Bug in Media Foundation MPEG-4 File Source related to timestamping video frames of a fragmented MP4 file

Some recent update in Media Foundation platform introduced a new bug related to fragmented MP4 files and H.264 video. The bug shows up consistently with file versions:

  • mfplat.dll – 10.0.14393.351 (rs1_release_inmarket.161014-1755)    15-Oct-16 05:48
  • mfmp4srcsnk.dll – 10.0.14393.351 (rs1_release_inmarket.161014-1755)    15-Oct-16 05:45

The nature of the problem is that MPEG-4 File Source is incorrectly time stamping the data: frame time stamps are incorrect, they seems to be getting wrong durations and increments, then quickly jumps into future… and on playback this leads to unobvious playback freezes. As Media Foundation is used by Windows Media Player, Windows 10 Movies & TV Player, the bug is present there as well.

The original report is on MSDN Forums.

Presumably it is possible to roll certain Windows Update package back, or alternatively one has to wait for Microsoft to fix the problem and deliver a new update deploying the fix.

Applying Hardsubs to H.264 Video with GPU

Video adapters currently offer a range of services which enables transcoding of H.264 content with certain modifications (including but not limited to flexible overlays, scaling, mirroring, effects and filters) end-to end on GPU keeping data as DirectX Graphics Infrastructure resource at all processing stages.

Such specialized processing capabilities are pretty powerful compared to traditional CPU processing, especially taking into consideration performance of low end low power-consumption systems still equipped with contemporary GPU.

For a test, I transcoded video H.264/MP4 files of different resolutions applying a text overlay having a time stamp of video frame being processed. The overlay is complex enough to be  varying frame to frame, be a standard font with respective rasterization (using DirectWrite). The test re-encoded H.264 content maintaining bitrate without giving too much care for other encoding details (defaults used).

Hardsub Performance Test

The roughly made test was successful with two video GPUs:

  1. Intel HD Graphics 4600 (7th gen; Desktop system; Core i7-4790 CPU)
  2. Intel HD Graphics (8th gen; Ultramobile system; Atom x5-Z8300 CPU) – the system is actually a $200 worth budget Chinese tablet Cube iWork 10 Ultimate

The test failed on other GPUs:

  • Intel HD Graphics 4000 (7th gen; Mobile system; Core i7-3517U CPU)
  • NVIDIA GeForce GTX 750

The problem – as it looks without getting into details – seems to be the inability of Media Foundation APIs to fit Direct3D-enabled pipelines out of the box, such as because of lack of certain conversion. It looks like transcoding can be achieved, with just putting some more effort into it.

As of now, Intel offers their 9th generation GPUs and the ones being tested are hardware of a few yeas in age…

Compared to real time performance of 100% (meaning that it takes one second to process one second of video of given metrics), both systems managed to do the transcoding relatively efficiently. With a roughly built test having a bottleneck at applying overlay, taking place serially in single thread, both systems showed performance sufficient to convert 1920×1080@60 video faster than in real time and without maxing CPU out.

Intel’s seventh generation desktop GPU managed to do the job way much faster.

It is interesting that even cheap tablet can process a Full HD video stream loading CPU less than 40%. Basically, the performance is sufficient for doing real time video processing (including using external web camera like Logitech C930E) with certain processing in 1080p resolution using budget grade hardware.

Re-encoding Performance

When there is no necessity to keep the real time processing pace, the cheap tablet showed the ability to do GPU processing on 2K video, which is also good news for those who wants to apply budget hardware to high resolution material.

Apparently, the key factor is ability of the process to keep data in video hardware. As Intel GPU H.264 abilities scale well when used for concurrent multi-stream processing, the performance numbers promise great performance recording video in several formats at a time: raw video, video with overlay, scaled down etc.

The table below gives more numbers for the tests concluded:

Re-encoding Performance Numbers

As mentioned above, the overlaying part itself is a single threaded bottleneck and presumably it is a reserve to be used to cut elapsed time down even further.

Another interesting observation is that while ultramobile system still uses much of CPU time (which is okay – it’s not a powerful system by design), the desktop GPU has minimal impact on CPU while doing pretty complicated task.

Applicability of Virtual DirectShow Sources

Virtual DirectShow  sources have been a long time synonym of software-only camera implementation exposed to applications along with physical cameras in a way that applications consume the sources without making a difference whether the camera is real or virtual. Vivek’s template was a starting point for many:

Capture Source Filter filter (version 0.1) 86 KB zipped, includes binaries.  A sample source filter that emulates a video capture device contributed by Vivek (rep movsd from the public newsgroups).  Thanks Vivek!  TMH has not tested this filter yet.  Ask questions about this on microsoft.public.win32.programmer.directx.video.

With API changes over years, the sample and the concept is still understood as the method of adding a virtual camera, however new scenarios exist where the concept no longer works. Typical problems:

  1. 64-bit applications cannot consume virtual 32-bit virtual sources
  2. Virtual sources are no visible and accessible to applications consuming video using Media Foundation API

The diagram below explains the applicability of virtual cameras:

Applicability of Virtual DirectShow Sources

Important is that virtual sources can only be consumed by the DirectShow-based applications of the same bitness.

If source developer needs to synchronize virtual source throughout multiple applications (e.g. video is synthesized by another application and needs to be deliverable to multiple clients), he needs to add interprocess synchronization on the backyard of virtual source.

If developer needs to support both 32- and 64-bit apps, he needs both variants of virtual sources registered, and possibly synchronization of the kind of the paragraph above.

The only virtual device which is visible to all video capture applications if implemented by kernel level driver (implementations are rare but exist).

C++ #import and x64 builds

I already wrote earlier on 32/64-bit issues with Visual Studio. The problems are not frequent but when they happen they are pretty confusing. Here is another one today.

C++ code is simple:

    #import "libid:59941706-0000-1111-2222-7EE5C88402D2" raw_interfaces_only no_namespace

    CComPtr<IObject> pObject;
    ATLENSURE_SUCCEEDED(pObject.CoCreateInstance(__uuidof(Object)));
    BYTE* pnData;
    ATLENSURE_SUCCEEDED(pObject->Method((ULONG_PTR) (BYTE*) pnData));

A COM method returns a pointer to data – pretty straightforward, what could have gone wrong?

COM server vendor designed the library for easy .NET integration and defined the pointer argument as an integer value. They suppose the values to be used further with System.Runtime.InteropServices.Marshal class.

32-bit builds worked well and 64-bit builds experienced memory access violations. An attempt to consume the COM server from C# project showed the same problem: unexpected exception in the call.

The problem is that cross-compiler importing COM type library using LIBID takes 32-bit library even when it builds 64-bit code. This is the problem for both C++ #import "libid:..." and .NET COM reference using the identifier.

The type library imports as the following IDL in 32-bits:

[id(1)]
HRESULT Method(
                [in] unsigned long bufPtr);

It is supposed that 64-bit builds get the following import:

[id(1)]
HRESULT Method(
                [in] uint64 bufPtr);

Effectively though, 64-bit builds get the 32-bit import and the argument which is supposed to carry casted pointer value is truncated to 32-bits, ULONG type. Cast to ULONG_PTR in 64-bit C++ code is, of course, not helpful since it’s trimmed anyway further fitting the IDL argument type.

The same happens with C# build.

It was developer’s choice to publish ordinal type argument, they wanted this to be “better” and ended up in bitness mess. If the argument remained a pointer type in the IDL then even incorrect bitness would not necessarily result in value truncation.

All together it is unsafe to import [an untrusted] type library using LIBID when it comes to 64-bit builds. It’s 32-bit library to be taken and it can result in incorrect import. Instead, such build should explicitly point to 64-bit type library, for example:

#if defined(_WIN64)
    #import "Win64\ThirdParty.DLL" raw_interfaces_only no_namespace
#else
    //#import "libid:59941706-0000-1111-2222-7EE5C88402D2" raw_interfaces_only no_namespace
    #import "Win32\ThirdParty.DLL" raw_interfaces_only no_namespace
#endif

Too bad! libid looked so nice and promising.

Small correction for DirectShow BaseClasses CTransInPlaceFilter::Copy

False assertion failure in DirectShow BaseClasses transip.cpp, in CTransInPlaceFilter::Copy:

ASSERT(lDestSize >= lSourceSize && lDestSize >= lDataLength);
[...]
CopyMemory((PVOID) pDestBuffer, (PVOID) pSourceBuffer, lDataLength);

The code asserts on “lDestSize >= lSourceSize” which is true in most cases, but it does not have or need to be true. The code below copies just lDataLength bytes and assertion on buffer sizes is, generally, excessive. Valid scenarios do exist with this assertion popping up for no reason.

Screen recording using Desktop Duplication API and hardware H.264 encoder

The application takes advantage of three powerful Windows APIs at a time:

MediaFoundationDesktopRecorder initializes a desktop duplication session and sends obtained desktop images to H.264 video encoder producing a standard MP4 recording. Optionally, it can add an audio track capturing data from one of the standard inputs.

The best performance is achieved when used with hardware H.264 encoder: not only the performance of hardware encoder is better, but additionally desktop images are transferred to the encoder efficiently, without being copied through system memory. With respective hardware, recording is pretty efficient.

There are certain limitations: duplication API is Windows 8+, encoder availability depends on hardware and OS versions. The application let API pick encoder automatically and in worth case scenario falls back to software encoder, which is typically a performance hit.

MediaFoundationDesktopRecorder UI

When started, the application prints initial information, esp. regarding availability of devices, and appends as actions and events take place.

The application uses configuration file with the same name and location as the application, and .INI extension. Changes to the configuration file take effect when the application is restarted.

The application registers Win+F5, Win+F8 hotkeys globally to start/stop recording when the application is in background (that is, when user interacts with another application).

The application generates .MP4 files in the directory of its own location. There will be a video track, and optionally one additional audio track – depending on settings. Video is taken from one of the monitors, and audio – from one of the available standard audio input devices.

The application also generates log files at one the locations:

  • C:\ProgramData\MediaFoundationDesktopRecorder.log
  • C:\Users\$(UserName)\AppData\Local\MediaFoundationDesktopRecorder.log (in case the first path above is inaccessible, esp. due to insufficient permissions)

Configuration

The configuration .INI file might contain a few settings that set up and alter the behavoir of the application:

[Input]
;Video Adapter Description=NVIDIA GeForce GTX 750
Video Output Device Name=\\.\DISPLAY2
;Audio Friendly Name=Stereo Mix (Realtek High Definition Audio)

When started, the application enumerates (“found video…”, “found audio…”) available video and audio inputs. These discoveries are compared against configuration file settings in order to identify monitor for recording, and possibly audio input device.

Default behavior is to take first available monitor, which happens when settings do not instruct otherwise. By default, no audio is recorded. Audio is recorded and added to resulting file if input device is provided explicitly.

The application also prints which devices are taken for further recording (“using adapter…”).

[Format]
;Video Frame Rate=30000
;Video Frame Rate Denominator=1001
Video Bitrate=4096000
Video Texture Pool Capacity=24
Video Throttle=70
Audio Bitrate=192000

Default behavior is to identify monitor’s refresh rate and produce output file with video at the same frame rate. Video Frame Rate and Video Frame Rate Denominator settings offer an override to target file frame rate. With the former value only, it is the frame rate. With both values they define a ratio, e.g. values of 30000 and 1001 result in 29.97 fps file.

Frame rate reduction is a good way to reduce encoding complexity and overall graphics subsystem load.

Bitrate values define respective bitrates for the encoded content.

Details

As recording goes, the application grabs new desktop snapshots and sends them to encoder. There are no specific expectations about frame rate stability and reduction in case of overload of graphics subsystem. When the complexity is excessive, it is expected that some frames might be lost without breaking the entire playability of the output file.

The application provides additional information when it creates a file, for example:

Using Direct3D 11 at feature level D3D_FEATURE_LEVEL_11_0
Using Desktop Duplication mode: Resolution 1680 x 1050, Refresh Rate 59954/1000, Format DXGI_FORMAT_B8G8R8A8_UNORM
Using path “D:\Projects\...\Output\20160707-070707.mp4”
Using video transform Direct3D 11 Aware, Category MFT_CATEGORY_VIDEO_PROCESSOR, Input MFVideoFormat_ARGB32, Output MFVideoFormat_NV12
Using video transform NVIDIA H.264 Encoder MFT, Direct3D 11 Aware, Category MFT_CATEGORY_VIDEO_ENCODER, Input MFVideoFormat_NV12, Output MFVideoFormat_H264
Started writing…
PPP frames written (QQQ frame timeouts, RRR early frame skips, SSS late frame skips)
Stopped writing
Output file size is TTT bytes

When started the application might experience a condition when certain hardware resource is no longer available, e.g. the desktop itself is locked by user. The application will close the file, and attempt to automatically restart recording into new file. The attempts keep going until user explicitly stops recording.

The application does NOT do the following (among things it could):

  • the application is limited to record from one monitor only; to record from two at a time it is possible to start several instances however the produced result will not be synchronized
  • the application does not provide options to record single window image, to cut a section of monitor image or to scale image down
  • the application does not offer choices for video encoders (e.g. there are two or more hardware H.264 encoders), it will always use encoder picked by the system
  • the application only offers bitrate setting for video encoding
  • the application does not provide flexibility in audio encoding settings, it also expects that audio device is available throughout the entire recording session (esp. is not unplugged as recording goes)

References (Informational)

Download links

Build Incrementing for Visual Studio C++ Projects

Over long time I used an automatic build incrementer add-in for Visual Studio and C++ projects, which proved to be helpful. Having increments in file information, the binaries were easy to identify. It was easy to find a matching symbol information etc. Long story short, a tool like this has been a must.

The add-in has problems or downsides though. It kept patching the .RC source and touched it when no other changes existed in the build, touching source code forced rebuilds on its own and reloaded resource-related files opened in Visual Studio editors. I was annoying even though more or less acceptable.

Visual Studio 2015 Community Edition does not support add-ins because of 2015 or because it’s Community Edition. Either way it was time to update the incrementer ot make things nicer overall.

This time I preferred to change things a bit. No longer source code patching: the incrementer can be attached as a post-build event and patch VERSIONINFO resource on the built binary. This requires that current build number is kept somewhere but not in the .RC text, so I am using an additional .INI file. The good thing is that this file can still be included in version control system and the version history can be tracked relatively easily. No longer source code modification which makes code base dirty and forces another rebuild.

Command line syntax:

C:\>IncrementBuild-Win32
Syntax: IncrementBuild-Win32.exe argument [argument...]

Arguments:
  help - displays syntax
  configuration <path> - path to .INI file holding configuration information (mandatory)
  binary <path> - path to binary to be patched with file version update (mandatory)
  string <name> <value> - add, update or remove specific version information string (optional; multiple arguments possible)
  dump - print version information data block dump before and after update

Additional feature is that incrementer can attach additional version strings (see example below – it adds build configuration as a version information string).

Setting up is easy. First, the project should have a version information resource, so that the binary has data to patch in first place.

Then, there should be an .INI file which tracks version numbers. The binary will be build with .RC numbers and then incrementer will apply the least significant number from the .INI file incrementing it along the way.

[General]

[VersionInformation]
;Language=133 ;MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_US)
;Version String Format=%d.%d.%d.%d
Current Build Number=4

Next thing, project post-build event needs a command for patching:

Post-Build Event in VS for C++ Project

"$(AlaxInfo_Common)\..\Utilities\IncrementBuild\_Bin\IncrementBuild-$(PlatformName).exe" configuration "$(ProjectDir)Module.ini" binary "$(TargetPath)" string "ConfigurationName" "$(ConfigurationName)" 

The command takes Module.ini from the projects directory for configuration file, patches build output and also attaches build configuration as an additional version information string.

Build output looks like this:

—— Rebuild All started: Project: EnumerateTransforms, Configuration: Release Win32 ——
stdafx.cpp
Application.cpp
Generating code
Finished generating code
EnumerateTransforms.vcxproj -> D:\Projects…_Bin\Win32\Release\EnumerateTransforms.exe
EnumerateTransforms.vcxproj -> D:\Projects…_Bin\Win32\Release\EnumerateTransforms.pdb (Full PDB)
Configuration Path: D:\Projects…\Module.ini
Binary Path: D:\Projects…_Bin\Win32\Release\EnumerateTransforms.exe
Incrementing build number, product version 1.0.0.1, file version 1.0.0.4
Applying version information string, name “ConfigurationName”, value “Release”

Presumably, it is not necessary to use same bitness tool for a binary, since version information patching API should be able to patch resources of mismatching build, but I normally use a matching tool anyway, why not?

Download links