The proposed codec run with a GOF size N=16 using the Haar transform in the temporal direction and the 5/3 spatial wavelet transform at three-levels of the decomposition. At the decoder side we use following error concealment. First, bitstream of each subband is decoded in progressive way until loss is detected. In this case, any losses in stream corresponds to higher quantization at the encoder side and does not lead to error propagation. Second, if main low-frequency subband, which contains brightness information for all GOF, is lost we copy corresponding subband from the previous GOF. In this case the error propagation can happens. To minimize probability of this event, at the encoder side for the main subband we use repetition of the highest bit-planes which can be placed into the two additional source packets.

In case of H.264/SVC we have used JSVM 9.8 reference software with two spatial and five temporal scalable layers. GOP size and intra-frame period are set to 16. For error-resilient coding we have used flexible macroblock ordering with two slice groups and loss-aware rate-distortion optimized macroblock mode decision. Frame copy error concealment method is used at the decoder side. For both codecs the packet length was set to 800 bytes and channel with independent packet losses was simulated.

This movie shows the performance of both codecs at the channel rate (2500 kbps) for different packet loss ratios. One can see that the proposed codec is significantly less sensitive to packet losses and provides much better visual quality while H.264/SVC has a lot of frames with unrecognizable objects.

If you cannot playback YouTube video above, please see this version:

Tampere_01 (640x480, 30 Hz | 300 frames) [download]

Tampere_02 (640x480, 30 Hz | 300 frames) [download]

Tampere_03 (640x480, 30 Hz | 300 frames) [download]

Tampere_04 (640x480, 30 Hz | 300 frames) [download]