The following is an extended version of the Delayed Choice Quantum Eraser called "The Envelope Experiment." This may be referred to as the "sealed envelop" or "very delayed quantum eraser."
See steps 1..8 above. This is in the video "Tom Campbell: MBT LA 2016 3 of 7 New Experiments" starting at 1:23:23: https://youtu.be/OZUqtxdT0QY?t=5003
Why is D2 needed? If you know the D1 data (a particle went through slit one) you can know it was not detected in D2 (or vice versa). Prior experiments prove that if you can infer the which-way path of one slit, pattern detection will still work (particle or difference). That would reduce the complexity down to only two detectors and two recorders.