRM in the video files then it will probably be AHX/ADX/ADH audio (some of those are ringing a bell for using this but it might be similar games)
https://wikivisually.com/wiki/ADX_(file_format)
There are more tools these days that will convert them. I don't know what goes as far as suggesting ones for it though.
There are further audio formats on the DS, I have a little list
https://gbatemp.net/threads/the-various-audio-formats-of-the-ds.305167/ though it is a bit outdated in some of the info (I think we now have a better description of some of the pokemon ones and I have since done a bit more with some of the cartoon network games and club penguin).
Option crazy man.
If this is to play on a PC emulator then desmume has lua support. You could have it watch to see where video is being played, mute the emulator and instead call the sound to be played from your PC.
http://fceux.com/web/help/LuaFunctionsList.html is for the NES but it is what Desmume bases its lua support on
http://tasvideos.org/LuaScripting.html
http://wiki.desmume.org/index.php?title=Faq#What_is_this_Lua_stuff_I_see.3F
