krontech / chronos-cam-app Goto Github PK
View Code? Open in Web Editor NEWChronos Camera App
License: GNU General Public License v3.0
Chronos Camera App
License: GNU General Public License v3.0
Performing black calibration generates a small recording segment in memory to acquire the black reference images. This should not be considered a valid recording data, and the Unsaved video warning just creates unecessary UI noise.
At horizontal resolution exactly 464, the overlay text looks "doubled" and hard to read.
At horizontal resolution below 464, the overlay text is unreadable.
Now that WT79 is the fastest wavetable for full-frame images, the factory calibration procedure never selects WT80 for anything (it always chooses the fastest setting). We should add a test case for underclocking the sensor.
Steps to reproduce:
Changing the focus aid status, either on the main screen or the Util screen, has no effect until OK is pressed on Record Settings. Entering and exiting the Play screen may also temporarily fix the focus aid status.
Setting custom white balance in warm or cool light just makes the image even warmer or cooler.
Presets also have the opposite effect intended.
camUpdate-v0.3.1-rc1-4-g112da3a
Enabling invert doesn't seem to have an influence on the trigger. Trigger 3 will always activate once the signal reaches 1.5V regardless if a rising or falling signal is applied. This behavior is different to how IO1 and IO2 reacts.
This behavior is also different to what was mentioned in the manual. Once 1.5V is reached the camera will recognize it as "HIGH", but this should only be "HIGH" when a 3 - 30V signal is applied.
Ethernet carrier detection doesn't appear to work reliably on some switches if the camera is powered up with the cable plugged in. This appears to be more prevalent on cheap consumer switches than it does on our bigger rackmount gear.
In particular, the 8-port D-Link switches seem to reproduce this on every boot, meanwhile the Mikrotik and HP rackmount switches boot up correctly every time.
Image brightness will increase/decrease every couple of frames when using resolutions that are exactly 480 pixels in height. More noticeable at resolutions with less horizontal pixels.
The PMIC firmware has a mode which can automatically power the camera on and off when AC power is applied and removed. However, there is no UI to configure this feature right now. This should be added to the Util window as a checkbox.
To reproduce:
Expected behavior:
How to reproduce:
If the trigger is pressed more times than there are segments, the video is sometimes corrupted. This seems to occur more often when the first and last trigger are separated by more frames than what the total length of the recording is supposed to be (eg. on a 8gb camera set to 1280x1024, first and last triggers separated by more than 4365 frames).
Symptoms:
Corrupted recordings appear 70k to about 170k frames long, at least at 1280x1024. The first n frames look normal, where n is the total record length specified on the record mode screen. The rest of the recording is various forms of garbage, including strangely colored images and vertical lines.
At 30 and 45fps, the image looks OK until standard black cal is performed, then the image looks very dark.
At 60fps, the image looks OK both before and after standard black cal.
tested on latest 3.2 beta
If the saved file frame rate is set lower than 60fps, the maximum bitrate will not adjust the max bitrate, but instead the max megabits per 60 frames.
Connect a storage device with multiple partitions. On the list of partitions for that device, the third line of text (which applies to the second partition) is partially cut off.
When the camApp encounters a failure, it simply displays an error number which is hard to make sense of. We should add a dictionary to convert error numbers into strings to make it easier to explain what the problem might be.
The new self-calibration code generally works pretty well, but it presumes that the user is always
able to perform a calibration before starting recording. This may not always be the case. We should store the most recent calibration results into a file and load them on boot and when switching resolutions.
However, the files can be considered a part of the user cal and can be regenerated whenever performing a manual black calibration.
This is also more important for the LUX2100, which has a really slow ADC offset training algorithm (for now).
Shutting Down is displayed until the camera is shut down manually.
After using the keyboard to enter a horizontal resolution below 192, or a vertical resolution below 96, the vertical or horizontal resolution just reverts to what it was before the keyboard was opened, instead of going to 192 or 96.
The TIFF tags for files saved by a monochrome camera were:
Photometric Interpretation: RGB color
Samples/Pixel: 3
whereas they should be:
Photometric Interpretation: min-is-black
Samples/Pixel: 1
The camera model is correctly shown on the About screen as Monochrome.
frame_000002.tiff.zip
I'm really not sure how often this happens or what it takes to reproduce it.
At least I've never seen it happen on playback mode or in saved videos.
Even with battery installed, No Batt is intermittently displayed in place of the battery % and voltage in the bottom right corner of the main screen.
Forum thread, started in early May: http://forum.krontech.ca/index.php?topic=468.0
When the text overlay is enabled, it shows T=0.00000000s for every frame in the recording.
Given the difficulty of opening both inputs at the same time (having both inputs open at the same time does not trigger the bug) as well as the infrequency of usage of the date input, we are not currently planning to fix this issue. However, while it remains as reported by Afion, it is filed here so that we know about it.
User's have reported that the exposure before and after a black cal has changed.
The normal way to set max exposure is either by tapping Max on record settings when setting a resolution, or dragging the shutter slider to the top on the main window. However, to set the camera to the absolute maximum available (which is only 3-5 microseconds higher), an extra step is currently needed:
Go to record settings and tap Max, even if you already did so in step 1. Then press OK.
This happens most or all of the time after booting my camera at 640x480, but not at 1280x1024.
The second frame in any saved video is not the second frame in the region that was selected on the play screen, but instead is from whichever frame was viewed on the playback window at the time the save was started.
Happens at least on h.264, CinemaDNG and TIFF. I can test the other formats if necessary.
May happen more often on faster storage mediums - I can test more for this too if necessary.
Focus aid and zebras are sometimes disabled after a reboot when they were enabled before.
There appears to be a rounding error when selecting 336x120 resolution somewhere in between the calculation of the maximum framerate and the selection of the wavetable for that rate that results in the selection of WT19 instead of WT20. This is results in a small loss of image quality at 336x120 due to the increased noise of WT19.
to reproduce this:
Edit the resolution and offset so that the rectangle that indicates the used portion of the sensor is at least partially obscured by the keyboard.
Tap the text in either of the boxes for editing the Offset, then edit it using either the keyboard or the up/down arrows.
Close the keyboard.
Visual bug only, not related to functionality.
After saving hundreds or thousands of videos, the camera may lock up on the playback/save window.
It has been observed that the starting frame for video being saved by the camera jitters by one frame. This causes difficulty when it comes to synchronizing multiple cameras together.
This can easily be avoided by not changing the exposure time during a recording.
When exposed to a high-end environment, it has been observed that false touch events can be generated on the LCD screen. This makes the camera difficult if not impossible to use in some environments.
The incidence of false events has been reported to depend on the strength and angle of attack of the wind.
When the camera is operated in gated burst mode, the live display does not appear to be buffered correctly when the trigger signal is inactive.
To reproduce this issue, set up the camera to film a fast-moving and repetitive event (a spinning fan makes an excellent subject), configure the camera for gated burst mode with 1 prerecord frame (the default), and then start recording. While the trigger signal is inactive the video appears to be torn at multiple horizontal positions across the fan. The frame tearing appears to go away when the trigger is activated.
Tested on FPGA build 3.21.
most noticeable at 1280x1024 60fps, as the range of exposures where this occurs is larger than the range at 1057fps. At 60fps, the range starts at about 16 degrees.
When in gated burst mode, the FPGA appears to stop detecting trigger events after several edges are received. Stoping and restarting the recording appears to get the FPGA back into a working state.
However, the date that is displayed in the file's name (if autoname was used) will be correct.
When the dialog box for record settings or any other window is tapped during a recording, the dialog box pops up asking if you want to go to save settings.
If a record end trigger happens while this is open, the playback screen comes up and the dialog box stays open, allowing you to open another window during a save.
If any of the 0%, 50% or 100% buttons are pushed on the trigger delay screen while the slider has jog wheel focus, the slider stays highlighted and the slider handle does not move. At least the buttons do have their intended effect of changing the numbers, and these numbers are still applied correctly if the OK button is pressed while the numbers do not match up with the position of the slider.
Image looks more yellow than it should at these framerates and gains:
framerate above 31.99kfps at gain 12
framerate above 11.11kfps at gain 24
camUpdate-v0.3.1-rc1-4-g112da3a
Occurs to all Trigger IO for the test
This was tested on 2 different cameras, 00020 (color 16GB) and 00031(color 16GB)
Performing the gated burst test resulted in recordings with 2 frames or recordings with high number of frames(70,000 - 400,000). The recording (3 seconds) would usually be around 3000 frames.
Video recording is damaged. Various vertical lines patterns shown in the video playback. Vertical lines on video playback is changing on its own. May be accessing previous recording data on the RAM since the video is showing some recordings at completely different times.
Reverting to defaults may allow you to get a correct recording, but subsequent recordings will have the errors mentioned above
After reverting user settings to default, Auto Power Mode should be reset too, but it isn't.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.