How to Perform AMP? Cubic Adjustments for Improving the QoE

[EN] Adaptive Media Playout (AMP) consists of smoothly and dynamically adjusting the media playout rate to recover from undesired (e.g., buffer overflow/underflow or out-of-sync) situations. The existing AMP solutions are mainly characterized by two main aspects. The first one is their goal (e.g., k...

Descripción completa

Detalles Bibliográficos
Autores: Montagud, Mario, Boronat, Fernando|||0000-0001-5525-3441, Roig, Bernardino|||0000-0002-9599-572X, Sapena Piera, Almanzor|||0000-0001-8473-6063
Tipo de recurso: artículo
Fecha de publicación:2017
País:España
Institución:Universitat Politècnica de València (UPV)
Repositorio:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglés
OAI Identifier:oai:riunet.upv.es:10251/98369
Acceso en línea:https://riunet.upv.es/handle/10251/98369
Access Level:acceso abierto
Palabra clave:Adaptive Media Playout
Cubic interpolation
Inter-Destination Media Synchronization
Media synchronization
QoE
MATEMATICA APLICADA
INGENIERIA TELEMATICA
Descripción
Sumario:[EN] Adaptive Media Playout (AMP) consists of smoothly and dynamically adjusting the media playout rate to recover from undesired (e.g., buffer overflow/underflow or out-of-sync) situations. The existing AMP solutions are mainly characterized by two main aspects. The first one is their goal (e.g., keeping the buffers¿ occupancy into safe ranges or enabling media synchronization). The second one is the criteria that determine the need for triggering the playout adjustments (e.g., buffer fullness or asynchrony levels). This paper instead focuses on a third key aspect, which has not been sufficiently investigated yet: the specific adjustment strategy to be performed. In particular, we propose a novel AMP strategy, called Cubic AMP, which is based on employing a cubic interpolation method to adjust a deviated playout point to a given reference. On the one hand, mathematical analysis and graphical examples show that our proposal provides superior performance than other existing linear and quadratic AMP strategies in terms of the smoothness of the playout curve, while significantly outperforming the quadratic AMP strategy regarding the duration of the adjustment period and without increasing the computational complexity. It has also been proved and discussed that higher-order polynomial interpolation methods are less convenient than cubic ones. On the other hand, the results of subjective tests confirm that our proposal provides better Quality of Experience (QoE) than the other existing AMP strategies.