sps_extension_6bits and sps_extension_data_flag are reserved for future use, but according to spec, a decoder should be able to decode bitstreams if they are not equal to zero.

"When present, sps_extension_6bits shall be equal to 0 in bitstreams conforming to this version of this Specification. Values of sps_extension_6bits not equal to 0 are reserved for future use by ITU-T | ISO/IEC. Decoders shall allow the value of sps_extension_6bits to be not equal to 0 and shall ignore all sps_extension_data_flag syntax elements in an SPS NAL unit."

The same is applied to pps_extension_6bits and pps_extension_data_flag as well. Look at the patch. It allows the values of the syntax elements to be non-zero.

1. num_add_layer_sets value is not set at decoder when NumIndependentLayers is less or equal to 1.
2. Decoder assertion when 2nd and 3rd enhancement layers only depend on the base layer.

Suggestion to fix SHM decoder on PPS scaling list inferring process.

When pps_infer_scaling_list_flag is equal to 1 in EL PPS, the current decoder makes MD5sum mismatch. I think it is due to the wrong if-branch condition of the decoder.

Attached zip includes a sample bitstream to reproduce the bug and a suggested fix.

The sample bitstream has two layers. The BL PPS has pps_scaling_list_data_present_flag equal to 1, and the EL PPS has pps_infer_scaling_list_flag equal to 1.

I confirmed MD5sum match by SHM-dev decoder (rev.1008) with the fix patch.