Red Scarlet Vs. 5D3 Raw Sharpness Comparison

R

Ryan Lightbourn

Well-known member
Many people have been commenting on the 5D3 not looking sharp enough, so I put this together.

I was swapping cameras & lenses back and forth, so conditions were extremely similar for both.

For more accurate results, hit the 'Download' tab and download 'Original .MOV file'
https://vimeo.com/69803484
 
Ryan Lightbourn said:
Many people have been commenting on the 5D3 not looking sharp enough, so I put this together.

I was swapping cameras & lenses back and forth, so conditions were extremely similar for both.

For more accurate results, hit the 'Download' tab and download 'Original .MOV file'
https://vimeo.com/69803484
Click to expand...

Always love people doing tests, but I've found red Scarlett , or any red for that matter, varies with debayering and compression settings. Great job, but I still think 4k is going to win if handled correctly every time.
 
Ryan Patrick O'Hara said:
Always love people doing tests, but I've found red Scarlett , or any red for that matter, varies with debayering and compression settings. Great job, but I still think 4k is going to win if handled correctly every time.
Click to expand...

For a 4k timeline/delivery or when shots need to be re-scaled, absolutely.

But keep in mind the 1080p raw is being downscaled 50% from a 4k-equivalent area of the 5D's sensor. If you look at Redcode 4k @ 100% crop, and 5D3 raw stills @ 100% crop, the 5D3 stills are much, much sharper. With a 1080p timeline/export, the 5D3-raw in theory should be sharper than Redcode raw @ 50% scale no matter the debayering process.
 
Plus if you use post sharpening, working with more information, like Scarlett, would likely make Red's ability to go sharper much much higher, probably well past what you'd even want! My .02, and I've only played with raw 5dm3 once.
 
I have to say this is a win for the 5DMKIII owners. I would have never guessed when this camera hit the market that it would be in a sharpness test with the RED Scarlet. Really great to see the images coming from RAW video.
 
Ryan Patrick O'Hara said:
We talking about video or stills? Of course the 5dm3 stills will be sharper! Haha
Click to expand...

That's my point...the 5D3 raw video @ 1080p in a 1080 timeline should no doubt be sharper than Red (it's essentially a series of raw stills). I'm not saying it's better than Red...just an observation that I tested and seemed to be true. A 4k timeline would be a different story.

I'd never sharpen R3Ds in a 1080p timeline...they're already plenty sharp enough downscaled. If anything I may reduce sharpness when working with 5D3-raw depending on the subjects in my scene. Oversharp images look too digital IMO.
 
Scarlet imo was never really that great, Red Epic is a different monster.
Also remember the 5D does not have an optical low pass filter, so on paper it should be sharper, heck using GH3 with Panasonic lenses is so sharp it can cut you......

BMCC is so sharp for this very reason as well
 
Last edited:
Ryan Lightbourn said:
Yes it does. And GH3 video with those panny lenses looks absolutely horrible...do the lenses have some sort of sharpening filter built in?
Click to expand...

Well I guess that's an opinion...... The Panasonic lenses do a lot of correction though.

Now regarding the footage they both look great, it seems though that the 5D is clipping a little on the rooftops where the Scarlet retains highlights better judging by your video, really nice footage though an really just content in general. From the videos to this, nice.
 
Ryan Lightbourn said:
That's my point...the 5D3 raw video @ 1080p in a 1080 timeline should no doubt be sharper than Red (it's essentially a series of raw stills). I'm not saying it's better than Red...just an observation that I tested and seemed to be true. A 4k timeline would be a different story.

I'd never sharpen R3Ds in a 1080p timeline...they're already plenty sharp enough downscaled. If anything I may reduce sharpness when working with 5D3-raw depending on the subjects in my scene. Oversharp images look too digital IMO.
Click to expand...

essentially a series of RAW stills? I will abdicate from this because that really wasn't how I thought it worked.

Ryan Lightbourn said:
Yes it does. And GH3 video with those panny lenses looks absolutely horrible...do the lenses have some sort of sharpening filter built in?
Click to expand...

I do however know this is not how that works! :)
 
mcgeedigital said:
Wrong. They both have the same sensor. I owned both. At 4k they are identical.
Click to expand...

Well the Epic can Handle higher Data rates , therefore allowing better compression. Also the Epic shoots at 5k which is very different beast than 4k. I have shot and used both cameras (not own) + handled work flow.
 
J.F.R. said:
it seems though that the 5D is clipping a little on the rooftops where the Scarlet retains highlights better judging by your video, really nice footage though an really just content in general. From the videos to this, nice.
Click to expand...

Appreciate it. Rooftops didn't clip at all though...I brought the contrast down a bit and re-upped the video so that you can see. If I bring contrast down even more, all detail is in the roofs in both 5D & Scarlet shots.

And dude, the Scarlet is an absolutely Epic camera (pun intended)...no clue what you're talking about there. I've used every popular sub-$100k camera under the sun (aside from Alexa) and Scarlet is easily in my top 5.
 
Ryan Lightbourn said:
Appreciate it. Rooftops didn't clip at all though...I brought the contrast down a bit and re-upped the video so that you can see. If I bring contrast down even more, all detail is in the roofs in both 5D & Scarlet shots.

And dude, the Scarlet is an absolutely Epic camera (pun intended)...no clue what you're talking about there. I've used every popular sub-$100k camera under the sun (aside from Alexa) and Scarlet is easily in my top 5.
Click to expand...


I have a project i worked on that never released that was shot on the Scarlet at 2k-3k. Personally I think the Scarlet at 2-3k looks horrible. Also for whatever reason the DP on the shoot didn't have too much light and the footage was insanely filled with grain. I know a 5D in the same situation would have been better in low light with less noise.


After that project I just never really liked the Scarlet again. Now on the surface yes, if it were lit correctly I wouldn't have had that problem but man even using the 5D in clubs with no lights it didn't look that bad. Also the scenes that were properly lit and shot at 2k looked very soft.

I need to if anything work more with the Scarlet but every other project after that we have used the Epic (which I love).


Now looking at the video again, the 5D Raw really looks nice. How does it compare to non hack (please be 100% honest)?
thanks

Also have you used the BMCC? I just did a simple shoot in the park with the BMCC and loved the image.
 
J.F.R. said:
I have a project i worked on that never released that was shot on the Scarlet at 2k-3k. Personally I think the Scarlet at 2-3k looks horrible. Also for whatever reason the DP on the shoot didn't have too much light and the footage was insanely filled with grain. I know a 5D in the same situation would have been better in low light with less noise.

Now looking at the video again, the 5D Raw really looks nice. How does it compare to non hack (please be 100% honest)?
thanks

Also have you used the BMCC? I just did a simple shoot in the park with the BMCC and loved the image.
Click to expand...

FYI, any Red MX camera sucks donkey balls @ 2k. 3k is ok...not as good as 4k-5k obviously, but 2k is just terrible. And it sounds like the DP was lousy...any camera will look bad in the hands of an inexperienced DP.

I bought the 5D3 strictly because of the raw hack. I haven't really used DSLRs since 2011...they always looked cheapish to me, but once the raw hack was announced I figured FF-raw was a no-brainer @ $3300. I accidentally shot .MOV for a few shots while I was in the Bahamas...it's no comparison to raw whatsoever. Slightly better than Mk2, but same old DSLR look. As Steve said, night & day when shooting raw.

I used the BMCC 2.5k version once. Moire and aliasing aren't so good, low light isn't so good, no 60p, and I'm not a fan of the 2.4x crop. Initially, it was a great way to get super cheap raw (with some side-effects), but IMO will be obsolete soon with the 5D-hack & 4k version coming out. The 4k has some exciting specs, so time will tell how that works out.

All-in-all these are exciting times to be a filmmaker. The point of this video was simply to stir up excitement about the ML team's revolutionary achievement. Lot's of butthurt going on...it's simply unnecessary.
 
Last edited:
@ JFR Why would you shoot a job @ 2k instead of 4K ? Was it for some upped frame rate or the DP only had B4 lenses ? Very telling the "DP" had minimal light then decided to enhance the effect @ half the resolution. IMO the "DP' = minimal experience.

As others have stated.........Scarlet = Epic , Epic = Scarlet there is no difference except in frame rate choice & FOV. Intercut between the 2 cams on a regular basis they look identical.

5D iii now with hack is back in the game, apart from the ridiculous workflow & that criticism is often levelled at Red (another myth).
 
Last edited:
You must log in or register to reply here.
Back
Top