Think of a disk drive as millions of pigeon holes, each capable of holding a small amount of data (the disk sectors). In order for the operating system to be able to store and later retrieve data on the drive, it needs to maintain an index to say which pigeon hole any particular piece of data is, and in the case of data larger than can be held in one pigeon hole, where the second pigeon hole is, and the third... Often consecutive pigeon holes can be used, but not always (there might not be enough empty pigeon holes in a row before you come to a full one, especially when data has been added and removed over a long period of time).
Writing to the disk means using the index to locate a series of consecutive or non-consecutive empty pigeon holes sufficient to hold the amount of data to be written, filling those pigeon holes (displacing any data that may have been lingering there from before), and updating the index to say what and where that data is. Reading from the disk means using the index to locate which pigeon holes the desired data is being held in, then copying out the content of those pigeon holes. Deleting data means going through the index and marking the pigeon holes that contain that data as "empty" (although they continue to have the data in them until new data is inserted).
The process of formatting a disk clears out the index, so although the pigeon holes might still contain the previous data (unless you do a disk erase) they have lost any link they had with the index - so you no longer have any knowledge of what a file was called or which group of pigeon holes made up the data (and in what order). Imagine going into a library where the books are shelved in any old order, and trying to find any particular book if the note of which shelf it is on has been lost, and then trying to find all the volumes of Encyclopaedia Britannica when every book in the scrambled library looks physically identical and has no title on the spine.
If the content of each individual pigeon hole is examined one by one, it might be possible to deduce what is in it. This is especially true of plain text which can be printed out and pieced together again, but very laborious (don't forget there are millions of pigeon holes, billions on a multi-terabyte disk). If the data is not something recognisable (which the binary data of a video file would not be - a short section of video probably wouldn't even play), the problem is multiplied.
So the bottom line is yes, in theory and with a lot of hard work it would be possible to recover video data from the disk. What you would do is take the whole drive as one long file and try playing it from the first sector and see what you get. If that doesn't work play it from the second sector, then the third... and gradually build up a knowledge of where each .ts file starts and then breaks off and resumes somewhere else. It just depends how much of your time (or money if you want to hand it over to specialists) you are prepared to dedicate to recovering "just telly". Wait for the repeats.
