I recently recorded a podcast with Peter, the creator of Open3DOLED – A hardware kit designed to allow certain high-refresh-rate displays to display 3D content like blu-rays and 3D video games. While it’s not currently an easy thing to accomplish, it may soon be the only way we can experience all of the really cool 3D content out there, as no new equipment supports it. Even if you’re not that into 3D, this was still a really cool conversation that I think my fellow nerds will enjoy. You can watch the YouTube video above, or search any podcast app for “RetroRGB Open3DOLED” to listen to an audio-only version:
Open3DOLED: https://github.com/open3doled/open-3d-oled
Peter and I also mentioned a few other projects related to this one and I wanted to link them here to make it easier:
Blur Busters: https://blurbusters.com/
MPC-BE: https://sourceforge.net/projects/mpcbe/
POT Player: https://github.com/open3doled/open-3d-oled/blob/master/PotPlayer/README.md
Wibble Wobble: https://github.com/PHARTGAMES/WibbleWobbleCore
After the interview, Peter was able to test and verify some of the BFI stuff we were talking about. He posted a pretty cool comparison video (embedded at the bottom of this post) and here’s what he had to say:
So I ran the tests with the 1000 fps slow motion camera comparing all the BFI modes on my LG C1 OLED. We were indeed both right! When operating in BFI 60hz, it pushes out frames at 120hz, and every intermediate frame is either black (for high BFI) or some brightness reduced version of a merged frame of the before and after frame (off, low, medium). However when in 120hz BFI (which your TV doesn’t support, it works like a rolling shutter as I described with varying persistence time.
The video below compares all the modes 60hz vs 120hz and all three BFI modes (off, low, medium, high), on the LG C1 OLED. It is interesting that at 120hz BFI it behaves like a camera rolling shutter, where it lowers the persistence time as the BFI strength is increased. Whilst at 60hz it seems to show a frame every other frame.
Then the intermediary frame appears to be either a merged version of the before and after frames (when off, low or medium), or a black frame (when on high).
The source video used in this test was a frame sequential video of flicker blue and red full screen images. The left side of the image corresponds to the top of the screen, and the right side corresponds to the bottom. A 2048×64 pixel ROI was used on a machine vision camera running at 1000 fps to capture the video. It is slowed down 33.33x to 30 fps. Each capture was phase aligned to within 1 frame starting at the start of the video. To compare with accurate phase offset it is recommended to reduce playback speed to 0.25 and start the video from the beginning. You can then pause and use “,” and “.” to step backwards and forward one frame at a time.
Also two other things.
1) I wasn’t able to see any increased brightness in the OLED peak brightness when the BFI was enabled this time. So it’s conceivable when I performed my original photo diode measurements the amplifier wasn’t behaving the same. So it may indeed be the case that when BFI is enabled and pixel persistence is 37% of when BFI is disabled, the brightness may be 37% lower. I thought that it was only 50% lower because it increased the brightness when BFI was on too compensate, but it isn’t obvious from my slow motion video test. It is possible the slow motion camera gain may have been hiding any possible brightness increase as it isn’t clear exactly how it does it’s gain in the firmware.
2) Regarding your DLP projector having a green tint when you turned on 3D mode. I still think this is most likely caused by DLP-link, but normally when the glasses are functioning properly, putting on the glasses should block the DLP-link light pulses (at least 95% of them). So if you were seeing more than 5% of the DLP-link light pulses it’s possible your glasses were faulty. It’s also possible that when it switched to 3d mode, it needed to adjust the color wheel timing, and perhaps there was some issue there, or perhaps with the modified color wheel timing to use the white segment for the DLP-link pulse, they hadn’t provided correct timing parameters for the DLP chip calibrated to the adjusted color wheel timings. I think I mentioned in the interview that I have a Xgimi Aura projector that has a green tint when DLP-link is turned on. about 95% of that light is blocked by my dlp-link glasses. It does still raise the black floor significantly, but the green hue isn’t that visible. It is a laser based light source which I think is why they use green as it probably has the highest conversion efficiency from the UV laser light source I’m guessing they are using to illuminate the phosphor wheel. For HID Xenon lamp DLP projectors, normally they use the white segment for the DLP-link light pulse, as they want the most possible light from the lamp to reach the dlp link glasses photo diode. Also when using a laser though they can pulse it at some over drive current for the 30 to 100 us interval the dlp-link light pulse occurs. Based on this I’m thinking the green was probably caused by some misconfiguration of the color timing parameters when the DLP link 3d mode was enabled requiring the white segment of the color wheel to be used for DLP-link instead of the viewable image. For anyone who wants to know more about DLP link you can check out the datasheet for the dlpc3432 on page 41-43 (it can be found easily online)
…and of course, if you want to check out more information about that awesome OLED I recently reviewed, here’s links to those videos as well:
