Boxes appear in place of certain symbols in .ass subtitles

[ibot]

I use MPC-HC, with Haali, madvr, LAV audio and video, and xy-vsfilter. All are up to date. Yet, in certain videos, I see boxes or boxes with question marks in them in place of certain symbols. I extracted the sub stream and the symbols in question are ☆ and ❤.

How can I get MPC to recognize these symbols? What is the problem here?

I've attached a picture that illustrates my issue. Here's a link to the same image, if that's what is preferred. These are two different videos, by the way.

[Baldrick]

Does it work in VLC? (it has built in subtitle decoders)

[ibot]

Does it work in VLC? (it has built in subtitle decoders)

Wow, I hadn't tried that before. But, yes, it looks like those symbols both show up in VLC.

Do you know what might be the problem with my current MPC setup?

[AlanHK]

Do you know what might be the problem with my current MPC setup?

Not using the right font.
Look at the subtitle settings and see if any allow you to choose a font.
Or possibly the font encoding does not match. Again, hopefully there should be a setting to fix that.

[ibot]

Not using the right font.
Look at the subtitle settings and see if any allow you to choose a font.
Or possibly the font encoding does not match. Again, hopefully there should be a setting to fix that.

I checked xy-vsfilter's settings, and it does give me options regarding the font, but any changes to those settings don't affect .ass subtitles. It's the same exact thing with MPC's settings. I think all of those settings apply to .srt subtitles.

I know the problem does not lie with the file, because VLC was able to display the subtitles flawlessly, more or less.

EDIT: Alright, so I uninstalled xy-vsfilter and switched to MPC's built-in subtitle renderer, and it still showed up as a box. Also, I installed DirectVobSub and switched to it, and it gave no different results. Then, I reinstalled xy-vsfilter, and, still, nothing changed.

[ibot]

Here's a screenshot of the options in question. Should any of this be changed?

[Selur]

Changing or enforcing another font won't help since not all fonts contain all symbols.
e.g. russian, chinese, japanese characters,...

It might help to look into the .ass file to see which fonts are used and install the necessary fonts,...
Also if the .ass file was extracted from an mkv, the mkv might also contain the fonts (not all mkv splitter support the loading of fonts that are present in the mkv container)

[ibot]

The heart symbol is styled in the Georgia font.

Here's an interesting thing I noticed: when I paste the heart symbol into Microsoft Word, it won't allow me to change its font; instead, it forces me to use a font called MS Mincho for it. The thing is, I already have this font installed on my computer, under the name "MS Mincho & MS PMincho."

Here's a screen shot of what it looks like in VLC, and when the .ass file is open in Aegisub (I'm not sure why I took this one, but oh well, it can't hurt):

[AlanHK]

double post

[AlanHK]

Also if the .ass file was extracted from an mkv, the mkv might also contain the fonts (not all mkv splitter support the loading of fonts that are present in the mkv container)

Use MKVExtractGUI2 and see if there are any fonts in the MKV, if so extract them and you can install them in Windows, then they will be available for all programs.

Also, look in the ASS file with a text editor (eg Notepad). You should find all the fonts required defined near the top, e.g.:

[Script Info]
; Script generated by Aegisub v1.10 
; http://www.aegisub.net
Title: Default Aegisub file
ScriptType: v4.00+
PlayResX: 704
PlayResY: 480
Video Aspect Ratio: 0
Video Zoom: 6
Video Position: 0
Last Style Storage: zip
WrapStyle: 0

[V4+ Styles]
Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
Style: style1,Humanst521 BT,20,&H00FFFFFF,&H0000FFFF,&H00000000,&H64000000,0,0,0,0,100,100,0,0,1,2,2,2,10,10,10,0
Style: op,Essai,24,&H00E9E9E9,&H0000FFFF,&H64000000,&H96000000,0,0,0,0,100,100,0,0,1,2,0,2,10,10,10,0
Style: main,Napapiiri,28,&H00D3C4C6,&H0000FFFF,&H32000000,&H96000000,-1,0,0,0,80,100,0,0,1,1.5,2,2,10,10,10,0
Style: ed,TimeScrDMed,24,&H00E0CBCD,&H00EED3D6,&H32000000,&HC8000000,0,0,0,0,100,110,0,0,1,2,2,1,10,10,10,0
Style: Default,Arial,20,&H00FFFFFF,&H0000FFFF,&H00000000,&H00000000,0,0,0,0,100,100,0,0,1,2,2,2,10,10,10,0

The fonts here are the second parameter in each style listed:
Humanst521 BT
Essai
Napapiiri
TimeScrDMed
Arial

If your system has a font of the same name, but a different version that doesn't have the same characters, it might be the problem. Replace it with the one extracted from the MKV in that case.

Also, the last parameter in the style is "encoding". I think "0" means the system default, but that may not be what you need. I'd need to look up the spec to see the other options. I've changed that previously to make punctuation display correctly.

[ibot]

Alright, so I extracted all fonts and installed all of them, and it solved one of the issues! The heart symbol still doesn't show up, but the star symbol does. Thanks!

Now, what would the problem be with the heart symbol? What makes it different? I find this really strange.

[AlanHK]

You can see in your ASS that the heart is in the "op trans" style. What is the definition of that?

[ibot]

You can see in your ASS that the heart is in the "op trans" style. What is the definition of that?

Georgia font. Here's a screenshot:

[Selur]

a. using another splitter might help if the needed font is embedded (to extracting and installing fonts is normally not necessary); since you mentioned Haali, you did check that a. in MPC-HC the internal mkv support is disabled (otherwise Haali will not be used) and b. that the configuration of Haalis Media Splitter allows the lookup of embedded fonts, didn't you?
b. strange thing is, when I look at (font map using mapOchar) the Georgia font in my Windows/Fonts folder, it does not seem to contain a heart -> seems like the styling is messed up.

[ibot]

a. using another splitter might help if the needed font is embedded (to extracting and installing fonts is normally not necessary)

Yeah, I tried MKV cleaver and MKVextract, and both gave me the same results. I do believe I've extracted every possible font.

b. strange thing is, when I look at (font map using mapOchar) the Georgia font in my Windows/Fonts folder, it does not seem to contain a heart -> seems like the styling is messed up.

It's true that Georgia font doesn't contain a heart. But, then, why would VLC be able to play it? This is such an enigma.

By the way, thank you guys so much for your help so far. I really appreciate the amount of thought and effort you've invested in this.

[Selur]

But, then, why would VLC be able to play it?

VLC is fontconfig based meaning it does use all the fonts fontconfig finds instead of just relying on the system fonts. I guess is that you got the required font (or a font that is marked as compatible to the font family) somewhere on your system and fontconfig finds that font and so VLC can use it. (btw. if I'm right about fontconfig MPlayer-based media players like SMPlayer should also be able to playback the file correctly,.. )

Cu Selur

[ibot]

VLC is fontconfig based meaning it does use all the fonts fontconfig finds instead of just relying on the system fonts. I guess is that you got the required font (or a font that is marked as compatible to the font family) somewhere on your system and fontconfig finds that font and so VLC can use it. (btw. if I'm right about fontconfig MPlayer-based media players like SMPlayer should also be able to playback the file correctly,.. )

Cu Selur

You're right! I downloaded SMPlayer, and it displays the heart, just like VLC.

So...now what? Is there any solution to this, now that we know the problem?

[Selur]

to make clear what the problem is:
the styling points to wrong front, VLC&SMPlayer only can playback the file 'properly' since they find a to the font-family compatible font, that contains the element you are looking for.
So there are two options (which spring to mind):
a. change the styling in a way that it points to a font that really does contain the character you are looking for (this is probably the easier option)
b. read up on how to figure out which folders fontconfig scans and check each of them for the font you want If fontconfig is installed on the system, atm. it probably only comes as a dll with SMPlayer/VLC, one should be able to trigger it to give out a list of all indexed fonts. (haven't every tried that on Windows )

Cu Selur

[ibot]

to make clear what the problem is:
the styling points to wrong front, VLC&SMPlayer only can playback the file 'properly' since they find a to the font-family compatible font, that contains the element you are looking for.
So there are two options (which spring to mind):
a. change the styling in a way that it points to a font that really does contain the character you are looking for (this is probably the easier option)
b. read up on how to figure out which folders fontconfig scans and check each of them for the font you want If fontconfig is installed on the system, atm. it probably only comes as a dll with SMPlayer/VLC, one should be able to trigger it to give out a list of all indexed fonts. (haven't every tried that on Windows )

Cu Selur

Alright, sounds good.

[AlanHK]

The heart symbol is being pulled from a symbol font, not Georgia.
Some apps can work that out, others not.
In Aegisub, can you select that single character and choose the symbol font for it explicitly?
If that works, you can search and replace the coding for that in the ASS and it should work more reliably.

[ibot]

Yeah, it looks like that's my only option, practically. Thanks for your help!