This database contains 170 videos sequences, the subjective results including votes and Mean Opinion Scores and auxiliary data such as the transmitted video files and their parsed version, the Hybrid Model Input XML (HMIX) files. 10 different video source contents were used. For each content, the reference (without degradation) and 16 different degradations called Hypothetical Reference Circuits (HRCs) were subjectively evaluated. The HRCs are based on H.264 coding with and without transmission errors. Several encoders, decoders and packet loss schemes were used including transcoding scenarios, spatial and temporal subsampling. A description of the SRC and the HRC is provided in the description file including the description of the viewing environment in which the video quality was evaluated. We used Absolute Category Rating with Hidden Reference (ACR-HR) with 5 point scale as subjective experiment methodology. We provide a spreadsheet with the individual scores and the Mean Opinion Score (MOS) for each of the 170 Processed Video Sequences (PVS). The PVS (avi videos) are freely available on our FTP server. The source contents stem from several sources, including the Consumer Digital Video Library (CDVL), user generated content from the University of Vienna, and an excerpt of Sita Sings the Blues by Nina Paley. Some sequences were obtained using the left view of our NAMA3DS1 database. To see further information on the sources and their respective copyright information please refer to the SRC description in the description file. The dataset was created in the context of the Joint Effort Group / Hybrid project of the Video Quality Experts Group (VQEG/JEG-Hybrid). It may be used for training hybrid video quality measurement algorithms and to prepare indicators for large-scale verification on the upcoming VQEG/JEG-Hybrid database. Additional to the decoded video file, we provide the network bitstreams in PCAP format and the parsed version of the PCAP files as defined by VQEG/JEG : the Hybrid Model Input XML file (HMIX) files. Preview images are provided for getting an impression on the content seen by the observers during the subjective experiment.
Video sequences and other materials available at FTP: Link: ftp://ftp.ivc.polytech.univ-nantes.fr/JEG264HMIX1/ Qualinet Databases Mirror: Link: ftpes://multimediatech.cz/IRCCyN_IVC/IRCCyN_IVC_JEG264HMIX1 Username: dbq-mirrors Password: kucykepe
The original link was: http://www.irccyn.ec-nantes.fr/spip.php?article1033
References and Citation
Please, cite the following paper in your reference if you use this database for your work [BSJ12].
BSJ12: Marcus Barkowsky, Nicolas Staelens, Lucjan Janowski, Yao Koudota, Mikołaj Leszczuk, Matthieu Urvoy, Patrik Hummelbrunner, Iñigo Sedano, Kjell Brunnström, "Subjective experiment dataset for joint development of hybrid video quality measurement algorithms", Third Workshop on Quality of Experience for Multimedia Content Sharing (QoEMCS), Berlin, July 2012.