Welcome to our community

Be a part of something great, join today!

  • 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 :)

Crop Factor Demystified

what about light sensibility when cropping? I heard that you will get more noise when you only use a part of the sensor.

Sensitivity remains the same, you just record less pixels which means you will be "pixel peeping" a lot more. Meaning, if you record half the pixels on the sensor you are watching them twice as big on the monitor, and when you do this, with any camera, it becomes easier to see the noise. Good thing is that high resolutions make noise look much more organic and pleasing. Since RED offers SO much resolution you still have a ton left when you crop, even if you crop in really hard. 3k is WAY bigger than 1080p. And with such a clean sensor you won't have much noise anyway.
 
Wow Paul, Jesper & Roberto big round of applause, this helps a lot to nube Red Camera owners like myself. Great visual descriptions to a highly technical issue, even a dumb director like me can understand Thanx
 
Imagine that I have a project set for HD delivery (1920x1080). Would the image quality "feel" any changes if I worked with different crops? I could work the "tele" shots with 3K, even 2K if faster frame rates are needed or HDRx. Then, do only the wide shots in 4K, in order to use traditional wide lenses. The data amount would be reduced considerably, but, at 1920x1080, I guess no one would suspect the difference...

Thats the way I am thinking too, espacially I want to film Action Sports where I need a lot of teleshots and high framerates.
So while filming a lot of stuff in 60fps and 2K I think a Canon 70-200mm lense would be enough, wouldnt it?
 
what about light sensibility when cropping? I heard that you will get more noise when you only use a part of the sensor.

the short answer is: no difference in light sensitivity.

When you crop in, nothing physically changes to what the photosites on the sensor are seeing. It's just which ones the signal processor(s) in the camera choose to sample for data. So whether you shoot a cropped region of the sensor or shoot 5K and then crop in post, there is no difference. That's assuming all other factors being equal -- as in you're using the same lens, setup, etc..
 
Imagine that I have a project set for HD delivery (1920x1080). Would the image quality "feel" any changes if I worked with different crops? I could work the "tele" shots with 3K, even 2K if faster frame rates are needed or HDRx. Then, do only the wide shots in 4K, in order to use traditional wide lenses. The data amount would be reduced considerably, but, at 1920x1080, I guess no one would suspect the difference...

think about it like this: you're performing your crop before you record rather than afterwards. your quality is essentially going to be the same either way, but you leave more options for post by shooting higher resolution and you leave more options for shooting by going ahead with the crop. technically, once everything is scaled down, you may have to do some noise management and sharpening to match everything up, but people have been doing this for awhile. just keep in mind that oversampled footage (i.e. 4k scaled to 1080) is going to be sharper and have smaller, less noticeable noise/grain than footage captured at 1080/2k.

Thats the way I am thinking too, espacially I want to film Action Sports where I need a lot of teleshots and high framerates.
So while filming a lot of stuff in 60fps and 2K I think a Canon 70-200mm lense would be enough, wouldnt it?

i'd say the 70-200 will be a good choice. if you don't already have one, i'd definitely recommend the image stabilization. f2.8 vs f4 is up to you and your needs/budget.
 
the short answer is: no difference in light sensitivity.

When you crop in, nothing physically changes to what the photosites on the sensor are seeing. It's just which ones the signal processor(s) in the camera choose to sample for data. So whether you shoot a cropped region of the sensor or shoot 5K and then crop in post, there is no difference. That's assuming all other factors being equal -- as in you're using the same lens, setup, etc..
The sensitivity issue does rear its head when you oversample, however. Taking a full 5K image and oversampling to 1080 will net you a less noisy image than acquiring at 2K and oversampling. So there are inherent quality issues related to cropping a sensor.
 
BIG HELP. Thanks, Jesper.

Scarlet-X%2BCrop%2Bfactor%2Bsheet.jpg
 
This is why I can't use a Scarlet X : at 4K, too many data to store, at 3K it "kills" any good wide lense and 2K is not usable (just imagine the lense and stabilisation you'll need) . I need really wide shots.

if 4k is too much for you, RED is not the place to be. funny thing is, the filesizes aren't all that bad. do what you will, buddy - but be careful when you start making excuses... that's the kind of thing that kills growth and development. i don't want that for anybody. this camera is meant to shoot 4k, and only be windowed down when necessary for framerates. you can get your wideshots. look into an external recorder, but odds are your files will be about the same size in the end, lol. i will highly recommend shooting 4k, no matter what your output. blanket statement, break it if you need to.
 
Quick questions about external recorders: do you get a flat image, or can you send to the HDMI output, a RedLog version of what you are recording, in order to have an approximate of what you want?
 
Thanks, interesting thread.

One of my chief concerns is data management for some of the work i do. I often shoot documentaries in remote locations without a crew, and upwards of two hours footage a day. Shooting red is a bit scary because I don't have 10k to drop on cards, and transferring off to a computer or drive is too much (laptops need power too). The external recorder option is great. I can just switch for the times I need to conserve, and shot 4:4:4 HD on CF cards. I'll shot 4k for any of scenic stuff, wildlife volcanos etc... If I can get away shooting host and interview stuff in it, I will. Drives are cheap, and I can convert anything else to a standard format like DNXHD 175 to edit with. 10 years from now I can do a new 4k master.

Can anyone tell me who masters for TV, do you do a standard color grade in REDCINE, then convert to Proress or DNXHD and further tweak it in Davinci or Color? Or do you work in your r3d files?
 
Wouldn't it be great to have a recorder that can do 2 or 3k REDCODE from SDI so if we wish at time shoot a lower image size (to conserve space) we can do it to avoid crop factor, but keep the RAW format?

Er... Totally impossible. As soon as you move to a scaled image, it becomes processed and can no longer be RAW. As PaulHerrin said above, this camera is meant to shoot 4K. You only window (crop) down to lower resolutions on the occasions you need the extra frame rates. This is the way it works with the RED One and EPIC as well. People, well, most people know know what they're doing, don't run around shooting 3K or 2K just to save space. The 4K REDCODE files are compact. Most people don't realize it, but 4K REDCODE is comparable in size to recording 1080p ProRes. If you're concerned about saving space, this is probably not the camera for you. And storage space after the fact -- HDDs, LTO tape, etc.. is cheap. OK, current supply issues are another matter, but it's a temporary situation. And the cost of an external recorder that does 1080p ProRes 422 is comparable to a 128GB RED SSD. One that does ProRes 4X4 will run you at least as much as much as a 256GB SSD and actually a bit closer to the price of two of them. 2 x 256GB SSD's will give you 4 hours of record time at 4K 2:1 24fps REDCODE 8:1...
 
Most people don't realize it, but 4K REDCODE is comparable in size to recording 1080p ProRes..

Then what I really need is a way to record the 4k data externally. It is a long story that I won't bore anyone with now, but I MUST record all of my data on an external high-performance DVR. Is there any way to do that and still get the 4K/redcode (some extra module, for example)?

If it is indeed true that the hd-sdi data are scaled to 1080 from the full sensor, then I would probably be happy with that, but I would need some official word from RED, or actual users, before deciding to go that route (I know, that answer will come in a few weeks, which is fine. Whenever someone knows that for sure please post a thread about it!)
 
Then what I really need is a way to record the 4k data externally. It is a long story that I won't bore anyone with now, but I MUST record all of my data on an external high-performance DVR. Is there any way to do that and still get the 4K/redcode (some extra module, for example)?

Not exactly sure why you would need to record externally... Unless you want 4K continuous feed for live events or something. That isn't possible at this time, perhaps with a future module offering multiple SDI or HDMI outputs or perhaps a 10G connector. And then, if you're going to record 4K output, do you want RAW recording or do you want RGB? But yeah, we'll need a different module for that. As REDCODE becomes more and more mainstream, I would like to see a REDCODE output module where the R3D stream could be captured by external recorders, switchers, etc.. for use in live feed situations and eventual Quad-HD / 4K broadcasts. It's all headed that way, but we're just not quite there yet.

If it is indeed true that the hd-sdi data are scaled to 1080 from the full sensor, then I would probably be happy with that, but I would need some official word from RED, or actual users, before deciding to go that route (I know, that answer will come in a few weeks, which is fine. Whenever someone knows that for sure please post a thread about it!)

Don't know what sort of "official word" you're looking for, but the brain has both HD-SDI and HDMI outputs. Both capable of 1080p and 720p output modes. Currently 23.98, 25 or 29.97 fps. No interlaced. No PSF. EDIT > This is referring to EPIC. Scarlet is the same camera, but with less internal bandwidth and capability for higher frame rates. The connectors and outputs are the same.
 
Last edited:
That's awesome and helpful! Now can someone explain to me how the crop factor affects the lens?
 
Back
Top