Good points!
I remember cameras 5-10 years ago having all these gimmicks, but then people started to correlate these features with lowend cameras, and manufacturers slowly started removing them.
But suddenly you add fancy words like machine learning to the exactly same words again, and it is cool again somehow.
Marketing always like put fancy words :
Deep learning AF
Previously, programmers manually write AF algorithm/logic.
Currently, AF algorithm too complicated for human written. So, R&D department try use AI (just input different scenarios data, then analysis it via AI for figure out common pattern) for composite AF algorithm/logic.
So, AI just a tools used to write program in R&D department, not used inside camera.
EVF & Rear LCD resolution
- 640x480x3=920K
- 800x600x3=1.44m
- 1280x960x3=3.68m
- X-T30 : 720x480x3=1.0368m dots
- X-T30 II : 900x600x3=1.62m dots
P.s. $270 Samsung Galaxy A32 smartphone has 6.4 inches
1080 x 2400 pixels Super AMOLED.
Dual-processor (used in Nikon Z6 II / Z7 II)
No fund for R&D new processor. So, just put 2 unit identical existing processor,
each fixed to handle different function. Compare with new processor, dual processor drain more battery power & produce more heat for same performance.
It different with multi-processor design in server, server will auto relocate task to processor which free.
Stacked sensor
Stacked sensor has been used in smartphone sensor.
Samsung stacked smartphone sensor consists of two chips: a 12MP backside-illuminated (BSI)
pixel chip on the top that uses
65nm process and a bottom chip for analog and
logic circuits that uses
14nm process. By using the super-fine 14nm process on the processing layer, Samsung says it could achieve a 29% drop in power consumption compared to current
conventional sensors that use a 65nm/
28nm process.
https://dpreview.com/news/218354003...power-efficiency-density-mobile-image-sensors
Similar faster smartphone processor chip build via smaller 7nm/5nm process, stacked sensor fast because built via smaller 12nm process instead of 28nm process.
$270 Samsung Galaxy A52 5G used Qualcomm Snapdragon 750G 5G (
8 nm).
Rumors said core tech used stick dual chip together licensed from other company to Sony/Nikon/Canon...
Compact camera built-in ND
Small grey film in lens, refer to 22:40 in below teardown
Film simulation
Customizable filter for provide better tone mappings. Similar to LUTs used in videography.
This thread has got me thinking, computational photography might be an excuse to manufacture cheaper lenses, and to just “correct” the defects in post-processing.
Are you means : "Sony a7 IV adds a
Breathing Compensation mode that crops and resizes the video to cancel-out any change in a lens's angle-of-view (AoV) as it focuses. The mode only works with select Sony lenses (all the GM lenses and some G series glass), as the camera needs a profile of the breathing characteristics. Video is cropped to match and maintain the narrowest AoV that might occur if you focused from minimum focus distance to infinity, meaning there's no distracting change of framing as you refocus.".
