flac.md : Some more fixes: errors, language.

By request, I have not submitted a full proposal, thus some "Everywhere" changes are not at all (yet) implemented everywhere. 

**Everywhere / general**
 - boldfaced **flac**-the-tool, capitalized FLAC-the-format 
 - italicized *warnings* and *errors*. 
 - What in the text was Warning-but-not-a "*warning*", is now "CAUTION:" (4 instances in this commit). "Note:" is now "NOTE:" Both start on newline.
 - Tried to stick to more consistent phrases about "print" a *warning*, and "exit with" an *error* (when it actually exits), and file already "exists" rather than a mix between "is present" and whatnot.
 - Got rid of a bunch of "specify"/"specifies", when the use was far from a _specification_ as in the FLAC specification or PICTURE specification. I just imagine that someone who text-searches for specification will be more interested in those which are left.

Apart from those, here are some of them explained by referreing to line numbers in *current* ("old") flac.md; also some possible changes that I have **not** taken on board (yet). Ignoring those listed above, and some I think are self-explanatory, and also skipping justification for some that have gotten a thumb up from @ktmf01 at xiph@b359d00
 - 182: Bugfix! Previous change of mine (my blame!) couldn't stick consistent on "Track04" vs Track07 ...
 - also: rewritten. 
 - 205-206: It does not "exit" with an error if processing multiple files. And the "already exists".
 - 222: Tried to improve language.
 - 231: Rewritten.
 - 245: As discussed earlier: Referring to the **Format options** 
 - 273 deviates from a suggestion from @ktmf01 - "print" a *warning* etc.  
 - 282: I didn't like "force" ... cf -f.
 - 298: Longer explanation. 
 - 312: Elaborates. Written to fit \--until.
 - 321: **FIX!** The syntax was wrong. Hopefully it is now right.
 - 322: Admittedly, this is now written in more detail *because* I first got it wrong.
 - 328 as mentioned earlier: "mode" is better reserved for encoding mode, decoding mode, ...
 - 335: Statement is not completely correct, but I put a reservation in the `--apply-replaygain-which-is-not-lossless` text.
 - 340: Another ref to **Format options**
 - 342: Elaborates. With CAUTION:.
 - 344: There are exceptions, at least \--apply-replaygain-which-is-not-lossless. 
 - 366 not yet done: Could need a rewrite. Have it drafted. 
 - 376: Encoding starts here. No need to mention the tukey, but something about format choice ... 
 - Not yet done: I have a draft with level-3 headlines for compression/audio options and for metadata options. 
 - 378: -V moved downwards.
 - 410 to 421: Have been discussed. Mostly, the "specifies" ... and then re-wrote.
 - 438: Reworded. **QUESTION:** Since it did say the wrong comma-separated, should it now outright say semicolon-separated rather than just removing it? 
 - Up to 463: I think these have been discussed.
 - 464-ish: Moved `--ignore-chunk-sizes` here, it is audio ... hopefully.
 - 484: ReplayGain is tags, moved down. Also: I couldn't provoke the warning about extra padding - omitted it, I don't think it is really necessary.
 - 498 to 511: Moved - and are now longer. Please read through. 
 - 516: Do I understand correctly that -T ... --no-utf-8 -T ... will apply no-utf-8 to the latter and not the first tagging?
 - 529: I don't think it has to say anything about UTF-8 when that is said under -T and this is like -T except ...
 - 537: Rewritten ... 
 - 554: -P is rewritten to get it on the format where it first says what it *does*. 
 - 605: Elaborated on `--force-raw-format` after testing. In particular that you should usually omit it when encoding.
 - 610 ff: Clearer what is input/output, rid of "Specifies", make clear that \--bps must be multiple of 8
 - 650: Also to negate defaults ... if that wasn't obvious.
 - 672 ff: I have drafted extensive changes from here on, but for now I am sticking to fewer changes.  
 - 720 ff: 0 to to are now taken from the RFC. 
 - 789 fixed the comma/semicolon error. (I have drafted a full rewrite, but not here). 
 - But: "subdivide_tukey(1/P)" is not implemented ... should be mentioned?
 - 827: Useful place to put the info on where it is maintained?

Author: H2Swine