dand
22nd July 2005, 10:55
Hi all,
Can someone explain the SBR (no PS) delay related to core AAC encoder?
For example, lets say we have:
Frame x+0: 2048 samples (0001 to 2048) sent to SBR's QMF bank
Frame x+1: 2048 samples (2049 to 4096) sent to SBR's QMF bank
Frame x+2: 2048 samples (4097 to 6144) sent to SBR's QMF bank
etc...
Frame x+0: What 2048 samples should be sent to 2:1 downsampler which would than produce 1024 samples for core AAC's MDCT filterbank?
Frame x+1: ?
Frame x+2: ?
And why (if possible)?
Thanks...
Can someone explain the SBR (no PS) delay related to core AAC encoder?
For example, lets say we have:
Frame x+0: 2048 samples (0001 to 2048) sent to SBR's QMF bank
Frame x+1: 2048 samples (2049 to 4096) sent to SBR's QMF bank
Frame x+2: 2048 samples (4097 to 6144) sent to SBR's QMF bank
etc...
Frame x+0: What 2048 samples should be sent to 2:1 downsampler which would than produce 1024 samples for core AAC's MDCT filterbank?
Frame x+1: ?
Frame x+2: ?
And why (if possible)?
Thanks...