This both allows factoring out size check for both MetaSound and TwinVQ-VQF
decoders and fixes the situation when there are several MetaSound frames
stuffed together (that happens in 8kHz @ 8kbps MetaSound in ASF for example).