- June 3, 2021 at 12:26 am
I recently had an 8K Red Camera project where I was able to push in to an interview shot pretty radically – in Frameflex it was down to 40, and in Resize it was 230! Since it was a standard 1920x 1080 delivery for the web, and no one saw any degradation of the image, everyone was happy.
Maybe the image was technically fine – who knows. The question is: how far can I take this in the future when I will be delivering to a good streaming service? If we shoot 4K, how far can I push in if they want a 2K delivery instead of standard HD? If it’s 8K again? Is there a mathematical formula I can run?
More and more I’m being asked to push-in on hi-rez footage – what is your workflow for keeping an undegraded image?
If there’s a thread in another forum about this, please let me know.
- June 3, 2021 at 6:50 am
Just stick to frameflex and avoid using image resize.
That way you are using whatever source pixels you have.
Once you have applied frameflex to a source clip you can key frame it in the timeline.
- June 5, 2021 at 2:47 pm
There is no perfect mathamatical formula. Most times i’m asked to punch in on a 4k or higher shot, I can start to tell fairly quickly that the focus isn’t perfectly sharp, so i try to do it as little as possible. When i hear the DOP or Director make the promise of punchins, I remind them that they will see quite quickly a soft focus issue when they do a punch in. Much less of an issue with scenics, but I can usually spot an interview with a punch in pretty easily.
And ofcourse you also must remind them, and yourself, if they ever want to deliver at 4k, they are no longer puching in, but blowing up the shot. Most broadcasters and delivery options i’m supplying are asking for 4k, not 2k.
- June 16, 2021 at 5:13 pm
You have to punch the numbers of the resolutions. Which is hard when terms like ‘HD’, ‘2K’ and ‘4K’ are thrown around, without specifying what the actual resolution is. Such terms are just umbrella terms for describing a range of resolutions.
HD usually means 1920×1080 (or ‘Full HD’). But sometimes people call 1280×720 HD also.
2k can mean either Native (2048 × 1080), Flat cropped (1998 × 1080), or Cinemascope cropped (2048 × 858). I’m going to assume that if you’re delivering for web, they want Native. But, it’s best to confirm with the client exactly what resolution they want.
4k can mean either Full Frame (4096 × 2160), Flat cropped (3996 × 2160), or Cinemascope (4096 × 1716). You should check with the DOP exactly what resolution they are shooting.
Once you have confirmation of the exact pixel resolution of how youre shooting and how you’re finishing, then you can punch the numbers. So lets assume they want a 2048 × 1080 delivery, and the DOP is shooting 4096 × 2160. Shooting resolution divide by finishing resolution is 2, meaning that you can scale the footage up by 200% without loosing any quality at all.
The general rule of thumb is that you can scale up an additional 10-20% and the quality loss will be negligible (ie. imperceptible to the naked eye). Although I personally don’t love doing that because even if quality loss is imperceptible, it still might compromise the colorist/VFX artist’s ability to work, also that you never know how someone later down the track will compress your film, and any compression will be aggravated by quality loss.
If you’re finishing a HD/2K format, and shooting 8k, your laughing. 8k includes a bunch of different resolution which I can’t be bothered listing, but it genrally just means a pixel width of approx. 8000 pixels. So you could essentially scale up 300-400% or thereabouts without any quality loss – this is approximate, and again you need to confirm the precise resolutions before getting precise calculations.<font color=”#888888″>
- June 16, 2021 at 8:35 pm
Thanks, this is just what I needed, and so clearly explained.
Log in to reply.