Cinema P3 Pro Camera for iPhone

I thought I'd open a new thread regarding this nice app instead of mixing the discussion with the one about BM's camera app. Seems a bit unfair and misplaced. There are pros and cons for either one.

Two features I like very much:
– You can change the data rate for HEVC to values even higher than in many expensive hybrid cameras.
This will give you a much richer signal structure for grading than the default and it's a great alternative to ProRes HQ.
– You can get a full screen waveform, and many other scopes too. My ageing eyes like that.

I’m commenting here to keep track of this thread. I am testing the Cinema P3 Pro app along side the BMD Camera App. There are things I like about both of them.

For sure. BM's app is great for teams, CP3 for the 'lone wolf', IMHO.

Unpopular opinion (as Marc says) - I like the quick and time limited way you can bring up false colors and other full-screen tools in Cinema P3 - I find it takes just a bit too long to do it with the BMD Camera App. Not a bit deal but a nice UI shortcut. I much prefer the ISO control on the BMD App - the way you can w quickly snap to 100, 200, 400, 800... and then fine tune if necessary is really great.

It's just the other way around for me.
I like to fix the exposure time at 180 degrees, bring ISO into the ballpark with ND and let CP3 fine-tune the ISO. I want to avoid any change of the shutter, to neither loose the EDR nor get into stutter territory.

True. I use the quick presets to jump to the approximate ISO and then use the variable ND to dial in the exposure (I also keep my shutter at 180 degrees). Full manual on exposure.

Time-lapse

Did another test today. All open apps closed, then restarting the phone.
– Doing a time-lapse in Apple's app, intentionally using HD in 30 fps. Resulting in 8.834 Mb/s at 30 constant fps.
– Then a time-lapse in CP3 without touching presets. Resulting bit rate 322 Mb/s at constant 25 fps in UHD.
– CP3 again, intentionally changing the bit rate: 201 Mb/s at constant 25 fps.
– Now a time-lapse in the BM app. Result: 5.733 Mb/s at constant 24 fps in UHD.
– Another one in CP3: 201 Mb/s at constant 25 fps.

So, I can't really see the issue here. Any step I missed that triggers it?

I use Cinema P3 on my iPhone 14 Pro and prefer this app over Filmic Pro.
Cinema P3 C-LOG looks much better for me than Filmic LOG, especially with the corresponding LUT.

I would ask the iPhone 15 Pro users here to do me a favour:
Could you please make a comparison of Apple LOG vs. Cinema P3 C-LOG?
This would be very interesting for users of "older" iPhones.

I usually make my films with Cameras from Sony or Fuji. But the sensor size of the iPhone 14 Pro (which did not change in the iPhone 15 Pro) and all the other features encourage me to hire the iPhone as well.
But the above mentioned comparison would be very interesting for me and others.

Kind regards,
Hartmut

I'm guessing CinemaP3's C-Log is not Canon Log but some other home-grown log curve, is that true? Can you use a CST in Resolve to convert this CinemaP3 C-Log or must you use one of their LUTs?

It’s a home brew, for which the app supplies LUTs. There’s no CST.

Regarding time-lapse issues as described by Kurt:
I did a time-lapse in CP3 today after opening, but not using BM Camera (and not quitting it).
Recording went smooth for close to three hours, got a stable time-lapse at constant(!) 25 fps, 200 mbps in HEVC 10 bit 4:2:2.
BTW, the charge went from 80% to 28% over that time, and I had frame averaging on (internal recording).
Those beasts have stamina!

Regarding time-lapse issues as described by Kurt:
I did a time-lapse in CP3 today after opening, but not using BM Camera (and not quitting it).
Recording went smooth for close to three hours, got a stable time-lapse at constant(!) 25 fps, 200 mbps in HEVC 10 bit 4:2:2.
BTW, the charge went from 80% to 28% over that time, and I had frame averaging on (internal recording).
Those beasts have stamina!

As I said in the original “test giving weird results”: “There’s an element of chance involved”, and I had to do “mindbending” operations, involving reboots to make sure, what happened and when it was caused how.

As you yourself discovered, deviations from the straight and narrow, also had - accidentally - happened to you, although to a lesser degree.

If you just chug along in CP3 chances are, you will not experience the problem.

If you look at the megabyte/second value for the particular recording, if in doubt, it is relatively easy to discover - over time - IF something went haywire and results are significantly below settings and your usual results in similar situations (then obviously very hard to remember, if and how anything else might have affected the outcome).

When I look in Settings, Security under Camera, 12 out of 22 Apps that have requested access have it (very few Apps have been uninstalled in my New Year clean-up). Two are strictly for Danish use only. There are now only 11 Apps out of 22, that need access to the Camera in my personal use cases - Panasonic’s access was completely superfluous as was the case for some other Apps. Access disabled after todays review.

Regards

Regarding time-lapse issues as described by Kurt:
I did a time-lapse in CP3 today after opening, but not using BM Camera (and not quitting it).
Recording went smooth for close to three hours, got a stable time-lapse at constant(!) 25 fps, 200 mbps in HEVC 10 bit 4:2:2.
BTW, the charge went from 80% to 28% over that time, and I had frame averaging on (internal recording).
Those beasts have stamina!

If you need more recording time, there is a simple solution, if you have the Apple MagSafe or any other wireless charging device at hand, connected to a powerbank and/or an USB-C charger.

This will also leave your USB-C connection free for those long endurance ProRES 422(HQ) recordings usually targeted to external SSD's.

I use one setup with my iPhone 15 Pro in a cage and Apple's wireless charging device connected to a powerbank in my small shoulder bag (or hung in a small pouch in a hook under my tripod, if I'm not using my full hub with internal SSD).

Recordings are typically made with my Zoom M3 (haven't met the much "advertised" side effects yet), that produces both standard LPCM stereo and "RAW" M-S (48kHz, 32-bit float) recordings on device. It has internal batteries, and no need to be connected to anything else. I use the M3 with the Sennheiser MKE 600 Dead Cat (MZH600). The Røde VideoMix NTG Dead "Wombat" is too short leaving the side channel open for extreme wind noise in even slight wind. I guess some YouTubers didn't deign to test their own advice in real life outdoor situations

Only needs syncing in post (especially simple in post in FCPX, since the my latest Davinchi Resolve Studio 18.6.4 build 6 seems to have some "quirks" for syncing audio - in my case especially M-S it seems). YMMV.

This will leave the USB-C completely free for external SSD use, and allow both high(est) quality audio and simultaneous external wireless power supply from any available source.

If I use the Acasis hub with internal SSD, I usually connect the USB-C output from the microphone/recorder directly to the hub, and then select Blackmagic 32-bit float input (not possible in Cinema P3) for inclusion of the 90 or 120 degree wide stereo recording directly into the ProRES/HEVC file, and still having access to the full Mid-Side RAW original recording (sometimes only using the mid channel shotgun signal from.the Mid-Side recording overlayed in post, where this has special importance).

Regards

Of course I also use a MagSafe power source when needed. And Acasis is building pretty nice hardware for sure, it should be a good suggestion for others around here.

To show what I mean by HEVC data rate being too low in Apple's and BM's camera app (and to take the discussion to the right thread, as Robedge suggested), some waveforms.

The scene was pretty much the hardest one you can use to torture a GOP codec: heavy rain on a backlit water surface, filling the screen. I underexposed by one stop intentionally and compensated for exposure and the low contrast of log with the same grade for both. Using PNGs not to loose detail, so these are not attached, but linked on my DropBox.

ProRes:
https://www.dropbox.com/s/ospv9wt7iad2e ... g.png?dl=0

HEVC:
https://www.dropbox.com/scl/fi/n9e275hw ... z4h8i&dl=0

ProRes HQ is around 85 MB/s. In Cinema P3 you can increase the data rate for HEVC, which is originally only around 4 MB/s, to more than 40. You won't need to go that high, though, since for most scenes it'll be far more efficient than ProRes. Do some experimentation to find your optimal setting (and report, if you like).

I'd be glad if BM could add such an option to their camera app, since it has its own advantages, for teamwork in particular.

Uli, at those links I don't see any actual video, just the waveform scopes, which are informative. I do see typical "combing" that results from manipulating low bandwidth images. I upped my h.265 bandwidth to 4 mbps, but I've not had time to seriously investigate the results. Too busy managing the fifty-year record low temperatures here in western Canada.

In other news, time lapse recordings were failing whenever I walked out of Bluetooth range. After I did some quick tests confirming that it was my hearing aids that were causing the problem, Tom fixed it and uploaded an update to the App Store. He's nothing if not responsive.

My friends in Portland told me that temperatures are about to rise. Hope the same for you.

4 mbps is far too low, or do you mean 4 MB/s? That's a factor of 8. Still a bit low, I'd go for 20 or 30 for UHD in 50/60 fps.

I don't know if this is really helpful, but here's a still from the video. Just found a scene that is one of the hardest for any GOP codec:
https://www.dropbox.com/scl/fi/rntazw0k ... fpxeh&dl=0

And Tom is doing a great job, if only all developers would be that responsive!

4 mbps is far too low, or do you mean 4 MB/s?

Thanks for keeping me on my toes, Uli. It's set at 8 MB/sec.

Too effing horrid to go out and shoot tests. It's still in the minus teens C and it's been snowing for nearly 24 hrs. straight.

8 MB/s should suffice for most scenes. The HEVC encoders in the A17 are really good.

Hopefully someone can give me some guidance…using the Cinema P3 Pro Camera app I recorded Apple Log footage. I am much more comfortable with the editing workflow in LumaFusion, but when I try to use an Apple Log to 709 LUT the colors come out looking wonky. If I drop the same footage into Resolve for iPad, and change the input gamma on the CST to Apple Log it looks great. Any ideas why LumaFusion is behaving that way with Cinema P3 Pro Camera Apple Log footage? It does fine with Blackmagic Camera app Apple Log footage.

LUTs are not as good as a CST. Where did you get your LUT?

LumaFusion comes with some AppleLog LUTs, and downloaded several others including Prolost, etc. Would it be that LumaFusion just doesn’t have the built-in color space option for AppleLog? If I add Cinema P3 Pro Camera DLog to 709 LUT into LumaFusion, then add contrast and saturation, it looks better.

LumaFusion comes with some AppleLog LUTs, and downloaded several others including Prolost, etc. Would it be that LumaFusion just doesn’t have the built-in color space option for AppleLog? If I add Cinema P3 Pro Camera DLog to 709 LUT into LumaFusion, then add contrast and saturation, it looks better.

If you look into a file produced in Apple Log (by Apple Camera or Blackmagic Camera), the Colorspace is BT.2020 and NOT the usual BT.709, as seen here:

Mediainfo data on an Apple Log recording made by Blackmagic camera earlier today.

Skærmbillede 2024-07-30 kl. 22.12.33.png (511.52 KiB) Viewed 5239 times

That may have an effect on editing tools, that may assume the most often used approach, like this:

Mediainfo from standard downsampled h264 internal recording in Linux S5 (no internal ProRES available - settings otherwise similar).

Skærmbillede 2024-07-30 kl. 22.19.25.png (366.08 KiB) Viewed 5239 times

from mty Panasonic Lumix S5 (internal V-Log recording).

Regards