Yes, only 0x27 is in the lower range of ASCII characters in the browsers code set. Everything in higher range has to be converted. To over-simplify a bit, each language has its own set of characters. The positions of the “standard” characters in the lower 128 ASCII range (95 printable characters that include digits 0-9, upper- and lower-case A-Z, the basic punctuation marks, etc.) is fixed, but characters with diacritical marks (accents, umlauts, etc.), and “fancy” stuff like curly apostrophes and quotes that are more typographically appealing, end up in the high range, and that high range can map differently for other languages.
Since each language’s character set can be different (the character at ordinal position 196 in one set is different from that in the same position in another), correctly displaying a string written in one character set could render the characters incorrectly in another, so in the attempt to resolve this, UTF encoding converts those characters to a standard mapping from a large, unified set of characters for all languages. Then, when a UTF-encoded string needs to be displayed in a different language, those characters are decoded and mapped back to where they are in that language’s character set.
All of this is primarily a function of support for non-Roman languages, like Greek, Cyrillic, Arabic, Thai, Japanese, Chinese, etc. I remember when Windows made this shift. It was a lot to wrap your head around, and many developers felt that we had done fine with basic 95 characters in 7-bit ASCII (the lower 128) up to that point, why change? Oh, maybe because most of the population of the planet uses non-Roman character sets. But simple things got harder, for sure. It was no longer possible to directly scan bytes of a stored string to find a substring, for example, you had to take extra steps, notably matching your encodings, and now searches for one character have to be multi-byte. It also introduces the possibility that characters may look very much the same, and be used interchangeably, like the lower ASCII 0x27 apostrophe and your curly apostrophe, but searching for one doesn’t match the other, in spite of their visual similarity and use. More work to fix that.
As for this forum (SMF), it apparently doesn’t support UTF encoding, although the browser will happily jam it down SMF’s throat by default.