Document the recording sources
The microphone section becomes the list of sources: inputs, output
devices, system sound. Track layout per source, every source on the
recording's clock, and the smaller mentions across the pages.
Link the manual on qblank.com
Describe encoding inside the process
- Recording: threads, muxer queue, closer, what the encoder is told,
mix and microphone clock, what in-process costs.
- ffmpeg: loading the libraries, the shared build download, staging,
updates within the header major, two frame probe.
- Encoder settings: options as name and value, qp on NVENC, AV1 times four.
- Audio, Building, Screenshots, HDR, Home: follow-ups.
Link the manual; add the About tab
The main window: tray menu drawn by qBlank, Start menu and desktop shortcuts
The main window: the tray icon
Add The main window page
Window size, Shift keeping the shape, borderless and the red taskbar dot
while recording. Linked from Home and Shortcuts.
Freeze and compare: a page for F11, F12 and Shift+F12
What each side of the split drops, where the two passes cut, the
direction submenu and the divider, why deinterlacing stays on, and why
none of it works while recording. Linked from Home.
The program is called qBlank
Every page now says the name the program carries from 4.0 on. What still says
CapView says it deliberately: the shared section names the camera is found
through, the migrator, and the sentences that are about a version which really
was called that.
Updates gains the section the rename needed -- the migrator shipped under the
old file name, the release asked for by repository number and asset label, the
executable that renames itself on its first start, and settings recognised by
their contents rather than by their file name. Home, Building and Configuration
point at it.
Drop the link to the deleted AmaRecTV page
Add a page for people coming from AmaRecTV
AmaRecTV was the answer to this problem for a decade and its last release with a bundled codec is from 2014. The page maps each habit onto where it lives here, spells out the three real differences -- ffmpeg instead of a codec of its own, the standard measured rather than chosen, the composite filter -- and says plainly what CapView does not have. Linked from Home.
Write up the three filters added in 3.3
The composite page gains a section each for following the movement,
restoring bandwidth and applying the chroma blur only where needed, with
the proof that motion compensation cannot remove dot crawl, the measured
response that shows why the bandwidth lift is built from triangles and
not boxes, and a table of what every standard resolves to. The artefact
table gains a column naming the control that handles each one.
Document how the video standard is found, and what every threshold is worth
New page: the two stages (the decoder settles the line count, the picture
settles the colour system), the candidate table, every threshold with the
measurement behind it, the lock times, the starvation rule, the retry
table, and the limits. Includes the round that ran on a 95 % black
picture and why a round now needs 10 % of the frame lit before it is
allowed to decide anything.
Source and signal loses its stale summary -- 250 ms poll, 0.6 s settle,
five candidates -- and links to the new page instead.
Shortcuts gains Shift+F5 and F8, and the enum snippet matches the eleven
actions that actually exist.
Cropping and geometry: the three ways to trigger Detect, and why a crop
is dropped rather than re-measured when the source changes size.
Rewrite the virtual camera page for the DirectShow filter
The page described a Media Foundation frame server that no longer exists. It
now describes what 3.0 actually ships, and why the argument that led to Media
Foundation in 1.9 was wrong: a source filter runs inside the consumer, so the
Global namespace and the privilege that goes with it are simply not needed.
New sections cover the source being passed through 1:1, the continuous ranges
that let a consumer ask for something in between, what OBS, Discord and the
browsers each end up with, the generation-named frame section, the seqlock and
the retry that fixed it, one wake event per consumer, and the consumer table.
Two things are recorded that were measured rather than reasoned about:
- The 8K120 collapse. A single torn read used to end the attempt, which at
large sizes turned ordinary contention into a black picture that never came
back: 94 frames delivered, 0 of them fresh. With three attempts it is 239 of
269.
- The producer half, exercised for the first time. A temporary per-user
registration got past the "installed?" gate and let CapView's own
VirtualCamera feed a real consumer off the SA7160, at native, at 720p30 and
at 640x480@15.
That per-user registration is written up as its own section, because it works:
the camera enumerates through the ordinary system device enumerator with no
administrator anywhere. 3.0 does not ship it, since nothing has shown that OBS
and Discord accept a registration without a FilterData blob, and a camera that
silently fails to appear is a bad thing to learn after release. It is the first
candidate for 3.1.
Building, Home, Updates and Shortcuts follow the same rename from "media
source" to "filter". Shortcuts also loses a wrong claim: the overlay does not
show the camera counters, Settings then Recording does.
Document the second grab point for screenshots
A screenshot used to be one thing: the picture, at source resolution, taken
between the first draw pass and the interface. The Recording tab now has a
switch that moves the grab to the far side of that line, and the page has to
say where the line is, what GrabBackBuffer does with the back buffer, and the
two things that change as a result -- window size instead of source size, and
SDR always, because the scRGB back buffer is refused rather than converted by
untested maths.
The JPEG XR paragraph no longer claims that screenshots never need ffmpeg. AVIF
sits three paragraphs below and needs it, which the same page had already said.
The two page descriptions on Home now match what those pages have grown into.
How the signal detector tells snow from a scene cut
The long version: one page per feature, and what the code does