Originally posted by LensBeginner
but are we sure the camera is only limited by firmware and not, for instance, by the image processor, or something else in the pipeline?
I would be inclined to think it's firmware - the HDMI encoder is a pretty limited bit of kit, It converts one form of digital data input to a second form on output.
You feed it 720, you get 720, feed it 1080, you get 1080.
IMHO, what's happening there, is that it's being fed downstream of the LCD screen, so we get what the LCD gets, which would explain why the HDMI output from 1080p25 footage, is 1080p30 on playback - framerate needed by the LCD.
Ditto why the K-3 outputs 720p60 when recording 1080p25.
Because the only other explanation, is that the sensor engine is always run at 720p60 in video or live view, and the vision is up-converted in firmware to 1080p - that has been done by some camera manufacturers in the past, but that was with small format sensor 'Camcorders' that didn't have the resolution in hardware for 1080p.
In a MILC or DSLR with 16Mpix sensors, that would just be daft.
The camera has an SD-XC controller - that is more then capable of handling much higher data rates in video and in stills then the current throughput. Given a fast enough SD-XC UHS1 card, the SD-XC controller should handle about Eight times more data throughput then the current h264 codec limit of 19mbits.
As for the peaking and Audio levels - the processing power Audio levels would be less then the histogram uses, and the peaking could use the processor bandwidth reserved for the 'Bright/Dark Area' display.