Due to crashes being caused by hardware acceleration in the media source
on mac, disable hardware acceleration of the media source being on by
default for the time being.

In transitions, because the 'to' and 'from' are always rendered to
textures, the end result will always have premultiplied alpha.  This
would cause alpha to have blackish edges during transition, and cause
semi-transparent images to appear darker than they were supposed to.

To replicate, simply set the transparency of a source to 50%, then
transition between that scene and another scene.  The source will appear
to "pop" in and out unnaturally due to the premultiplied alpha effect of
the render targets.

To fix this, the solution is to simply convert premultiplied alpha to
straight alpha in the transition pixel shaders.

(Note: This commit also modifies mac-capture and linux-pulseaudio)

This prevents outputs from being monitored, preventing a potential
feedback loop.

Twitch added the following ingests:
- South America: Lima, Peru
- EU: Lisbon, Portugal
- South America: Medellin, Columbia
- EU: Milan, Italy
- NA: Mexico City

The following ingest server is no longer listed in hitbox'  API:
- US-East: Washington

YouTube does not list a requirement that states that only the main profile can be used.

At request of c3r1c3 to be a bit more specific.

static should come before const.

Closes jp9000/obs-studio#778

Closes jp9000/obs-studio#750

Closes jp9000/obs-studio#779

EU: Oslo
South Korea: Seoul
India: Chennai

This allows for color grading based on a color look-up table. Commonly
used to color correct and/or grade video that has been shot in Log.

Closes jp9000/obs-studio#747

Adds an option to the aspect/scale filter to undistort the center of the
image when scaling from ultrawide to wide.

Closes jp9000/obs-studio#762

500ms is a bit unnecessarily low.

Update video ingest servers for Vaughn Live.  Also split Breakers.TV
into its own since it is possibly becoming its own entity soon.

Closes jp9000/obs-studio#771
Closes jp9000/obs-studio#772

This reverts commit f4d7f0cb.

Never thought a revert of a revert would happen.

Fixes a hang on startup.

Due to the fact that people follow terrible guides that ignorantly
recommend opencl, the "opencl" custom x264 option was blocked.  However,
because some people still apparently want to use it, it's being renamed
to "opencl_is_experimental_and_potentially_unstable" so that anyone who
uses this option realizes that they should probably think twice about
using the option rather than just follow some terrible guide that tells
them to use it.

When rehooking, even if scaling is off, it'll unintentionally overwrite
the game's backbuffer texture size with the scale size.

The Saturation option was unbalanced, giving more weight to the color
red and a TON more weight to the color blue.  This patch balances the
colors to provide a more even increase in saturation across the RGB
channels.

Closes jp9000/obs-studio#739

When using this filter in/on OpenGL situations can cause crashes with an
"int to vec4" cast.  This fix should resolve that issue.

Closes jp9000/obs-studio#739

This reverts commit bd67c18c.

Causes performance issues when the window isn't found.  The situation it
fixes is also very rare anyway.

Using and creating a window can use issues in game capture if multiple
game captures are active, so revert back to using a mutex, and just
ignore the keepalive check failure if injected inside a UWP program
(only check to see if GetLastError reports that it's not found -- if it
returns access denied or any other error, assume it's in a UWP program,
and ignore the keepalive check).

Silently ignores opencl parameter for x264 to protect users
following bad youtube guides.

Hopefully this will fix a crash from CEF on 10.9

There is a build error on freebsd:
    error: invalid application of 'sizeof' to an incomplete type
    'struct sockaddr_in'

Include proper header file to solve this.

Ref:
https://www.freebsd.org/doc/en/books/developers-handbook/sockets-essential-functions.html

YouTube now recommends a max video bitrate of 51.000 Kbps.

These aren't referenced encoder packets, so they must be freed manually.

There's no need to keep a reference to the backbuffers.  That and the
backbuffer references weren't being released anyway, so that fixes that
bug.

Executing the dxgi 1.4 code causes it to cycle through backbuffers,
backbuffers which may not exist if in discard mode.

It was undesirable to pass arrays via function parameters, so a
structure should be used instead.

Also increases total allowable backbuffers to 8.

If the backbuffer count is larger than 3, it could still try to assign
backbuffers to pointers beyond the variable's array size when calling
swap->GetBuffer.

Use a d3d9 device and allocator to encode in QSV.

This fixes a random crash that could only happen on Windows 7. The QSV
Deviced returned a DEVICE_FAILURE after a random amount of time with the
old method.

This fix is totally based on Shinck's QSVHelper.exe patch for OBS
Classic (see
https://obsproject.com/forum/threads/0-633b-qsvhelper-exe-was-killed-encode-failed.19230/page-3#post-161984
for more information)

This is more like a proof of concept, but that fix is currently stable
and tested more than 50 hours, with a single session of +14 hours.

That commit doesn't respect all OBS Guidelines. It is currently
recommended to wait for a more "cleaner" implementation.