BREAKTHROUGH: Hand-waved iPhone 13 pursuit camera outperforms rails for rapid field testing

Many sites including LinusTechTips, RTINGS, TomsHardware, and others use the free Blur Busters pursuit camera invention. Now also available as a rail-less smartphone wave, too!
Post Reply
User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

BREAKTHROUGH: Hand-waved iPhone 13 pursuit camera outperforms rails for rapid field testing

Post by Chief Blur Buster » 05 May 2022, 17:45

Handwave Smartphones Now Outperform Rails for Field Testing

After collecting data for 5 years, we have discovered that amazingly, handwave pursuit camera can sometimes outperform “less-than-high-end” field rail setups.

An expert permanent headquarters rail will always outperform hand-waves, but field testing often use tripods that are shaky and too lightweight, e.g. capturing pursuit photos of a display at a convention or at a store (like Best Buy).

Pursuit Camera Rail Problems For Rapid Field Testing

- Shaky light tripods
- Carpets at convention floors
- Heavy camera that flexes cheap travel-friendly camera rails
- Some conventions won't let you enter with large metal objects (security).
- Rail vibrations while camera slides on cheap lightweight rails
- Only few photo attempts are possible in 5 minutes
- You have to "fight" the camera settings sometimes.

Handwave Smartphone Advantages For Rapid Field Testing

- 4K Video as burstshoot substitute (1000+ photographs in 30 seconds)
- Custom apps can be downloaded
- Brute-force thousands of photographs compensates for error of hand-shaky
- Easy to jog-shuttle forward/back in video players to find the good photos
- TestUFO "Sync Track" invention is almost as good as a defacto a cryptographic certificate of hand tracking accuracy
- Much less setup/teardown required

Why good practiced iPhone/Galaxy handwaves now outperform rapidly-setup field rail builds

- Field testing requires rapid setup and teardown.
- Sometimes tripods have to be put into a carpet in a convention.
- Sometimes a SLR camera is too heavy for a lightweight mobile camera rail.
- Cheap lightweight camera rail flexes and vibrates a lot.
- Cheap tripods flexes too much.
- Easier media management of originals; one video file instead of a pollution of unusable burstshoot JPGs
- No time to validate a field build of a camera rail (convention, TV store, prototype)
- Handwave smartphone means no airline baggage, no carryons needed! Travel to conventions like CES/DisplayWeek/Computex/CEDIA/etc via ultra low fare airlines like RyanAir, Swoop, AlaskaAir, etc!

New 8K Phones Do Excellent 4K Video Almost SLR Quality Per Freezeframe

- 30fps video burst-shoots 1800 photographs per minute (brute-force your way out of shaky hand error margin)
- 60fps video burst-shoots 3600 photographs per minute (brute-force your way out of shaky hand error margin)
- Phone technology has now improved massively as of year 2022
- Some phones have amazing excellent image stabilization that does not affect freezeframe quality
- Camera apps on iPhone such as DSLRCamera or ProCam that turns video into a manual burst-shoot emulator

Handwave Professionals Now Reporting 0.1 Pixel Error Margins Now!

A few years ago, handwaved smartphones were achieving 0.25 pixel error margins.

Two years ago, cheaper phones did this:
Image

This already finally matched the rail quality of a mid-range field test rail-based pursuit camera rig with its cons listed above.

Today, with newer 8K optical-stabilized smartphones (iPhone 13) but using downconverted high-bitrate 4K settings to its high speed SSD-quality flash memory...

It is now possible to get 0.1 pixel error margin with handwaves, perhaps less, on some newer 8K-capable smartphones running at its conservative 4K settings for near-SLR-quality per freezeframe. Apparently, its optical image stabilization responds fast enough to sometimes do a near-perfect pursuit greatly compensating for the error margins of a shaky hands.

This exceeds rail-quality for field rigs!
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
Discorz
VIP Member
Posts: 999
Joined: 06 Sep 2019, 02:39
Location: Europe, Croatia
Contact:

Re: BREAKTHROUGH: Hand-waved iPhone 13 pursuit camera outperforms rails for rapid field testing

Post by Discorz » 06 May 2022, 14:23

Lovin' it. Its amazing what can be captured with these days phones.

These two are best hand-waved photos I could get with my smartphone. Took many attempts.

Image

Image

Gigabyte M32Q @170Hz, 170 FPS
Aim Stabilizer Sync at middle screen
Overdrive Balance
960 pps
Xiaomi Redmi Note 9S camera, photo mode, no stabilization
Compare UFOs | Do you use Blur Reduction? | Smooth Frog | Latency Split Test
Alienware AW2521H, Gigabyte M32Q, Asus VG279QM, Alienware AW2518HF, AOC C24G1, AOC G2790PX, Setup

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: BREAKTHROUGH: Hand-waved iPhone 13 pursuit camera outperforms rails for rapid field testing

Post by Chief Blur Buster » 06 May 2022, 14:45

Discorz wrote:
06 May 2022, 14:23
Xiaomi Redmi Note 9S camera, photo mode, no stabilization
Good attempt!

Even though this is not as SLR-quality as an iPhone 13:

I find that the iPhone 13 is much better though, it doesn't show any camera-artifacts (like horizontal bars or photo curving), and the optical stabilization helps noticeably.

Also, incredibly, in one of the two attempts, you achieved roughly ~0.1 pixel hand-tracking accuracy with that old phone:

Zoomed version of your first photo:
subpixel-accuracy.png
subpixel-accuracy.png (1.05 MiB) Viewed 8134 times
Given sufficient attempts, guess you don't need an iPhone 13 optical stabilization to handwave that accurately, thanks to the bruteforce spray of video frames, giving many "chances" of chance-accuracies in your handwave pan.

As you swung as moving smartphone sideways, you even captured 4 refresh cycles' worth of screendoor artifact darn nearly perfectly on top of each other -- preserving the screendoor artifact on a camera exposure of 4 refresh cycles long!

I even rarely see rail-based pursuits this accurate (from a sync-track perspective). The chief problem with your (probably cheap) smartphone camera is the camera artifacts (bowing, fisheye, horizontal lines), but that's simply fixed by newer superior smartphone cameras, combined with optical image stabilization. Which are much easier in fewer attempts.

It just continues to prove my point: You don't always need a rail anymore for high-quality indie pursuit camera tests.

That being said, reviewers and professional testers will need a good higher end smartphone camera (e.g. iPhone 13) to check all boxes (no camera artifacts). But your attempt is great for indie/forum member sharing.

TFTCentral is already using the railless method, AFAIK.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
Discorz
VIP Member
Posts: 999
Joined: 06 Sep 2019, 02:39
Location: Europe, Croatia
Contact:

Re: BREAKTHROUGH: Hand-waved iPhone 13 pursuit camera outperforms rails for rapid field testing

Post by Discorz » 14 May 2022, 06:12

Just tried to improve. Used less fisheyey photo mode with compromise of lower resolution, zoomed in a bit to trim some of the image distortion, still no stabilization but this time with random DIY "rail" made out of glued guides on top of wooden panel which helped with horizontal tracking but not as much with vertical. I reduced refresh rate to 120 Hz for perfect shutter-refresh match to remove flicker banding. Tracking accuracy is still very similar, perhaps a hair better on previous attempts.

IMG_20220514_123959_1.jpg
IMG_20220514_123959_1.jpg (2.69 MiB) Viewed 8001 times
Gigabyte M32Q @120Hz 120 FPS
Aim Stabilizer at middle screen
Overdrive Picture Quality
960 pps
EXIF included

Measured persistence from image approx. 1.2-1.4 ms
Compare UFOs | Do you use Blur Reduction? | Smooth Frog | Latency Split Test
Alienware AW2521H, Gigabyte M32Q, Asus VG279QM, Alienware AW2518HF, AOC C24G1, AOC G2790PX, Setup

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: BREAKTHROUGH: Hand-waved iPhone 13 pursuit camera outperforms rails for rapid field testing

Post by Chief Blur Buster » 17 May 2022, 19:40

That's pretty good.

I also used makeshift wooden rails in many of my early pursuit camera experiments ten years ago. I used scotch tape to make surfaces more slippery and easier to slide on without vibrations, though the tape became rough after about 100 pursuits and had to be repeatedly replaced.

Very ghetto jerryrigged setups until I purchased a sliding camera rail from Amazon that performed better (as long as vise-mounted on top of stiff wood, rather than on wobbly tripods).

I'm amazed how many reviewers use wobbly tripods on carpet, and I can outperform them with an optical-stabilized-camera iPhone 13 handwave!
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: BREAKTHROUGH: Hand-waved iPhone 13 pursuit camera outperforms rails for rapid field testing

Post by Chief Blur Buster » 29 Oct 2022, 01:26

Future: Automatic Smartphone Pursuit Camera App With Automatic GtG Graphing!

Long-term, I want to make a smartphone app that measures pursuit camera images & also doubles as a photodiode oscilloscope (up to 10,000 samples/sec). I've already discovered it is mathematically possible on modern camera sensors now, on all recent iPhone/Galaxy phones.

User Experience
1. A user just waves a camera
2. Camera goes beep
3. Resulting perfect pursuit image + LCD GtG graph!

Behind-The-Scene Algorithm
1. The camera is continuously recording in video mode at automatically preset settings (just like this thread)
2. A neural network AI built into the app will recognize the sync track (pretrained on known good/bad sync track results -- probably a single video file containing hundreds of handwaves is actually adequate training material for a starter app)
3. The app will pay attention to the sync track until it manages to capture a perfect freezeframe within desired error margin
4. The freezeframe is saved. That becomes the officially saved pursuit camera image.
5. To extract LCD GtG, you simply take a pixel row from the camera image (of a moving edge -- you could key on a single tickmark, or I can modify the sync track to include a built-in moving edge convenient for LCD GtG measurements), and subtract perfect linear straight-line MPRT100% from it. Anything leftover is LCD GtG identical to a photodiode oscilloscope!
6. Display resulting photo & display resulting GtG graph (like a photodiode oscilloscope).

A software developer can even use multiple pixel rows in resulting photograph, scanskew-correct it (based on moving-blur-edge tilt in photograph) and challenge the user to zoom in closer with faster pans, to spread more motionblur onto the camera sensor over a single refresh cycle, to get more "photodiode oscilloscope sample rate equivalent". I was able to determine that current smartphone sensors in iPhone 10-13 is now accurate enough to become a 10,000 samples/sec GtG photodiode oscilloscope, via the automated pursuit camera trick above.

And GtG heatmapping could be done on the same photo (simply by displaying multiple moving blur-edges of different color pairs). Heatmap 25 LCD GtGs in one pursuit photograph! As long as the blur edge can be blurred significantly across the horizontal resolution of the photo. Say, the blur edge is spread over about 10% the photograph width, and has an accurate sync track to 1 photography pixel, then a 4000 pixel-wide photo essentially has 400 photodiode oscilloscope samples per refresh cycle. At 144Hz, that's 144*400 = equivalent to a photodiode oscilloscope doing 57,600 samples/sec. Obviously bayer filtering, camera tracking error, lens blur, and image scaling adds error margins, but I am underscaling to 4000x3000 get a tack sharp photo, and from that you can get 5-digit samples/sec photodiode oscilloscope data from a single 1/30sec 4000-pixel-wide photograph!!! Amazing how math can pull this off. And you can get better than 16-bit accuracy (like a good Tektronix oscilloscope) simply by averaging/stacking multiple pixel rows from the photograph -- that's a 3000 pixel tall photograph, representing defacto 3000 separate unique runs of a photodiode oscilloscope to be averaged (stacked) on each other. You'd have to skew-correct the photograph (a simple math matter, since we know all pixel rows represents the same GtG run attempt), but then:

Theoretically iPhone 13 can be translated to a ~10,000 samples/sec 16-bit GtG photodiode oscillscope!
(with these algorithms above).

All it requires would be a custom TestUFO motion test, the only thing an end user does is keep hand-waving along the UFOs until the smartphone until the app goes beep, then the math magic pulls a clean GtG oscilloscope curve from a single camera image!

Amazingly, it appears mathematically clear that an iPhone 13 camera can outperform a single-run low-samplerate oscilloscope on an op-amp'd photodiode (albiet only for display measurements)! Once you've done all the math tricks above.

You could pursuit-photograph multiple moving blur-edges simultaneously of different color-pairs. Then LCD GtG heatmapping is just an app that any user can download! (a GtG matrix measurement from a SINGLE pursuit camera photo).

You're welcome. I believe I'm the world first person to suggest this "iPhone GtG photodiode oscillcope" idea for GtG graphing automatically. Please credit me for the free idea. (And don't patent it, thank you.) Researchers, contact me if you'd like to open-source collaborate on this idea -- www.blurbusters.com/contact
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

Post Reply