LoadCultures() in LocalizationManager skipped all iso6392.txt entries
without a two-letter ISO 639-1 code, dropping 302 of 496 languages
including mul (Multiple languages), und (Undetermined), mis (Uncoded
languages), zxx, and many real languages like Achinese, Akkadian, etc.
This caused FindLanguageInfo() to return null for these codes, which
meant:
- ExternalPathParser could not recognize them as valid language codes
in subtitle filenames, so the Language field was never set
- DisplayTitle fell back to the raw code string (e.g. "Mul")
Fix by allowing entries without two-letter codes to be loaded with an
empty TwoLetterISOLanguageName. Also set LocalizedLanguage in
ProbeResultNormalizer for ffprobe-detected streams (the DB repository
path was already handled on master).
Wire up EnableSubtitleExtraction config to MediaEncoder.CanExtractSubtitles
so the setting is actually respected. Gate subtitle extraction check behind
PlayMethod.Transcode since DirectPlay has no competing ffmpeg process.
Add parameterized tests for StreamBuilder.GetSubtitleProfile covering
text and graphical codecs, profile format matching, and extraction
setting behavior. Remove misplaced SubtitleEncoder extraction test.
- Add missing param and returns XML doc tags (SA1611, SA1615)
- Remove trailing alignment whitespace in test attributes (SA1025)
- Use nullable string parameter for null test case (xUnit1012)
Encoders sometimes produce sample aspect ratios like 3201:3200
(0.03% off square) for content that has effectively square pixels.
The exact string comparison against "1:1" marks these as anamorphic,
which triggers unnecessary transcoding on clients that require
non-anamorphic video.
Parse the SAR ratio numerically and treat values within 1% of 1:1
as square pixels. This threshold is well clear of the nearest real
anamorphic SAR (PAL 4:3 at 16:15 = 6.67% off).
Fix video lacking SAR and DAR are marked as anamorphic
Original-merge: 9470439cfa1eaf7cb9717f16031b020cedab516a
Merged-by: crobibero <cody@robibe.ro>
Backported-by: Bond_009 <bond.009@outlook.com>
Fix font extraction for certain transcoding settings
Original-merge: ee34c75386cc1a0ca0e15196a43b685fa0e73130
Merged-by: crobibero <cody@robibe.ro>
Backported-by: Bond_009 <bond.009@outlook.com>
Make priority class setting more robust
Original-merge: 70c32a26fa9f16db513a92cdd2dcafa7ee15a80d
Merged-by: crobibero <cody@robibe.ro>
Backported-by: Bond_009 <bond.009@outlook.com>
Follow TrickPlay example of running other background ffmpeg tasks as ProcessPriorityClass.BelowNormal:
- Keyframe extraction
- Media info probing during library scans
- Audio normalization
- Image extraction
Co-authored-by: Evan <evan@MacBook-Pro.local>
Follow TrickPlay example of running other background ffmpeg tasks as ProcessPriorityClass.BelowNormal:
- Keyframe extraction
- Media info probing during library scans
- Audio normalization
- Image extraction
Keyframe-only trickplay image extraction can fail for some media
files. The current behavior is to skip the media file and try again
on the next run, which will fail again.
This adds a fallback to regular non-keyframe-only extraction for
failed runs, so the extraction can complete.
