Reducing the shot noise requires more photons arriving at each pixel. Getting more photons per pixel requires either (i) bigger pixels on the detector, (ii) better illumination of the subject, or (iii) better optics.
(iv) Increase photon capture efficiency.
The article says that in conventional CMOS sensors, three quarters of the incident photons are either absorbed by a metal layer or hit a spot between photo diodes, not contributing to photo diode charge and read-out signal. The new coating can convert those photons into charge, increasing the signal by a factor of four without changing pixel size, optics or illumination. Noise will be lower.
Hamburg central station started playing classical music about 10 years ago. The reason was to drive off junkies; there's a saying thay certain drugs in combination with classical music lead to a bad trip.
You'd still need ECC RAM. Otherwise, when reading a broken page from 2 modules, you'd get different data, but don't know which one is correct. With ECC modules, one of the pages read would fail the ECC check and you'd just use the other one.
MTBF is defined quite unintuitive. It does not say anything about a single drive, but average failure of a *lot* of drives which are only used in their specified lifetime.
So, you take some thousand drives and let them work. If a drive fails, you replace it and count a failure for the MTBF statistics. If a drive works to the end of its specified lifetime (e.g. five years), you also replace it, but it does *not* count as a defect for MTBF.
Then, you calculate (number of drives) times (hours of service) and divide by the number of defects replaced, and this is the MTBF.
You can't compare this. A real 512MB machine would swap to disk. With a 512MB VM on a 1GB machine, the
VM's swap file accesses go to the host's file system cache. This is much faster.
I measured time-to-display (input lag + grey-to-grey transition time) on a Dell 20" in these three modes.
These modes differ in some visual aspects (gamma curve?) but not in time-to-display.
A cap per LED is a bad idea. It had to be in parallel to the LED (bridging the LED diode), so you'd need another diode per LED.
It might reduce flicker somewhat, but the peak current will be even higher as it has to light the LED and load the cap.
This multiplexing scheme already is limited by pin current at
the microcontroller, so you don't gain brightness.
There are surely cheaper ways. Like buying a controller with more outputs, increasing the number of LED's that can be illuminated at the same time, thus increasing the amount of time that can be spent illuminating each LED.
You also could add some cheap latch ICs (74LS373, 20 cents in single quantities here) to make a lot of digital outputs (though not tristate ones).
These have 8 inputs and one control signal. The 8 outputs copy the input pattern from when the control signal was last strobed (0->1 transition).
With n latches and 8+n microcontroller pins you get 8*n (static, non-multiplexed) outputs.
Of course, you could use these outputs to multiplex some larger LED array.
Refresh rate is configured by the video mode. It is independent of display response time.
TFT response time is the time it takes the display to switch from dark to light.
If is not the time between [new pixel is sent from the GPU to display electronics] and [pixel intensity on screen changes according to data sent]. This is called Input Lag.
Typical TN displays (the cheap consumer ones) have a response time of 5-10ms and an input lag of 30ms.
Slashdot Mirror
This raises a question I've been wondering about for a while, which is whether it would make sense to make a 64-bit ARM?
I don't know whether it made sense, but it was already done in 2001.
http://www.zdnet.co.uk/news/processors/2001/10/16/arm-to-unveil-64-bit-jaguar-chip-for-handhelds-2097352/
Reducing the shot noise requires more photons arriving at each pixel. Getting more photons per pixel requires either (i) bigger pixels on the detector, (ii) better illumination of the subject, or (iii) better optics.
(iv) Increase photon capture efficiency.
The article says that in conventional CMOS sensors, three quarters of the incident photons are either absorbed by a metal layer or hit a spot between photo diodes, not contributing to photo diode charge and read-out signal. The new coating can convert those photons into charge, increasing the signal by a factor of four without changing pixel size, optics or illumination. Noise will be lower.
If it works as advertised, this is a good thing.
Hamburg central station started playing classical music about 10 years ago. The reason was to drive off junkies; there's a saying thay certain drugs in combination with classical music lead to a bad trip.
You'd still need ECC RAM. Otherwise, when reading a broken page from 2 modules,
you'd get different data, but don't know which one is correct.
With ECC modules, one of the pages read would fail the ECC check
and you'd just use the other one.
MTBF is defined quite unintuitive. It does not say anything about a single drive, but average failure of a *lot* of drives which are only used in their specified lifetime.
So, you take some thousand drives and let them work. If a drive fails, you replace it and count a failure for the MTBF statistics. If a drive works to the end of its specified lifetime (e.g. five years), you also replace it, but it does *not* count as a defect for MTBF.
Then, you calculate (number of drives) times (hours of service) and divide by the number of defects replaced, and this is the MTBF.
You can't compare this. A real 512MB machine would swap to disk. With a 512MB VM on a 1GB machine, the VM's swap file accesses go to the host's file system cache. This is much faster.
I measured time-to-display (input lag + grey-to-grey transition time) on a Dell 20" in these three modes. These modes differ in some visual aspects (gamma curve?) but not in time-to-display.
I wanted Apple's poorly-coded (for Windows at least) proprietary video player.
Nah. You don't want the player, you just want the codec. (Google for Quicktime Alternative.)
There are surely cheaper ways. Like buying a controller with more outputs, increasing the number of LED's that can be illuminated at the same time, thus increasing the amount of time that can be spent illuminating each LED.
You also could add some cheap latch ICs (74LS373, 20 cents in single quantities here) to make a lot of digital outputs (though not tristate ones). These have 8 inputs and one control signal. The 8 outputs copy the input pattern from when the control signal was last strobed (0->1 transition). With n latches and 8+n microcontroller pins you get 8*n (static, non-multiplexed) outputs. Of course, you could use these outputs to multiplex some larger LED array.
Refresh rate is configured by the video mode. It is independent of display response time. TFT response time is the time it takes the display to switch from dark to light. If is not the time between [new pixel is sent from the GPU to display electronics] and [pixel intensity on screen changes according to data sent]. This is called Input Lag. Typical TN displays (the cheap consumer ones) have a response time of 5-10ms and an input lag of 30ms.