Hi there,
Recently I realized that the EDMA TC allocation is critical to the throughput of H.264 encoder and come up with more concerns about the IPNC bandwidth demand.
In the following figure:
Q1. what's the resource demand (EDMA channel/PaRAM no.? and what else ?) of SALDRE?
Q2. what's the resource demand (EDMA channel/PaRAM no.? and what else ?) of face recognition?
Q3. how can VSTAB trigger LDC ? How to obtain lens specific LDC parameters if VSTAB is not integrated ?
Q4. what's the NF quality gain compared with the noise filter in IPIPE ?
Q5. how many EDMA channels and PaRAM sets are required by NF and VSTAB respectively ?
Q6. what's the RSZ1 on the right side for ?
Expecting and thanks for your answers.
BR,
Yulin.
, 