Welcome to our community

Be a part of something great, join today!

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 :)

I want to love the Raven but.....

Steven Abrams said:
Red lists the Helium and Dragon sensors in their tech specs as having 80db signal to noise ratio, which is around 13.2 stops.
http://www.red.com/products/weapon-8k#tech-specs

16+ is what they rate HDRx being. If the ursa 4.6K is 14-15 stops it makes sense the OP sees more useable dynamic range on the ursa if not using HDRx.
Click to expand...

Im not sure I follow.

Signal to noise ratio does not really dictates how many usable stops the camera has.

And if you use HDRx then even with a half miss you can get out far more than 16+ stops out of any red camera that has hdrx. Lets say for simple calculation that a red camera has 10 stops. Then you turn on HDRx and set it to... for example 2 stops. Then you got one exposure with 10stops then you got another exposure that covers highlights 4 times as good as the A exposure.... now if you combined those two expos in post you for sure can drive out far more than those 10 stops. Now the debate is possibly not if dragon has 10 stops is more about where it is bettween 14 and 16+ stops all depending on how much noise or low level of detail you consider a usable image. But people have mesured 16+ stops with dragon without Hdrx on. Which I think is not done with the ursas. But I might mix it up the mini with the orginial ursa but again I think if the mini would trump the dragon in that regard then there would be quite a few test charts to prove it.


And red always, as long as i can remember, listed their cameras with 80Db signal to noise ratio. Now its quite clear that epic scarlet has more noise and less DR than helium still both got 80Db noise ratio listed so would be great if you explain how the 80Db number in the listings correlate to anything. exept per pixel noise levels.
 
Björn Benckert said:
Signal to noise ratio does not really dictates how many usable stops the camera has.
Click to expand...

Well, I honestly can't see how it won't.

If we simplify the discussion as much as possible: a sensor at full well capacity will reach a certain charge level. If you then compare that to the sensor's noise floor, that will have to be defined, you get a de facto signal to noise ratio. According to the principles of each stop of light doubling the intensity of the previous stop, we get the well known exponential curve. But this also dictates you can only double the signal strength so many times within a certain range (like 80dB) before there are no 'stops' left. Each stop correlates to a 6dB increase and if you take 80 dB divided by 6 dB you get 13.3. And oh, for the record, Canon rates their cinema sensors at around 64-68 dB or something...

So let's say you can double the signal strength 13.3 times in linear light before running out of dBs.... Where do the other stops come from? No amount of spreading the 'linear light' out over a logarithmic cure will change the underlying DR potential. Making an image flat doesn't increase true DR per se.

Many cameras today stipulate they are set at ISO 800 or so.... that would give you (?) 3 additional stops compared to an assumed base ISO/gain of 100/0.

I'm not at all interested in a "brand A" vs "brand B" discussion, but I would like to learn more about how a manufacturer's claimed native sensor SNR compares to captured DR. Right now, it doesn't quite make sense to me.
 
Just some info from Alen Roberts:
Red Epic Dragon https://tech.ebu.ch/docs/tech/tech3335_s16.pdf
Arri Amira https://tech.ebu.ch/docs/tech/tech3335_s15.pdf
Arri Alexa https://tech.ebu.ch/docs/tech/tech3335_s11.pdf
Ursa Mini 4,6k http://www.newsshooter.com/2017/01/02/blackmagic-design-ursa-mini-4-6k-newsshooter-ebu-lab-test-by-alan-roberts-14-stops-of-dynamic-range-and-best-used-in-raw/
Ursa min pro 4,6k not measured yet, but better then Ursa Mini 4,6k according to slashcam.de https://www.slashcam.de/4K-Kamera-Vergleich.html (it has better de-bayer).

Do with it what you want, I'm happy with my BMD ursa mini pro and for higher res. I rent.
 
I couldn't see what settings were used in REDcine for the test you highlighted...

And to the original poster - you're really going to want to learn more about post.

Prior to REDs ipp2 pipeline I'd say REDs on set look (their various 709 presets) have been more contrasty, saturated, and abrupt in the highlight roll off than some people prefer. But in post you could use Redlog Film with a LUT and find gentle, low contrast, soft roll off to suit ones taste. We didn't typically just grade in Redlog film without some form of a LUT because it takes a lot of wrestling to really bring out the balance and saturation (which is what the LUT would do).

But now with ipp2 you can use specific settings in the release build of your camera and specific LUTs from RED that work in Resolve (on top of Redlog film) and you should be able to dial in your base settings quite quickly. REDs done a really good job here. Or use a beta build and you can handle this in camera without the LUTs

Hope that helps...
 
Can you post comparative .dng and .r3d frames?

I agree with Paul above; I'm seeing easily distinguishable improvements with IPP2 even on MX footage! It's not a miracle worker or anything, but definitely a good 5% better colour/highlight retention (and better roll-off too).

That said, the 4.6k is definitely a much larger sensor to Raven, especially in 16:9 frame sizes... Actually it has over 60% larger surface area than dragon at 4k 16:9 (~25.34x14.25mm compared to ~20.5x10.8mm); as in, there's a 63% more surface area an 1.32x crop factor between the two. The 4.6k is a smidgeon smaller than 5k MX and 5.5k dragon. It also does 16:9 in full 4k DCI (4096x2304) like the R1 did... And while it "only does" 60p at 4.6k, it can do so uncompressed, 3:1, or 4:1 (whereas Raven goes down into 8:1 at 4k HD 60p, which isn't too bad... But 13:1 at 120p is taxing, even for REDcode).

OH, and you may wanna refrain from saying RED has "a look"; it's raw and can do whatever you want (and, as crazy as it sounds, people have gotten banned for saying such things.)
 
Last edited:
Stefan Antonescu said:
The Alexa 65 doesn't have an OLPF ? Do you have a source ?
Click to expand...

Most of the time you state what you have, not what you don't have.
I can't find a source that the ARRI Alexa65 doesn't have a spare tire or a boat, Arri however doesn't state that it has.
The last time I was in Germany they didn't have it and I can't find it on their website so I think they still don't have it.
 
Steven Abrams said:
Red lists the Helium and Dragon sensors in their tech specs as having 80db signal to noise ratio, which is around 13.2 stops.
http://www.red.com/products/weapon-8k#tech-specs

16+ is what they rate HDRx being.
Click to expand...

Really just need to step in here and say this is incorrect. 16+ stops is the rating without HDRx...... And it's been measured by many.

I've found 16+ securely and publicly. Most who have a conflicting opinion state 15 stops.

You can't always use the "good old standard" SNR for straight DR measurements these days for a variety of reasons. i.e. How REDCODE RAW is encoded.

Continue on.
 
Misha Engel said:
Yes Arri Alexa 65 and BMD UMP don't have an OLPF (so the have the same, just air and it's excellent you can even breath it).
Click to expand...

Stefan Antonescu said:
The Alexa 65 doesn't have an OLPF ? Do you have a source ?
Click to expand...

Don't do drugs. The Alexa 65 does indeed have an OLPF......

My source is the last time I shot with it and if my word is somehow not good enough, I guess I can link to the interview with Franz Kraus where he discusses the OLPF design in the Alexa 65.....


To the original poster, Douglas.

As many have mentioned Raven's sensor size at 4.5K 2.1:1 is indeed smaller than the BMD Ursa Mini Pro 4.6K. Here's a visual comparison.

RED Raven 4.5K 2.1:1 format compared to BMD Ursa Mini Pro 4.6K


Raven is more of a direct-ish comparison to Academy Super 35mm in my mind.

If you are looking to have more of an FOV equality between BMD and RED, checkout the RED Scarlet-W 5K S35. Or any 6K Dragon camera at 5K. Here's a comparison again with BMD.

Dragon 5K format compared to BMD Ursa Mini Pro 4.6K
 
Phil Holland said:
As many have mentioned Raven's sensor size at 4.5K 2.1:1 is indeed smaller than the BMD Ursa Mini Pro 4.6K. Here's a visual comparison.

RED Raven 4.5K 2.1:1 format compared to BMD Ursa Mini Pro 4.6K
Click to expand...

That's a little misleading since 2.1:1 is a non-format, so anyway you cut it, you're gonna have to crop... 3840x2160 is the largest 16:9 res that Raven (~20.5x10.5mm) does and that's a ~1.32x diagonal crop compared to the BUMP4.6k's 4608x2592 (~25.5x14.25mm), not 1.1x, as your link implies. The 4.6k also makes for better anamorphic than Dragon at 5k Ana (in terms of sensor size and resolution).

Same thing goes for Dragon at 5k; when you crop to 16:9, it's actually smaller than the 4.6k... So unless you're solely going DCI and/or have no intention of using broadcast standards, it's a bigger sensor in practical usage.
 
Last edited:
Mike P. said:
That's a little misleading since 2.1:1 is a non-format, so anyway you cut it, you're gonna have to crop... 3840x2160 is the largest 16:9 res that Raven (~20.5x10.5mm) does and that's a ~1.32x diagonal crop compared to the BUMP4.6k's 4608x2592 (~25.5x14.25mm), not 1.1x, as your link implies. The 4.6k also makes for better anamorphic than Dragon at 5k Ana (in terms of sensor size and resolution).

Same thing goes for Dragon at 5k; when you crop to 16:9, it's actually smaller than the 4.6k... So unless you're solely going DCI and/or have no intention of using broadcast standards, it's a bigger sensor in practical usage.
Click to expand...

RED cameras are indeed multi-format, so yes the HD aspect ratio concept you present is true. As is using the 2.1:1 full format of Raven for say widescreen.

That's an interesting other conversation all together. I mostly film in DCI aspect ratio for nearly everything, minus commercial work which is usually 16x9.
 
Phil Holland said:
That's an interesting other conversation all together. I mostly film in DCI aspect ratio for nearly everything, minus commercial work which is usually 16x9.
Click to expand...

Phil,

What things are you taking into consideration when you are choosing your aspect ratio on your projects? Why is most of your commercial work in 16x9 (~1.78:1) and not in the full container (~1.9:1)? I would think that doing the 1.9:1 aspect ratio gives you more options, no?
 
Ryan Williams said:
Phil,

What things are you taking into consideration when you are choosing your aspect ratio on your projects? Why is most of your commercial work in 16x9 (~1.78:1) and not in the full container (~1.9:1)? I would think that doing the 1.9:1 aspect ratio gives you more options, no?
Click to expand...

Most of the my commercial work is destined for 16x9 screens. While there's more flexibility to crop or re-frame with the wider frame, that's extra work for a post house. Most of what I shoot is shot "as it will be" in terms of framing and I'm a stickler for composition, so not much ever truly needs to be done. That's not how everybody's path is though.

This is one of those things we discuss before any shoot to make sure the pipeline is optimal for the production.
 
Phil Holland said:
Don't do drugs. The Alexa 65 does indeed have an OLPF......

My source is the last time I shot with it and if my word is somehow not good enough, I guess I can link to the interview with Franz Kraus where he discusses the OLPF design in the Alexa 65.....
Click to expand...

Well I kind was on legal drugs (beer) and it tasted good.
But you are right about the OLPF and I must have mixed up the difference on a debate we had here with the Nikon D800/D800e/D810/D850.

And here the requested link http://www.arri.com/camera/alexa/technology/arri_imaging_technology/lowpass_filter_pack/

That what I like about this community, at the end, most of the time we get the right info.
 
Phil Holland said:
Really just need to step in here and say this is incorrect. 16+ stops is the rating without HDRx...... And it's been measured by many.

I've found 16+ securely and publicly. Most who have a conflicting opinion state 15 stops.

You can't always use the "good old standard" SNR for straight DR measurements these days for a variety of reasons. i.e. How REDCODE RAW is encoded.

Continue on.
Click to expand...

This is an endless debate, it all comes down how you measure and who measures and what you think is usable. There is no commen accepted standard for this.

https://www.dxomark.com/Cameras/Compare/Side-by-side/Red-Helium-8K-S35-versus-Red-Epic-Dragon___1110_898
https://tech.ebu.ch/docs/tech/tech3335_s16.pdf

Lets say 14+ like ARRI does (human eye is around 13 stops DR with a native ISO between 500..1000).

Lucky for us, lots of camera's can do 14 stops these days.
When there is to much light we us ND's and when there is not enough we use lights (or Sony A7SII and/or Canon ME20F-SH).
 
You must log in or register to reply here.
Back
Top