Welcome to our community

Be a part of something great, join today!

Register Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

The difference in shooting FF1080P and 3k?

Russ Blaise

Russ Blaise

Well-known member
Joined
Jul 15, 2009
Messages
101
Reaction score
0
Points
0
Location
Tucson, AZ
I'm real new to shooting video, (be easy on me please). I'm looking into getting the Scarlett 2/3" Fixed lens camera. It's says it will shoot at FF1080P and 3k. Can someone explain the real difference between the two. Is this something you can turn a switch to go from FF1080P and 3k? Is the difference in the resolution and quality only? If I shoot in 1080P rather than 3k, is the file size smaller?

Thanks,
 
It's merely a different type of file format -- 3K RAW is an immediate, unprocessed, unconverted bayer-filtered monochrome image off of the sensor (plus Redcode compression) and 1080P is basically 1.9K RGB, a converted color image made from 3K RAW. Resolution would be similar since conversion to color gives you about 3/4 of the original RAW resolution. But 1080P is a video format while 3K RAW is a data file format.

They probably will end up both taking up about the same amount of space/data for recording, though 1080P RGB actually is more data to handle (conversion to RAW to RGB automatically triples the data but you are also reducing the size of the frame from 3K to 1.9K, so that's 3X more data reduced by 1/3 again... but that's assuming the same level of compression, 1080P could use more compression to get back down to the level of 3K RAW). So there's a possibility that shooting in 1080P mode may be more limiting in terms of available frame rates, etc. because it is more processing-intensive in the camera.

You have to remember that RAW is naturally storing color information in 1/3 the amount of data as RGB since RGB has three color channels for one image, assuming the same size frame, pixel-wise.
 
Last edited:
We do not know all of the details behind the proposed specs. It would be resonable to assume however, that the 3k vs FF1080P is a simple translation of the current Red One design.

It takes an enormous amount of processing power to downscale video frame by frame to a new size, real time. Given there is only a limted amount of processing power on the camera, it seems that it takes the same amount of power to create a 3k image at up to 120fps as it does to make a 1080P image at up to 30fps from the original 3k sensor image.

This means that you can capture the full amount of light hitting the sensor at either 3k or downscaled to 1080P. As a contrast, the RED One 'windows' down to the resolution; effectively magnifying your Field of View. Since this is not desireable when trying to obtain wide angle shots, the new idea to down sample, probably made true by the very specific ASICs (those that are putting some delays in the plan) planned for the new design.

Given a 1080p image is less than 3k, theoretically it would be a smaller file size. However, we do not know if the 1080P image will be a R3D file format or something completely different. I can't imagine though, that the camera could scale from 3k to 1080P and encode to R3D at the same time; maybe it can, who knows?

Again, none of this is for sure. We will know more soon.

KO
 
I don't think they are talking about downscaling 3K RAW to 1.9K RAW, which would be very hard... they are proposing a conversion from 3K RAW to 1.9K RGB (1080P).

The Red One already does a quickie in-camera partial debayer to 720P RGB for monitoring purposes only. You wouldn't really "look" at a RAW image actually since it is pre-conversion to color.
 
Russ Blaise said:
I'm real new to shooting video, (be easy on me please). I'm looking into getting the Scarlett 2/3" Fixed lens camera. It's says it will shoot at FF1080P and 3k. Can someone explain the real difference between the two. Is this something you can turn a switch to go from FF1080P and 3k? Is the difference in the resolution and quality only? If I shoot in 1080P rather than 3k, is the file size smaller?

Thanks,
Click to expand...

If the 3K RAW vs. 1080p RGB results in file sizes that are similar, why would you shoot to 1080p? I mean, I guess I could understand if you were trying to optimize a workflow for a fast delivery of recording but I would think that 3K raw capture would give you much more to play with in post.
 
Thanks, David and Kevin. 3k kinda of sounds like shooting Camera Raw with my Nikon D200. It's just a data file that you have to run through a software like Adobe Camera Raw. If I shoot footage in 3k with the Scarlet, do you have to run it through a software to transcode it? And is that idea behind the new version of FCP 7, it can now handle Red? I'm I way off about this?
 
The only reason I can think of is that it'll playback faster and by extension transcode faster.

But the format seems kind of pointless personally I would prefer the transistors be put into something else since:

1) If your app supports .r3d 1080p it surely also supports .r3d RAW. So it's not like recording to something like mxf which pretty much everything plays back. You're still going to deal with r3d compatibility.
2) It's the same data rate so you're losing a lot of image but not making any gains in data.
3) REDRocket debayers in real time.

The only market which it seems to attract is the low-low end which can't afford a REDRocket. But if you aren't getting payed well enough to afford a RED Rocket are your schedules really bad enough to need a RED Rocket?

edit: lol the profanity filter must be enabled to convert any "RED Rocket™" into well... <---- that.
 
Russ Blaise said:
Thanks, David and Kevin. 3k kinda of sounds like shooting Camera Raw with my Nikon D200. It's just a data file that you have to run through a software like Adobe Camera Raw. If I shoot footage in 3k with the Scarlet, do you have to run it through a software to transcode it? And is that idea behind the new version of FCP 7, it can now handle Red? I'm I way off about this?
Click to expand...

You wouldn't transcode it exactly. You'll use a workflow that keeps the RED r3d files all the way through the post process. At the very end, your final edited, corrected output file will be encoded from the original captured r3d files. By not transcoding up front in post, you maintain all the raw goodness in the r3d file.

Yes the new FCP supports RED files. So does the old FCP. Download the software yourself and check it out:

http://www.red.com/support

You'll probably need some RED footage to really check it out though.
 
We don't know what the 1080p file format will be yet. But it should output a standard encoded RGB format via HDSDI or HDMI to fit into normal HD video production work flows more efficiently without rendering or transcoding first.
 
Aaron Newsome said:
Yes the new FCP supports RED files. So does the old FCP. Download the software yourself and check it out:

http://www.red.com/support

You'll probably need some RED footage to really check it out though.
Click to expand...

Hey Aaron, Do you know anywhere I can download some RED r3d test footage to try out? I'd like to get familiar with working with Red footage before jumping in... Also I do have the new FCP 7.

Thanks :smile5:
 
My hope for the 1080P mode is that it would allow for you to push the sensitivity up and or maybe squeeze some more dynamic range from it, while giving you less noise.

Not to mention making things a lot smoother for things that have to be delivered in 1080P, since the camera is doing the conversion for you, you don't have to spend time rendering out the 3K source.
 
I really put my foot into now... This is way over my head but I'm having fun.

I downloaded a RED r3d file that was 615.5MB. I didn't know how to get it into FCP 7 so I opened Log and Transfer and dropped the file in there. Clicked "Add Clip to Queue" and in a few it showed up in my FCP project. Dropped the file into the timeline and it was fine. Now, it shows up as a Apple ProRes 4444. Is that still a RAW RED file or by doing it this way it got converted?
 
Don't forget that the 1080P is FF and that it is supposed to cover the whole sensor. Because of this I don't think it will need to be recorded in 3k and then downconverted, it will likely just have larger pixels.
 
Russ Blaise said:
I really put my foot into now... This is way over my head but I'm having fun.

I downloaded a RED r3d file that was 615.5MB. I didn't know how to get it into FCP 7 so I opened Log and Transfer and dropped the file in there. Clicked "Add Clip to Queue" and in a few it showed up in my FCP project. Dropped the file into the timeline and it was fine. Now, it shows up as a Apple ProRes 4444. Is that still a RAW RED file or by doing it this way it got converted?
Click to expand...

Hmm. I would think that doing a log and transfer would transcode the file and leave you with a ProRes file to work with. Unless there is a way to conform your final delivery file, using the original r3d, I don't think this is the right way to do it.

Since I've never seen an r3d file (I don't own a RED), I can't say for sure. Hopefully someone with experience will comment on the proper workflow.

I'll tell you what,... as soon as the Scarlet is released, I'll tell you what I find.
 
Wesley Scoggins said:
My hope for the 1080P mode is that it would allow for you to push the sensitivity up and or maybe squeeze some more dynamic range from it, while giving you less noise.
Click to expand...

Been thinking about this one for a few minutes now. Trying to wrap my head around how and why this would be possible?

I can't think of any logical train of thought that would allow me to understand how and why shooting in 1080p mode would make the Scarlet sensor more sensitive and have more dynamic range.

Help me understand this train of thought ...
 
Averaging pixels = less noise.

Less noise = more effective sensitivity in the shadows (no help in the highlights).
More sensitive = more stops of DR.

But you would get the same effect from doing it in post as you would in camera.
 
Gavin nailed it. But yea, because you can turn up the sensitivity of the Red, you get a noisy image, but compress it down from 4K to 1080P, and you're able to scale it down and get a cleaner looking picture.

The same idea applies to the Scarlet, only what would REALLY be neat is if it could be done in camera, so you turn on FF1080 mode, and are able to hike up the sensitivity and hack down the noise. Then just download a very clean looking 1080P file.
 
Russ Blaise said:
I really put my foot into now... This is way over my head but I'm having fun.

I downloaded a RED r3d file that was 615.5MB. I didn't know how to get it into FCP 7 so I opened Log and Transfer and dropped the file in there. Clicked "Add Clip to Queue" and in a few it showed up in my FCP project. Dropped the file into the timeline and it was fine. Now, it shows up as a Apple ProRes 4444. Is that still a RAW RED file or by doing it this way it got converted?
Click to expand...


It won't be a RED RAW file, it's now a quicktime with the apple prores codec. Basically now it's converted to an UNCOMPRESSED RGB file, no longer a RAW.
I still work off FCP6, so unless Apple changed this (which i doubt) it should still be the same type of file in 7.
 
You must log in or register to reply here.
Back
Top