Bad linebreaks will now cause cues to be skipped (with a warning),
rather than throwing an error.
Closes#2358
Co-authored-by: Álvaro Velad Galván <ladvan91@hotmail.com>
Adds color support for SimpleTextDisplayer and WebVttGenerator (only one
place to fix both now thanks to #4941).
It's limited to the [8 colors
classes](https://w3c.github.io/webvtt/#default-text-color) supported by
the WebVTT specification, and also works with their 3 or 6-digit hex
variants (if the stream has TTML subtitles).
It does not support rgb, rgba or any colors other than these 8.
Fixes#4545
---------
Co-authored-by: Alvaro Velad Galvan <ladvan91@hotmail.com>
When the VTT size setting is used, the horizontal positioning was wrong
in both native and UI display.
The native display is wrong on Chrome and Edge because of a layout bug
in Chrome, where the shadow DOM for the cue box has conflicting
(redundant) styles. For example, these VTT settings:
`line: 85% position: 50% size: 63%`
result in these styles in the shadow DOM:
`top: 85%; left: 18.5%; transform: translate(-18.5%, -85%)`.
The `translate` style is what breaks the positioning. Unfortunately,
there is no way to fix that in JavaScript.
The UI display, however, was buggy for different reasons and is fixable.
The styles `left: 0; top: 0;` were applied by default, and then `top:
85%;` and `width: 63%;` were set based on the cue settings. The default
of `left: 0` was what broke the positioning. Removing this leaves `left`
set implicitly to `auto`, which is correct.
No other test cases were broken (or fixed) by this change.
This also adds a filter parameter to the lab workflow to run a subset of
tests for quicker results. This is useful for updating screenshots.
Closes b/259121343
Many of these are rendered incorrectly in some way.
This PR contains only the tests, but not the screenshots, to make the
tests easier to review. The screenshots will appear in a separate PR.
Related to work on PR #4767
There is currently only one suite, but that is about to change. This
renames all screenshots to include the name of the suite, and updates
the review page with a temporary hard-coded suite name.
Related to work on PR #4767
Also fixes a compiler error that comes up for some reason when renaming
the other tests. The compiler will accept an assignment to nestedCues,
but not pushing to it. I was unable to fix the missing type information
for this.
Related to work on PR #4767
Captions split over stream segment boundaries are repeated over more
than one segment.
Regions in which they sit maybe described the same but allocated
different ids and vice versa.
Proposal to treat internal region ids uniqely by encoding the dimensions
in the id.
fixes#4839
Reinstate previously removed region elements when next caption finds it
is not there: Detect the absence and ensure `updateDOM` is true so its
reinstated and the next caption can be shown.
Closes#4680
Co-authored-by: Joey Parrish <joeyparrish@users.noreply.github.com>
Subtitle timing and offsets were not calculated correctly for HLS live
streams. This issue appears to have affected all v4.x releases up to
v4.0.5, v4.1.5, and v4.2.2.
Both TTML and VTT regions should be thought of as separate elements
from the cue structures they contain. To this end, the UI text
displayer now creates separate region elements to represent CueRegion
objects, and the Cues attached to them nest inside those region
elements in the DOM.
Closes#4381
This changes the `SimpleTextDisplayer` to flatten text payloads
only starting at non-container cues. Not doing so poses an issue with
'rolled over' caption lines.
E.g., consider:
- TTML file `N` with caption line `A` from time `[1.4, 2.1]`.
- TTML file `N+1` with caption lines `A` from time `[1.4, 2.4]` and `B` from time `[2.4, 3.4]`.
If captions `A` and `B` are descendants of two different `<p>` elements, Shaka
currently flattens the text payload from `A` and `B` into a single cue that is
rendered onto the display from time `[1.4, 3.4]`. Therefore, caption `A`'s text
payload is incorrectly showing during time `[2.4, 3.4]`.
The fix updates the `SimpleTextDisplayer#append()` logic to split up individual
non-container cues into their own `VTTCue` object. This will not be done at the
`TTMLTextParser` level to ensure proper style inheritance still occurs for the
(actual) text cues.
Closes#4381
Since the transition to sequence mode for HLS in v4.0.0, VTT cue
timings were broken. This is mainly because VTT cue timing in HLS is
meant to be based on an offset from the media timestamps, and we
generally don't know those now that we use sequence mode.
To fix it, this change uses MediaSource segment mode for the very
first video segment as a way to extract the timestamp, then clears the
buffer, switches to sequence mode, and appends it again. This lets us
get the timing data we need, while avoiding major drawbacks of the
previous HLS implementation:
- We don't need to fetch segments upfront (which is high latency)
- We don't need to fetch segments twice (once for timestamps, and
once again to buffer)
- We don't need to maintain parsers (which were complex and limited
the formats we could support)
Closes#4191
This seems to reduce flakiness in the TextDisplayer tests. The
theory is that Safari throttles timers when the system is very busy,
which it often is during full test runs across multiple browsers. By
not relying on timers and triggering an explicit update in the
TextDisplayer, the tests seem to become more reliable.
This change fixes tests on Chromecast by loading tests later in the process. Test scripts are now dynamically inserted by boot.js, rather than loaded by Karma. The bootstrapping code then awaits the completion of that before starting the Karma frameworks (Jasmine) to run the tests.
This also removes the use of goog.provide/goog.require in tests and test utils. We don't need to load test utils or library sources dynamically in each test, and this gives us more explicit control over script loading and ordering.
Closes#4094
This PR fixes#3242 where for some live streams using segmented VTT, text timings are relative to segment start instead of being absolute.
The PR introduces a new setting: `manifest.segmentRelativeVttTiming: boolean` allowing such alternative timing offset calculation.
The setting is off by default, preserving the current player behaviour.
Co-authored-by: Joey Parrish <joeyparrish@users.noreply.github.com>
When running in sequence mode, we ignore the normal timestamps
of video and audio segments. This lead to problems in some Apple-
encoded webvtt content, which used the X-TIMESTAMP-MAP tag to account
for the timestamp offsets in their video. Thus, those subtitles would
end up 10 seconds offset.
This changes the webvtt parser to ignore the X-TIMESTAMP-MAP when in
sequence mode.
Issue #2337
The recent changes to TTML parsing, to not inherit regions,
inadvertently ended up breaking text alignment in situations
where a region with alignment was on the p or div above a span.
Previously, we only inherited the text and display alignment
from a region on leaf nodes... which was a problem, since we
also didn't apply any styles to text nodes.
Change-Id: I62ac155bc4310a5f7da52c10ca2dd434f8015c97
This changes the TTML parser to not allow cue regions to be inherited
to the children of the element the region was originally assigned on,
except for the purposes of styles (colors, etc).
To allow regions on elements "above" the cues in TTML, such as the
<body> or <div> elements, this also changes the TTML parser to render
the full structure of the TTML file as a tree of cues. The end result
will be a single cue representing the <body>, with children
representing the <div> elements inside it, and those <divs> will have
children that represent the actual cues. Now that our text displayer
can intelligently update child cues as they enter or leave the display
window, this approach should be possible.
Closes#3850Closes#3741
Change-Id: Ia8d750daa06920610c04e9b26e29d2d304eaf8a9
Previously, in the updateCaptions_ method in the UI text displayer,
when trying to determine if a cue should be removed from the container,
we checked to see if it was out of the time range.
This worked in most cases, but it did have the side-effect that the
displayer would not remove the old cues if they were removed manually
via the remove() method. This was most visible if the user changed the
text language; the currently-displayed cue of the old language would
linger until it reached the point it would stop displaying normally.
This changes the UI text displayer to force an update when remove() is
called, if necessary.
Change-Id: I84a4847b67c5fb7597342a943abe13a3cc9e826e
First, the positioning of cues was incorrect. We need to explicitly
set all the position attributes when we position elements with the
"absolute" position.
Second, if we position a text <span> manually, the background will fill
the whole region. So to keep the background wrapping the text, we need
to add another <span> for the text.
Third, the background and font color should not be set on every <span>
element since it won't allow parent cues to set the inherited value.
So this moves the defaults to the top-level text element and allows
parent cue elements to override this. It also would make app CSS
easier to override. Because background color isn't inherited through
CSS, the default is set in JavaScript instead.
Fixes#3521Fixes#3600Closes#3713
Change-Id: I45fc88dcac4a0a062e1474087f24c80d98eef619
A new UI test added in a recent CL involved manually setting the
currentTime value of a videoElement in tests.
For some reason that wasn't working on Safari, so this changes that
UI test to use a fake video instead of a real one.
Closes#3689
Change-Id: I412f677f6cda49bbfc850a9102edfe3b75bca302
Previously, we added support for respecting the time constraints
of nested cues. However, the UI text displayer did not take the
time constraints of nested cues into account when determining when
and how to update the cues.
This changes the UI text displayer to also do that.
Issue #3524
Issue #3643
Change-Id: I6b643f2aa21f367a8e40a8aca2ebb62492c071c2
When trying to detect if it is appropriate to ignore a non-break
element, the ttml parser would base the decision on the presence
of timing info in the cue itself, and on the presence of
textContent within the node. If neither is present, the element
might be ignored.
This works in most cases of nested cues, as the presence of text
within a child node will show up in the textContent of the parent
node. However, not all valid ttml cues have textContent; image
cues, for example, have their contents defined as a DOM attribute.
This changes the ttml parser to not ignore any cue that has child
cues. It also changes the ttml parser to cause cues with no defined
timing to base their timing on their child cues, if present.
This also changes our UI text displayer to respect the timing info
of child cues.
Closes#3643
Change-Id: I9a017f53398bbed8dbeeeebca2cff76dd3666c64
Unfortunately, modern browsers do not support any CSS property that
directly maps to tts:textOutline. However, the non-standard (but still
commonly implemented) -webkit-text-stroke-width and
-webkit-text-stroke-color properties, when used together, can
replicate the functionality of tts:textOutline, with the exception of
the optional blur radius.
Closes#3612
Change-Id: I863d09fc447765646bd405c59b0b20960362a594
When an external subtitle track is added and you are using DAI, the external track does not take into account the ads that the video has, so this PR makes this internally take into account when generating the external track.
This uses flexbox once again to get proper positioning of cues. To
compensate for the issues that originally made us remove flexbox, this
adds a wrapper span inside the flexbox element.
Summary of screenshot changes:
- slight change to background sizing
- ui-basic-cue
- ui-cue-with-controls
- ui-duplicate-cues
- ui-end-time-edge-case
- ui-flat-cue-bg
- ui-two-basic-cues
- background fills block with literal newline in text
- ui-cue-with-newline
- region anchored without padding
- ui-region-position
- new screenshots
- *-nested-cues-with-linebreak
- *-region-with-display-alignment (regression test for this issue)
Closes#3379
Change-Id: I8c678721d96662e0f8940cda12df4f5b5e5baf1e
Old versions of Chromium, which may be found in certain versions of
Tizen and WebOS, may require the prefixed version of "writingMode":
"webkitWritingMode".
However, testing shows that Tizen 3, at least, has a "writingMode"
property, but the setter for it does nothing.
Therefore, instead of just detecting the existence of the property, we
need to additionally check if setting it had any effect. If it is
either missing or non-functional, then we fall back to the prefixed
version of the property. This fixes the functionality on Tizen 3.
We also need to change the conditions we check for in the tests, since
property existence is not enough to set the correct test expectation.
See also PR #3330
Change-Id: Ic906f3c5af956b5edd1788e95e1978eb4b3098ac
TTML allows "smpte:backgroundImage" on both the div level and p level. Adding parsing the background images into cues on the div level.
Closes: #3097Closes#3248
Álvaro Velad Galván
Tom Bloom
Albin Larsson
Joey Parrish
David HM Morgan
Julian Domingo
Álvaro Velad Galván
Philippe
theodab
Jacob Trimble
bcupac
Joey Parrish
Michelle Zhuo