skip-logo.lua: fix skipping in the first two frames

mpv typically decodes and filters at least 2 frames before starting
playback. This happens during seeks, as well as when playback starts
from the beginning of the file.

skip-logo.lua receives notifications for all filtered frames, even
during seeking. It should interrupt during seeking, so as a crude
heuristic, it ignored all frames while the player was seeking. This does
not mean all these frames are skipped due to seeking (thus it's a "crude
hueristic"). In particular, it means that the first 2 frames of a video
cannot be skipped, since they're filtered within the playback restart
phase (equivalent to "seeking").

Fix this by making the heuristic slightly less crude. Since we observe
the property as "none", the property is not actually read until we do it
explicitly. By not reading it during seeking, we can let the frames
internally queue up (vf_fingerprint discards them in a ringbuffer-like
fashion if they're too many). Then, if seeking ends, we get the current
playback timestamp, and check queued up frames that are at or after that
timestamp. (In some ways, this duplicates what the player's seeking
logic does.)

A disadvantage is that this is racy. While playback-time is guaranteed
to be set when seeking changes from false to true, playback could
already have progressed to the next frame (or more) before the script
gets time to react. In theory, we could add a seek restart hook or so,
but I don't want to. A property that returns the last playback restart
time would also do it, but feels to special. Not an important problem
in practice anyway.

1 files changed, 36 insertions, 15 deletions

diff --git a/TOOLS/lua/skip-logo.lua b/TOOLS/lua/skip-logo.lua
index 2c624a04e8..34cbff06f2 100644
--- a/TOOLS/lua/skip-logo.lua
+++ b/TOOLS/lua/skip-logo.lua
@@ -79,6 +79,8 @@ local meta_property = string.format("vf-metadata/%s", label)
 local config = {}
 local cases = {}
 local cur_bmp
+local seeking = false
+local playback_start_pts = nil
 
 -- Convert a  hex string to an array. Convert each byte to a [0,1] float by
 -- interpreting it as normalized uint8_t.
@@ -194,24 +196,12 @@ local function check_fingerprint(hex, pts)
     return false
 end
 
-mp.observe_property(meta_property, "none", function()
+local function read_frames()
     local result = mp.get_property_native(meta_property)
     if result == nil then
         return
     end
 
-    -- Disable matching while seeking. This is not always ideal. For example,
-    -- the filter chain may filter frames ahead of where it will resume
-    -- playback (if something prefetches frames). On the other hand, the
-    -- skipping logic shouldn't activate when the user is trying to seek past
-    -- the skip frame anyway. You could be more fancy and here, and store all
-    -- seen frames, then apply the skipping when it's actually displayed (by
-    -- observing the playback time). But for now, the naive and not-always-
-    -- correct way seems to suffice.
-    if mp.get_property_bool("seeking", false) then
-        return
-    end
-
     -- Try to get all entries. Out of laziness, assume that there are at most
     -- 100 entries. (In fact, vf_fingerprint limits it to 10.)
     for i = 0, 99 do
@@ -223,10 +213,41 @@ mp.observe_property(meta_property, "none", function()
             break
         end
 
-        if check_fingerprint(hex, pts) then
-            break
+        local skip = false -- blame Lua for not having "continue" or "goto", not me
+
+        -- If seeking just stopped, there will be frames before the seek target,
+        -- ignore them by checking the timestamps.
+        if playback_start_pts ~= nil then
+            if pts >= playback_start_pts then
+                playback_start_pts = nil -- just for robustness
+            else
+                skip = true
+            end
+        end
+
+        if not skip then
+            if check_fingerprint(hex, pts) then
+                break
+            end
         end
     end
+end
+
+mp.observe_property(meta_property, "none", function()
+    -- Ignore frames that are decoded/filtered during seeking.
+    if seeking then
+        return
+    end
+
+    read_frames()
+end)
+
+mp.observe_property("seeking", "bool", function(name, val)
+    seeking = val
+    if seeking == false then
+        playback_start_pts = mp.get_property_number("playback-time")
+        read_frames()
+    end
 end)
 
 local filters = mp.get_property_native("option-info/vf/choices", {})