iPhone X and FCPX: HEVC or not....

Hey Folks:

Have a project coming up that will incorporate some iPhone X footage. Delivery is 1080 HD for the web.

Assuming we should request iPhone set to record 4K, so we can scale shots in the edit. Beyond that, questions are:

Is there an advantage to capturing in HEVC? And if so, ingesting the straight HEVC as is or should we transcode to Pro Res before ingest?

Or, should we request iPhone X capture to "most compatible," which would be H.264 .MOV? And then transcode to Pro Res before ingest?

Big picture, my understanding is that HEVC is a "smaller/lighter" size file format--but from a quality standpoint, is it better? And if so, is it worth working with the native files until export?

Thanks for your input!

Find out how your iPhone X footage will be shot. We handle a lot of footage shot on various iPhones using the Filmic Pro app ( www.filmicpro.com ).

Filmic Pro allows the user to set capture bit-rates that are higher than the iPhone would normally record. (50mbps at 1080 and 100mbps at 4K for example). The recordings are .mov files in H.264.

We usually transcode these files to ProRes 422HQ as this is our usual workflow. But often recently we just work directly with H.264 in the FCPX timeline and the quality is excellent. We grade in Color Finale Pro and occasionally add mild noise-reduction with Neat Video - all using the H.264 footage.

The higher bit-rate Filmic allows compared to the standard iPhone settings is definitely a factor in maintaining the quality during post / grading etc.

Shoot & edit in HVEC/H.265 native, no problem, if you're on High Sierra. And yes, it is better quality than H.264.

ProRes "HQ" is a bit of overkill for iPhone clips, ProRes 422 is plenty enough. "HQ" isn't really buying you anything extra, considering you're recording really high compression to start with..

Ben is right, 422HQ is way overkill for iPhone footage. I wasn't recommending it - it's just a workflow we use 'cos we deal with a lot of HQ footage recorded on a Samuri Blade.

Shooting H.265 is sometimes beneficial (depending on your subject matter). Having compared both H.264 & H.265 footage side-by-side with drone footage, I can say that sometimes there is very little visual difference. Assuming the bit rate is the same, you can really go either way. I'm not sure if the iPhone H.265 footage adds any additional benefits that my P4P H.265 footage had. Of course, the best bet is always to do camera tests ahead of time, especially with a new workflow like this.

I think what hardware you are using plays a big part in whether to use H265.
Only the latest Macs have hardware decoding/transcoding (Kaby Lake?? 2016 and later??) and without the hardware support, transcoding is much slower than if using the software only in High Sierra. From articles I have seen, it is very slow as software only. So factor that in as well. Might be better to use H264.
HEVC can save about 40% of disk space or increase bit rate for same size so this is great especially if sending anywhere, but if it takes up to 8-10X longer to transcode?? on older Macs?

Just thought I'd mention the hit if no hardware support.
Greg

Thanks, folks.

It won't be a huge amount of footage so I'm not that concerned with file size. Given that we don't usually work with phone footage, my concern is quality.

Seems like the jury may still be out as to whether HEVC is qualitatively better than H264, regardless of file size.

But appreciate the insights and any additional info.

screendor wrote: ...Seems like the jury may still be out as to whether HEVC is qualitatively better than H264, regardless of file size...

The H264 and HEVC both support different encoding bit rates, color bit depths and GOP size. The issue of image quality will be strongly affected by this. At acquisition time these encoding parameters are determined by the camera. So what quality you can expect is more determined by the camera and encoding than by FCPX.

The encoding parameters are not always well documented for a "real" video camera, much less a smart phone. So your question can only be answered by shooting identical scenes with an iPhone X using both HEVC and H264, then (1) examining the playback quality using various playback tools which themselves may inject apparent quality differences, and (2) Inspecting the video files using MediaInfo or Invisor to determine the bit rate and encoding parameters.

For my iPhone 7 running iOS 11.2.5, 4k HEVC is captured at 21.2 megabits/sec, whereas 4k H264 is 50.6 megabits/sec, both 8-bits per color channel.

In particular look for certain difficult encoding cases during playback such as low light, rapid subject movement, rapid dark/light transitions, water in motion, smoke/particles, low light, photographic flash/strobe units, etc.

Examining the 4k HEVC and H264 material from my iPhone 7, it might appear HEVC has a little less low-light noise, which would be consistent with what FCPX.guru said, but it would require more extensive testing to be sure. I just did a quick casual test.

Even this has some uncertainty across various playback methods. E.g, VLC uses all-internal playback codecs whereas other tools may not. Another unknown is whether the playback tool uses hardware acceleration which can affect smoothness or even image quality, depending on the method. So just playing back a few seconds of footage in one player on one machine using one version of macOS might not be revealing.

Note all of this is *before* the material gets to FCPX. The quality of the captured footage is pretty much locked in at that point, so this more a technical cinematography question not an editing question.

However as VidGreg said, you ultimately want to edit this so at that point there can be issues with performance and compatibility. HEVC is extremely compute-intensive so whether hardware acceleration is available can make a huge difference. E.g, on my 10-core Vega64 iMac Pro, FCPX encoding to 4k 10-bit HEVC is 30 times slower than encoding to 8-bit HEVC.

On the decoding side, it appears FCPX 10.4 handles both 8 and 10-bit 4k HEVC pretty well (on the iMP), since transcoding to optimized or proxy takes the same time whether the input file is 8 or 10 bits. How it would be handled on various other Macs could differ -- this must be determined by testing.

However your ability to even evaluate and characterize HEVC -- outside of FCPX -- will vary based on what machine you have, how it's configured, what version of macOS it's running, and how the HEVC file is encoded.

Another tricky area is how you share HEVC content from an iPhone 7 or above. If shared via AirDrop, Messages, Apple Mail or downloaded to a Mac via Image Capture, by default the HEVC will be transparently converted to H264. To keep this from happening, you must change the iOS setting Settings>Photos>Transfer to a Mac or PC and select "Keep Originals". Otherwise during the download process, your HEVC material might be converted to H264 without you even knowing. Read this KB article for details: support.apple.com/en-us/HT207022

In general I don't see the benefit of rushing to HEVC unless someone in your organization is willing to do all the required testing to validate and incorporate this into your workflow. My documentary team has several cameras which can shoot HEVC but each one would have to be tested separately, and for now H264 works OK.

Thanks for the detailed insights and analysis—much appreciated.