H264 vs H265

[F3honda4me]

I took video back to back close up of a few trees, snow, brick etc switching between h264 and h265 (4K 30fps). Looking at the raw video side by side naked eye I can see no immediate discernible difference. So is the h265 really more for post then? Does it add more dynamic range or what is it providing in post that h264 does not?

Also I can edit raw h264 in Sony vegas but not h265. So if I have to convert h265 to something like prores 422 am I just losing what I would have gained using h265 over h264 in the first place? Don't I lose some data converting to prores etc?

 

[Tomas Wangen]

I took video back to back close up of a few trees, snow, brick etc switching between h264 and h265 (4K 30fps). Looking at the raw video side by side naked eye I can see no immediate discernible difference. So is the h265 really more for post then? Does it add more dynamic range or what is it providing in post that h264 does not?

Also I can edit raw h264 in Sony vegas but not h265. So if I have to convert h265 to something like prores 422 am I just losing what I would have gained using h265 over h264 in the first place? Don't I lose some data converting to prores etc?

This is still an open question and I have certainly not made a final choice yet, you will find several discussion on this here already. We can probably expect h265 to have more potential than h264, but Dji need to make some improvements to the encoding and profiles (crossing fingers).

(edit: the reason I mention profiles, is that the none-profile has very few shadow levels introducing banding, this will most likely be mitigated by a future, correctly calibrated dlog profile).

 

[smiller]

H.265 will not add dynamic range, it is intended to provide increased bandwidth, or more accurately in this case allow more data to be transmitted within an established maximum bandwidth. This might result in lower artifacting and/or better resolution especially when there's lots of image movement/changes, but any difference may or may not be visible depending on circumstances. In stationary scenes without movement any difference would be smaller or nonexistent.

Regarding transcoding to ProRes or other intermediate file formats, in theory there has to be information loss but in reality it is negligible. Many/most high end production processes use this technique.

 

[F3honda4me]

H.265 will not add dynamic range, it is intended to provide increased bandwidth, or more accurately in this case allow more data to be transmitted within an established maximum bandwidth. This might result in lower artifacting and/or better resolution especially when there's lots of image movement/changes, but any difference may or may not be visible depending on circumstances. In stationary scenes without movement any difference would be smaller or nonexistent.

Regarding transcoding to ProRes or other intermediate file formats, in theory there has to be information loss but in reality it is negligible. Many/most high end production processes use this technique.

So am I ok to just transcode to prores 422 LT? Or should I be doing SQ or HQ?

 

[Tomas Wangen]

So am I ok to just transcode to prores 422 LT? Or should I be doing SQ or HQ?

Standard quality is sufficient for the p4p output (no visible loss of quality). Reducing sharpness to -1 or -2 will probably help the transcoding as well due to less high-frequency detail/noise (no proof, only theory). HQ files are nice but really enormous.

 

[smiller]

So am I ok to just transcode to prores 422 LT? Or should I be doing SQ or HQ?

There's endless debate on that and of course it depends on individual circumstances. Personally most of the more rational discussions I've seen say that standard ProRes is fine for work with prosumer (dSLR/P4P) level output so I've settled on that. If you want to pixel-peep yourself to insanity you can experiment with your own content though.

 

[Dronefriend]

Prores LT gave me small, but visible artefacting when I converted my P3S footage. I would go for plain 422 Prores with "normal" shots, and if theres anything I'm really satisfied with I would go with HQ. Especially if there's a lot of movement in the material.

 

[Tomas Wangen]

Prores LT gave me small, but visible artefacting when I converted my P3S footage. I would go for plain 422 Prores with "normal" shots, and if theres anything I'm really satisfied with I would go with HQ. Especially if there's a lot of movement in the material.

LT is only sufficient for proxy files and not for final render.

 

[smiller]

Here is what Apple says, might provide a good idea what's what

Final Cut Pro 7 Professional Formats and Workflows

 

[Dronefriend]

LT is only sufficient for proxy files and not for final render.

Yes. What I meant with this post was that on a P3 Standard, plain 422 (called SQ on earlier versions? Approx 200 mbit) was sufficent. This probably means plain 422 is the minimum on P4P material. I would definitely go HQ if it's an important shot. Can't recover once you degrade the quality, better stay safe until final version is exported and delivered.

 

[Tomas Wangen]

I agree. I actually a/b tested some challenging footage (high contrast, plenty of details combined with neutral zones) with the hq and normal, and due to the initial compression of the source-files (which is considerable), there was no discernible difference between the two. I have therefore settled on 422 (non hq) in my workflow to save some space. (also note that LT is great for proxy files for some super-responsive editing).

 

[richr]

I took video back to back close up of a few trees, snow, brick etc switching between h264 and h265 (4K 30fps). Looking at the raw video side by side naked eye I can see no immediate discernible difference. So is the h265 really more for post then? Does it add more dynamic range or what is it providing in post that h264 does not?

Also I can edit raw h264 in Sony vegas but not h265. So if I have to convert h265 to something like prores 422 am I just losing what I would have gained using h265 over h264 in the first place? Don't I lose some data converting to prores etc?

Vegas Pro 14 will handle h.265 in mov but not mp4. I tested both and only the mov container was recognized. And of course, Vegas will handle proxies automatically.

Sent from my Pixel XL using PhantomPilots mobile app

 

[Tron_X]

Vegas Pro 14 will handle h.265 in mov but not mp4. I tested both and only the mov container was recognized. And of course, Vegas will handle proxies automatically.

Sent from my Pixel XL using PhantomPilots mobile app

Do you have Build 201 installed?

 

[richr]

Do you have Build 201 installed?

Yes. Windows 10 and Vegas pro 14 build 201.

 

[Roamer105]

So am I ok to just transcode to prores 422 LT? Or should I be doing SQ or HQ?

420 at 8bit depth is enough. That is what the encoder actually encodes (either 264 or 265 at this point). If you HAVE to transcode, then don't bother using 265, simply go for 264 at 100mb/s.

 

[Roamer105]

Just did a little test. I used some H265 100mb/s video from a P4P (original) and transcoded it to H265 35mb/s with the Adobe Encoder (very good encoder). The resulting video does not show ANY loss or additional artifacts looking at 400x blowup as well as scopes. This hopefully spells good news in that DJI has some headroom to improve the H265 encoder provided the computational resources are available.