SPRUJ28E November 2021 – September 2024 AM68 , AM68A , TDA4AL-Q1 , TDA4VE-Q1 , TDA4VL-Q1
The merge block supports merging up to 3 exposures to generate a cumulative 20 bit frame.
The merge block is based on 2 instances of the Motion-Adaptive merge block, each of which is independently capable of supporting the merging of 2 exposures (10 – 16 bits each) to a combined bit width of up to 20 bits. In a more generic scenario the merge block should be designed and verified to support the following scenarios
Using two instances of the motion adaptive merge block provides the functionality to merge up to 3 independent exposures, each of 12 bits, to a combined bitwidth of 20 bits. The block also provides capability to only merge 2 exposures using a combination of mux and clipping blocks.
The merge block comprises of two instances of the motion adaptive merge (MA1/MA2) as well as an LUT based companding block and mux structures to enable different data flow. To keep the design complexity under control some restrictions are applied to the data connectivity on the merge blocks. A key restriction is that for the two exposure merge only the MA-2 block should be used and the MA-2 should be used along with MA-1 for 3 exposure merge. Further the position of the long and short exposures on each block are fixed. Table 6-136 summarizes some of the possible combinations which can be realized using the logic. (Note, this is not an exhaustive list from a DV perspective).
The MA block when set to bypass/disabled mode, will pass the input stream marked as ‘short’ (refer to Figure 6-60) to the output.
Merge Mode | MA-1 | MA-2 | LUT | |
---|---|---|---|---|
2 Exposures (12 bit) to 20 bit | Bypassed | Merge Long / Short | Use VS and Short Exposures | Set to 20-16 |
2 Exposures (12 bit) to 16 bit | Bypassed | Merge Long /Short | Use VS and Short Exposures | Bypassed |
2 Exposures (12 bit) to 16 bit with tone mapping | Bypassed | Merge Long /Short | Use VS and Short Exposures | Set to 16-16 mode |
2 Exposures (12 & 16) to 20 bit | Bypassed | Merge Long /Short | Use VS and Short Exposures | Set to 20-16 |
3 Exposures (12 bit each) to 20 bit | Merge Long and Short | Merge VS / MA-1 output | Use all 3 exposures | Set to 20-16 |
Single exposure companded | Bypass | Bypass | Use VS exposure only | Bypassed |