SoftCast Test Videos

Introduction

We compare SoftCast to MPEG-4 part 10 (H.264/AVC). MPEG is encapsulated in MPEG-TS and encoded using Reed-Solomon 188/204 code and 3/4-rate convolutional code and modulated with QPSK (equivalent to 802.11a/g at 18Mbps).

Both SoftCast and MPEG use the same wireless resources, i.e. the transmitter uses the same average power and channel bandwidth. Both SoftCast and MPEG use GOP of 16 frames.

Mobility

In this simulation, we show one video as the receiver moves away from the sender so that the channel SNR drops from 12dB to 4dB. As the quality of the channel gets worse, at some point MPEG crashes due to bit errors. In contrast, SoftCast video "adapts" to the available channel quality, but it happens without any actual adaptation at the transmitter or feedback from the receiver.

We perform this comparison for different effective channel bandwidth:

• 1.30 MHz
• 0.32 MHz
• 0.24 MHz
• 0.16 MHz

Same experiment for the Table Tennis video [ no chart ]

Packet Loss

In this simulation, we run a composite video at fixed channel bandwidth (1.3 MHz) and SNR (12dB) but with induced packet loss. Note that this packet loss would not be caused by noise resulting in bit errors, but rather interference or deep fading (whole packets are erased).

We perform this comparison for packet loss rate of 1% and 10%. Note that for better exposition the video is slowed down to 15 fps (0.5x speed):

• Packet Loss