Unicode Supplementary Multilingual Plane/Supplementary Ideographic Plane Support

[GoogleCodeExporter]

What steps will reproduce the problem?

1. Have a subtitle line with a unicode code-point of 10000 or higher present. 
Commented or uncommented.
2. Have a following line with a unicode character of 100 to FFFF (Ascii seem to 
be fine).

See attached script. Putting checks in place to prevent this behaviour if 
support is too difficult to add would be much appreciated.

What is the expected output? What do you see instead?

Broken rendering of non BMP Unicode glyphs.

What version of the product are you using? On what operating system?

211, Windows 7.

Font with the Musical Character SMP range is included for correct rendering. 
This actually works fine with LibASS on Aegisub in Windows.

Original issue reported on code.google.com by i.am.the...@gmail.com on 25 Apr 2013 at 4:41

Attachments:

• broken_unicode_smp_rendering.txt
• Musica306.zip

[GoogleCodeExporter]

This is already supported (somewhat).

VSFilter 2.39, MPC-HC VSFilter, and xy-VSFilter all render the attached MKV 
sample correctly in DirectShow.

VSFilter implementations only seem to have issues with this when CSRI is used, 
or when loaded as an external script.

We'll look into this at some point. For the time being, I can only recommended 
muxing such scripts into MKV as a workaround.

Original comment by cyber.sp...@gmail.com on 26 Apr 2013 at 1:02

• Changed state: Accepted
• Added labels: Priority-Low
• Removed labels: Priority-Medium

Attachments:

• musica_test.mkv

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

In our next release, both lines in your example should render correctly in a 
UTF-8 script. We've already fixed both bug listed below.

1) The UTF-8 auto-detection code we merged from MPC-HC was designed to abort 
and fallback to ASCII when it found an invalid Unicode character. This was 
recently fixed in MPC-HC earlier this month to not fallback to ASCII when the 
script was signed UTF-8.

2) All VSFilter implementations incorrectly abort on valid 3-byte Unicode 
characters in UTF-8. Only when the script was saved as UTF-16 would these 
character render correctly.

Original comment by cyber.sp...@gmail.com on 28 Apr 2013 at 7:59