In addition to the FASTR1 and SLOWR1 readout patterns, a handful of FASTGRPAVG patterns are available that reduce data volume.

The target acquisition process requires the target acquisition file to be stored on-board. Since on-board memory can only store the equivalent of 10 full frame groups of MIRI data at a time, target acquisition in ungrouped FAST mode is limited to a maximum exposure time that might be shorter than desired by the user. To get around these data volume issues, the FASTGRPAVG readout patterns are provided that co-add 4, 8, 16, 32, or 64 FAST frames.

The FASTGRPAVG readout modes allow for longer exposure times. This is done by averaging a given number of MIRI frames together in the ISIM (Integrated Science Instrument Module) FPAP (Focal Plane Array Processor) as a group, with a total exposure duration of:

Exposure_duration (for full array) = (N_frames × N_groups × 2.775 s × N_ints) + (2.775 s × (N_ints-1))

For target acquisition, N_ints = 1, so the above equation reduces to only the first term. N_frames can be 4, 8, 16, 32, or 64.

Most of these patterns are only available for target acquisition, but starting in Cycle 6 the FASTGRPAVG8 pattern is available for science observations. It offers several benefits over the SLOWR1 mode, with very similar group time. The SLOWR1 mode samples each pixel 8 times before moving to the next pixel and repeating the process.

Switching between FASTR1 and SLOWR1 is a state change of the detector, which affects the detector stability; as a result, programs using FASTR1 after a SLOWR1 program may see artifacts from the thermal settling following the mode switch. In addition, the SLOWR1 mode does not reduce the read noise compared with FASTR1 to the expected extent. For FASTRGRPAVG8, in contrast, the frames are averaged outside the detector itself; the readout process is identical to FASTR1. There is therefore no state change for the detector, and no settling artifacts. The read noise behaviour is also improved, and the data volume savings are also the same as for SLOWR1. While the SLOWR1 mode will remain available for use in Cycle 6, FASTGRPAVG8 should be the default choice for programs requiring data volume mitigation (parallel programs and/or very deep observations) going forward. A detailed description of the modes and the rationale for this change is provided in Regan et al. (2025).

---

Subarray-dependent group times

Rows for the FASTGRPAVG8 readout pattern are highlighted as it is the only averaged readout pattern that is available for science applications.

Table 1. Readout times for the imager and coronagraph subarrays

Table 2. Readout times for LRS subarrays

---

References

Regan, M. 2025, JWST-STScI-008930
Replacing MIRI Slow Mode with Fast Group Averaging

---

---