embed spanish subs... need some help here

[afteryouwho]

Hi all.
I am trying to embed subs in Spanish on an avi file but am having a problem I can't see how to fix.
I have the movie file in .avi and a .srt file in Spanish. When I embed the srt in the avi file, it works fine but all the Spanish words with an accent or the ñ come out with the accented letter or ñ blanked out.
I know I am missing something here in the configuration of the recording but can't see what.
Anyone out there to throw me a line?
Thanks.

[hech54]

I'd imagine it's the same as German umlauts. I don't recall ever seeing umlauts in subtitles. Schröder is changed to Schroeder, etc etc etc.

[afteryouwho]

Thanks for the reply.

After trying a hundred options, I got things to sort of work by using the text encoding "maccentraleurope". I was able to get all the accents in but the ñ still comes up empty. You might get the umlauts to work that way, too.

[hech54]

My wife tells me that she HAS seen umlauts in subtitles....but not very often. In my opinion the decision to replace umlauts with "AE" or "OE" is done ahead of time to ensure that something is shown on the screen of everyone's DVD player.

[afteryouwho]

Not being a German speaker, I have no idea about this.

I just hope someone sees this thread and can shed some light on my problem.

[Case]

To get proper display of characters, you need two things: the right text encoding and a matching font.

• ISO Latin 1 (ISO 8859-1) has defined a text encoding of most accented characters for western european languages, including all spanish glyphs like ñ, á, é, í, ó, ú, Á, É, Ú, Í, Ó, Ñ, ¿, ¡, ª, º.
• ISO 8859-15 is a update/variation on ISO 8859-1, to include these symbols: €, Š, š, Ž, ž, Œ, œ, Ÿ (at the cost of missing: ¤, ¦, ¨, ´, ¸, ¼, ½, ¾).
• Unicode is a character superset, containing most of the worlds characters, with about 100,000 glyphs. The most common text encoding with Unicode is UTF-8.

The second requisite is a font that has these glyphs. If the font that is used is missing these glyphs, then the subtitle rendering engine won't be able to show some characters. This may show as a space character, as an underscore, as a question mark, or as just nothing -- depending on programmed error handling.

A spanish .srt subtitle file found on a subtitle site probably has text encoding ISO 8859-1 or (Windows) CP-1252, which are very similar, and often interchangeable.
The .srt file is just a text file without meta data, like which text encoding is used. So most of the times you have to make assumptions based on the language.

If you edit your own subtile file, make sure to at least know which encoding is used when saving, and prefer Western (ISO Latin1) or Western (Windows Latin 1) or Unicode (UTF-8).

To use subtitles in ffmpegX, you are limited to the provided fonts. Some of those fonts have a limited character set. However, the default Arial has all the characters for spanish.
You also need to specify which text encoding is used for the file. A mismatch may result in some characters not being displayed, as character 241 is not always mapped to ñ.

[Case]

Sample rendering of spanish .srt file where the text encoding matches between the options of saving the file and the subtitle settings in ffmpegX:


The fact the the text encodings match is more important than which text encoding you choose.

[afteryouwho]

Hi Case.
Thanks for the very clear explanation of what I did wrong:

I used TextEdit to redo the subs but did not realize that it was set to the American Typewriter font which, although it shows up on my computer, is not compatible with ffmpegx rendering as far as the ñ is concerned.

I mention this in case someone else has the same problem.

And thanks again for your help.

[afteryouwho]

This is an annex to my former reply. Using Case's great explanation, I changed the font in my .srt file to Arial thinking that was the problem but it was not.
Using text encoding ISO 8859-1 and ISO 8850-15 ffmpegX is still unable to recognize accents, ¿ and ñ when it embeds the sub-titles. If I once again use Maccentraleurope text encoding I get the accented letters but no ñ or ¿.

I am definitely missing something here because the example image Case includes in this thread shows that it is possible to get the process right.

To get proper display of characters, you need two things: the right text encoding and a matching font.

If you edit your own subtile file, make sure to at least know which encoding is used when saving, and prefer Western (ISO Latin1) or Western (Windows Latin 1) or Unicode (UTF-8).

To use subtitles in ffmpegX, you are limited to the provided fonts. Some of those fonts have a limited character set. However, the default Arial has all the characters for spanish.
You also need to specify which text encoding is used for the file. A mismatch may result in some characters not being displayed, as character 241 is not always mapped to ñ.

I have tried to see what encoding TextEdit is using when I save the file but am also unable to see how to do that. Or is there somewhere in ffmpegX that let's me set the default text to Arial?

Thanks again for all the time and trouble.

[Case]

I'm not exactly sure where it goes wrong.

About text encoding. Think of your subtitle file as a bunch of ones and zeros, or a bunch of hexadecimal numbers, to make it a bit more organized. The numbers gets familiar meaning for the ASCII values (character 32 is a space, character 41 is a capital 'A', character 61 is an lower case 'a'. Beyond ASCII (not-accented letters) things get uncertain, as it depends on what "rules" you choose to follow. Such a rule would be the text encoding. Think of it as a pre-defined set of mapping numbers to glyphs. (This "encoding" is nothing like video encoding or file encryption, despite using the same word.)

TextEdit has two basic modes: plain text and rich text. Plain text doesn't allow any formatting, like selecting a font, while rich text does, including fonts, color, sizes, paragraph alignment, etc. You can switch between the two modes using Cmd-Shift-T, or its menu equivalent in the Format menu.
You should use plain text for (.srt) subtitles. This defaults to what looks like the Monaco monospaced font in TextEdit, although that is a program setting, not specified in the file. Don't worry about the look of the font, only character completeness counts now.
Edit as you please.
Save the file. The Save As dialog box should mention "Plain text encoding", with a drop down list to choose from, including a customize option to edit the list (to your needs). I changed my list a couple of times, so I forget what the default list is.
If you used characters that are not available in a certain text encoding, then those encodings are grayed out.
If you're not sharing this file with others, then choose what you please, as long as that text encoding is available in the other apps of your workflow.
If you are sharing this file, you may want to use a text encoding that is expected.
As it is hard to tell which encoding is used in a plain text file, I often append that to the file name for reference. Like so: some_movie.es.isolatin1.srt

Does this help a bit?

[afteryouwho]

Hi Case:

Thanks for taking all this time with me. I think I have finally got a hold on things.

I saved my TextEdit .srt file as MacRoman and then in ffmpegx put the encoding in the same thing and it all came out.
From all your helpful advice I figured out that it is important that the two encoding coincide.

Thanks again.