I'd disagree that a lower resolution 7D II would be better. ISO performance can be improved in other ways besides larger photodiode area these days. Only when we hit the maximum achievable Q.E. will we have no other option but to use larger photodiodes. Reach is the key benefit of the 7D line...and reach has to do with pixel density, not sensor size. A 24.1mp APS-C 7D II would bring in the buyers more than anything else, as that is what the 7D line is all about...reach with sports-level performance.The 7D is also famed for its frame rate and buffer depth. If the MP is kept in check, there's no reason why Canon can't push the boundaries and make that small mirror and shutter move much faster in the mk II. But if they increase it to 24 MP, it's likely to remain at about 8 FPS.
Well, I just ran these calculations the other day. People have gravely miscalculated the data throughput rate of the 1D X at 144-165mb/s per DIGIC 5+ chip. Most of the calculations involve an 18mp image size, and do not account for metadata or other overhead. Additionally, not all throughput calculations people do factor in the 14fps frame rate of the 1D X (which it can do with mirror lockup.) The "real" pixel count of the 1D X is 19.3mp, as Canon masks off a border of pixels for calibration purposes, and those pixels ARE included in the RAW image when it is saved to the CF card. Accounting for a higher "real" RAW image pixel count, 14fps frame rate, and a buffer for any overhead:
19,300,000pixels * 14bit / 8bit/byte * 14fps = 473mb/s
Assuming there is some overhead, and the need for a little bit of leeway for metadata for each image, and the need to account for performing compression and the like, I'd say the total throughput of the 1D X is at least 500mb/s. That is 250mb/s per DIGIC 5+ dsp chip. Now, if we run a similar calculation for the 7D II assuming a 24.1mp image size, and a similar 7% additional pixels for the masked border pixels, we have a "real" 7D II pixel count of 25.8mp (25,787,000 pixels). Running the same calculation, only for 10fps instead of 14fps:
25,787,000 pixels * 14bit / 8bit/byte * 10fps = 452mb/s
Even with a considerably higher pixel count, the 7D II with a pair of DIGIC 5+ processors should be able to handle 10fps no problem. That would actually be a LOWER data throughput rate than the 1D X at 14fps! We can run the numbers for 11fps, too:
25,787,000 pixels * 14bit / 8bit/byte * 11fps = 497mb/s
Even at 11fps, we are still below the 500mb/s total that would be allowed if each DIGIC 5+ chip could process at a throughput rate of 250mb/s (which, to me, seems like a more plausible throughput rate than 165mb/s that you get at 18mp and 12fps). If you factor in the facts that the DIGIC 5+ processors have to not only perform ADC on the incoming sensor pixel data, but also compress and write the output pixels to the memory cards, the raw I/O throughput rate of these chips has to be very high.
I see no reason the 7D II couldn't have both a higher pixel count AND the additional 2fps bonus over the 7D I. The math certainly adds up...