Video Quality Measurement Software Development Kit
(VQM SDK)

No-reference perceived audio/video quality measurement technologies for third-party applications

Now available for Windows and Linux

 

Video Quality Measurement Software Development Kit (VQM SDK) is a library which contains the perceived video quality measurement and perceived audio quality measurement technologies available in Video Quality Monitor (VQM).

VQM SDK produces video and/or audio quality scores which are expressed on a MOS scale (Mean Opinion Score).

VQM SDK's perceived video quality metrics are dedicated to MPEG-2 and MPEG-4/AVC (H.264) video encoding formats. These video quality metrics have been optimized to produce quality scores which are highly correlated with human judgments collected during subjective quality assessment tests in normalized conditions (ITU BT.500 and ITU BT.710).

VQM SDK's perceived audio quality metric can be used with any audio encoding formats (like AAC and MP3 for example). This audio quality metric has also been optimized using the results of audio subjective tests in normalized conditions.

VQM SDK enables to:

  • add perceived video quality measurement capabilities in an existing application
  • add perceived audio quality measurement capabilities in an existing application
  • compute video MOS in real time
  • compute audio MOS in real time
  • extract video frames from virtually any video file format
  • extract audio samples from virtually any audio file format
  • measure blockiness (sometimes called "blocking effect"), blur (sometimes called "blurring"), contrast and jerkiness (sometimes called "image freezing")
  • measure audio bandwidth, silences, saturations, signal breaks and presence of high frequencies
  • compute the impact of jerkiness (image freezing) on quality of experience (QoE)

VQM SDK is a software product which exists for both Windows and Linux.

Overview of VQM SDK

VQM SDK contains the following elements:

  • for Windows: six DLL (dynamic-link libraries) files and one .lib file
  • for Linux: two libraries (.so)
  • one C header file (VQMSDK.h) which must be included in your application
  • documentation (developer manual)
  • C and C++ source code of sample applications for:
    • audio quality measurement
    • video quality measurement
    • audio and video quality measurement

Decoding

The video frames which need to be measured can be:

  • either decoded by VQM SDK from media files (*.avi, *.ts, *.mov, *.mp4, *.mxf, etc.)
  • or submitted directly to VQM SDK (in case they have been captured or decoded by your application)

The audio samples which need to be processed can be:

  • either decoded by VQM SDK from media files (*.mp3, *.aac, *.avi, *.ts, *.mov, *.mp4, *.mxf, etc.)
  • or submitted directly to VQM SDK (in case they have been captured or decoded by your application)

The following pixel formats are supported:

  • YUV 4:2:0 planar (YV12)
  • YUV 4:2:2 packed UYVY
  • YUV 4:2:2 packed YUYV
  • YUV 4:2:2 packed YVYU
  • YUV 4:4:4 YUVA
  • RGB32
  • BGR32
  • Y component only

The following audio sample formats are supported:

  • signed 16 bits
  • signed 32 bits
  • float
  • double

Metrics

VQM SDK contains the following metrics:

  • our no-reference perceived video quality metric for H.264
  • our no-reference perceived video quality metric for MPEG-2
  • our no-reference perceived audio quality metric for any audio codec
  • our no-reference jerkiness metric

Our jerkiness (image freezing) metric takes into account the number and durations of freezing events (generally due to rebuffering) to compute the perceived video quality due to these problems. Subjective tests show that human observers prefer a long freezing event rather than several shorter freezing events. Our jerkiness metric reproduces that too! So now you can measure the impact of rebuffering events of YouTube (or any other streaming service) on perceived video quality!

Measured features

For video, VQM SDK reports (for each frame):

  • perceived video quality (MOS value)
  • blockiness
  • blur
  • contrast (for H.264)
  • jerkiness (image freezing)
  • impact of jerkiness (image freezing) on quality of experience
  • minimum value, maximum value, mean value and standard deviation value of Y samples
  • minimum and maximum values of U and V samples
  • temporal activity

For audio, VQM SDK reports (for each second of audio):

  • perceived audio quality (MOS value)
  • audio bandwidth
  • silences
  • saturations
  • audio signal breaks
  • presence of high frequencies

Conclusion

VQM SDK is a very easy way to integrate state-of-the-art perceived audio/video quality measurement technologies in your application.

Moreover, VQM SDK contains a unique metric to compute the impact of jerkiness events on quality of experience.

Ask for an evaluation version today!