Bit rate estimation for H.264/AVC video encoding based on temporal and spatial activities

Abstract

We present a novel bit rate model for H.264/AVC video encoding which is based on the quantization parameter, the frame rate as well as temporal and spatial activity measures. With the proposed model, it is possible to trade-off the frame rate versus the quantization parameter to achieve a target bit rate. Our model depends on video activity measures that can be easily calculated from the uncompressed video. In our experiments, the model achieves a Pearson correlation of 0.99 and a root-mean-square error of less than 5% with the measured bit rate values, as verified by statistical analysis.

Modeling the bit rate of H.264/AVC video encoding as a function of quantization parameter, frame rate and GoP characteristics

“…In addition, our model captures the influence of the GoP length and GoP structure. Our model has structural similarity with the model proposed in [7] as well as our previously proposed model in [8]. However, we extend the analytical model by two additional factors that consider the GoP length and GoP structure.…”

“…In our model, we follow a similar approach as in [7] and [8]. Therefore, we separate the influence of each parameter and create an impact factor for each individual parameter: spatial correction factor, SCF (q p ), temporal correction factor, TCF (f ), GoP length correction factor, NCF (n), and GoP structure correction factor, MCF (m).…”

“…The main advantage of our approach, compared to the previously proposed rate models of [7] and [8] is that the influence of the GoP structure and length is taken into consideration for the encoding, which has a significant influence on the overall video bit rate.…”

“…This, however, is of special interest for the creation of the video segments in DASH based video streaming, where the GoP length and structure need to be selected for a video segment of a fixed length. In a pre-study we figured out that on average additional bit rate savings of 8% in comparison to [8] can be realized while considering the influence of the GoP structure on the rate.…”

“…The authors extended this metric in [7] with an estimation of the content-dependent parameters based on features which are extracted by a codec-dependent pre-processor. In [8], we developed a bit rate model for H.264/AVC based video encoding which is based on the quantization parameter, frame rate as well as the temporal and spatial activity measures. However, these rate models do not take other parameters into consideration that have a significant impact on the video bit rate, such as the length and structure of the group-ofpictures (GoP).…”