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 :)

Opinion POLL: EPIC native ISO = 800 or ISO = 320?

Opinion POLL: EPIC native ISO = 800 or ISO = 320?

  • Total voters
    159
Graeme Nattress said:
Mid grey in the scene for any given ISO is always mapped to the same point in the image post development curve.

Graeme
Click to expand...


Are you saying that if I light an 18% Grey Card to 800 ISO (with a light meter) and record it with the EPIC at 800 ISO it will show up at around 45% on a waveform (assumption: middle-grey) AND if I then change the EPIC to 320 ISO and record the same Grey Card it will show up at the same 45% point in a waveform scale?
 
Graeme Nattress said:
ISO800 gives you 4.5 stops above mid grey, ISO320 therefore gives 3.2 stops above mid grey.
Click to expand...

With 4.5 stops above mid grey at 800 how many stops below mid grey at 800 does that leave? Generally speaking, when people discuss what the native ISO of a camera/film stock is it usually is where there is approximately the same DR above and below mid grey. Does that mean you're saying Epic only has 7-9 stops of DR? That doesn't make sense.
 
Graeme Nattress said:
ISO for RAW (and only RAW) cameras is a recommendation.

RAW view has a gain and curve applied to show you what's going on in the highlights. ALL development ISOs have neutral gain because we don't change the analogue gain on ISO. It's all in the digital development curve.


Graeme
Click to expand...

I would think this debunks the Raw view looks like ISO 320 theory. Raw in fact has a gain and a curve
 
shashbugu said:
Dude! don't confuse things. GN has clearly stated the ISO 800 should be the way to go in most situations Alexa's type of 800 ASA is a different story, different sensor, different color science. EPIC is 800 and that is that. Spending time answering your question ridiculous 1600 ISO question is a waste of time.
Click to expand...


I disagree. The EPIC and the ALEXA use THE SAME UNDERLYING CMOS SENSOR TECHNOLOGY. Every camera is it's own flavor but fundamentally they really aren't that differnt (some just are a bit more refined than others).

And I do think GN's response just confuses the issue more. Most camera manufacturers define "native sensor sensitivity" as the point in which the sensor can handle (IE: sample without clipping) an EVEN distribution of tonal range (IE: exposure f-stops) above and below a mid-point (IE: 18% Grey Card). If we take Graeme at face value this suggests the EPIC handles an even range of tonal values when it is set closer to 1600 ISO.
 
"Not every camera can record RAW but every camera "sees" light in the same black & white, linear way (read RAW). " - you need to take that up with ISO, which only specifies equations for shooting JPEGs on DSLRs, and for RAW shooting specifies that you "go with the manufacturer suggestion".

" what's the difference between analogue and digital "Gain". Aren't we adding noise either way?" - digital gain doesn't add noise, but it can make noise in the image more visible. Analogue gain adds noise and also removes highlight detail through clipping. As all reasonable (ie ones that make a viewable image) development curves add some gain to some code values, to suggest that the one with the least gain is somehow native is absurd as it would produce an ugly image and that helps nobody.

Yes ISO1600 would balance above and below stops more evenly but we feel that ISO800 is a more appropriate starting point for people shooting the camera. I guess that takes us back to the ISO recommendation....

Ian, Adjusting ISO doesn't effect noise or noise in any way, but it does help you visualize exposure. Exposing to avoid clipped highlights is like a form of ETTR (expose to the right) and it's a valid method. If you ETTR and the image appears too dark in the mid-tones for comfortable viewing, up the ISO. If you think the ISO is too high for your noise tolerance, open up your exposure more.

With respect to 18% grey and code values.... Say you shoot a scene at ISO800 and you have your aperture set so that a grey card appears to be about 45% in the developed image and now you change ISO to 400, and open up your aperture a stop to compensate that grey card will still be at 45% in the developed image.

Adam, At ISO800 you have more stops below mid grey than above - everyone has a different noise tolerance, we suggest ISO800 as a starting point. We don't use term native ISO as it doesn't really make sense in a fixed analogue gain camera, and even in DLSRs where you have combos of both digital and analogue gain it gets hyper-confusing.....

Graeme
 
I didn't think about that. You are suggestion the "noise" I have seen is in fact compression artifacts?

Karl H said:
In the thread you're quoting that was specifically a solution for someone experiencing blockies in the shadows oh high fps. If you do get these problems then one solutions is to decrease compression (not always possible),
Click to expand...
 
Yes a true raw image would be unusable and unviewable. I have such a feature in my R&D app should I need to look at raw code values in their native state, but it's not practical for any other purpose. You can think of the raw view as a pure curve (no tonal adjustments), neutral colorimetry and negative gain to allow you to see the highlights without any compression.

Graeme
 
Adrian, yes, and if you like the texture there it's not a problem at all. Remember ISO800 is a suggested starting point and the others are reasonable to use too!

BTW, replying to yourself, saying you're confused with your own response causes me concern... :-)

Graeme
 
Adrian Jebef said:
Technically speaking, a true RAW image has no gain or curve BUT it is also NOT a usable image. If "Raw Check" actually showed you a true RAW image it would look like a messed up, black and white, mosaiced mess.

Check http://thedigitalparade.blogspot.com/2010/03/video-vs-raw-log-vs-lin.html for a brief explanation.
Click to expand...

I realize that...just making clear that setting ISO to 320 and then checking Raw as some have suggested is not a real indication of the raw exposure to the image. A true raw image may not be usable until developed....but Raw check is not the even the RAW image debayered.....it is just an exposure tool.
 
I never use the RAW check, the histogram and other exposure tools are much more useful.

Adrian, you need to get a grip dude. Your subtle allusions to "more refined" cameras are not helping the discussion, more importantly they're inaccurate.
 
Graeme Nattress said:
Ian, Adjusting ISO doesn't effect noise or noise in any way, but it does help you visualize exposure. Exposing to avoid clipped highlights is like a form of ETTR (expose to the right) and it's a valid method. If you ETTR and the image appears too dark in the mid-tones for comfortable viewing, up the ISO. If you think the ISO is too high for your noise tolerance, open up your exposure more.


Graeme
Click to expand...

So is that what the ISO 800 recommendation is? just another way of ETTR? If thats the case, Thanks for clearing it up! I was always curious why the ISO800 MX was showing approx 320 in raw view. Who would have guessed that my trip on here to day would make me want to go play again.
 
Graeme Nattress said:
Adrian, yes, and if you like the texture there it's not a problem at all. Remember ISO800 is a suggested starting point and the others are reasonable to use too!

BTW, replying to yourself, saying you're confused with your own response causes me concern... :-)

Graeme
Click to expand...


I constantly confuse myself... that's the BEST part !
 
Will Keir said:
I didn't think about that. You are suggestion the "noise" I have seen is in fact compression artifacts?
Click to expand...


yup that's exactly what I'm saying. However let my clarify and say it 'could' be noise. In dominiks case it look distinctly like compression artifacts to me, but I can't see your shot so who knows..... Check my 5K image from the original thread I linked to : http://www.darklineonline.co.uk/uploads/5k120test.jpg .This is not noise, the entire image is full of artifacts due to such a high contrast shot and also high compression. This isn't even 2K, but 5K.

Compression works best on midtones and highlights. To put it bluntly stuff in the shadows get the dregs of the compression bandwidth. I see it all the time on a lot of people's slo-mo footage. It's the shadows that dance. So if you're getting this then there's several options. Evin suggested a low con filter which sounds like it would help to lift those blacks. More light would also help (tho clipping would suffer). Buy a 512GB card and lower the compression ratio (not for me just yet!).

You learn to follow a few rules - for example a shot with lots of detail and high contrast @ high fps seems to be a nightmare for redcode. One with a shallow DOF and less details will probably be fine. Rating at 320 helps.
 
Are the RED exposure tools needlessly complex or wonderfully powerful?

Are the RED exposure tools needlessly complex or wonderfully powerful?

The "native" ISO debate seems to be part technical, part philosophical and almost religious. (When someone claimed to be in the 800 club I LMFAO)

Should RED have structured their nomenclature and technical notes to suit the existing framework of a particular group of shooters? Video guys think in IRE, film guys want shutter angles not fractions of a second, histograms are for stills guys, etc. IMO this thread has morphed into a microcosm of the Alexa vs Epic culture wars. In Adrian's world (I've done plenty of DIT days on national TVCs in the LA market myself) the Alexa is very popular and lauded for its fealty to film style conventions right down to the color matrix.

ARRI was determined to make a digital camera for the people who hated digital so they made choices in the Alexa's design intended to make the experience as film like as possible. They chose technically inaccurate, but familiar, terminology and feedback indicators. The result has been generally good exposure choices by most DPs (and their trusty DITs) that have led to consistently successful outcomes. It's the "if it ain't broke don't fix it" approach and in the short term, as many have noted, it has served the Alexa well.

That said, I vastly prefer the RED approach because its based on the actual dynamics and operational characteristics of a digital system. The pain is having to make some effort to learn something new, maybe expand your understanding of imaging metrics, do a couple tests, etc. The gain is that you can become more capable of getting the most out of digital imaging systems in a wide range of scenarios if you understand the interrelationships well enough to better predict the impact of your choices.

When Evin chooses his exposure he has specific targets in terms of code values that will allow him to create the look he wants in post. When I am considering my exposure choice, what I want from the camera system is as much usable feedback as I can get. In terms of ISO setting for the monitoring path I find using 800 consistently offers a valuable reference point and allows me to expose to the right without fear of clipping. By toggling to the RAW view I am quickly able to get a sense of any areas of the sensor that I might be "starving", then either help them or write them off depending on creative intent.

At the risk of feeding those who contend that defining "proper" exposure on the RED cams should be more cut and dried I do have a note that serves me well, YMMV. One of my favorite methods for getting the most usable code values into the RAW data set is to adjust the iris while watching the 800ISO RGB histogram. What I'm looking for is how the values stack up or spread out, particularly when I take it too far to either extreme - how does it behave? Early on I learned the hard way that there were times that exposing to the right as dogma made noiseless but flat images that lacked guts no matter how much I jacked them around in post. Bottom line, each scene and lighting condition has a natural distribution of values and in most cases maintaining an accurate representation of those relationships will yield the most useful starting point for grading.

Cheers - #19
 
Graeme Nattress said:
ISO 800 gives you 4.5 stops above mid grey.

Graeme
Click to expand...

Thanks for clarifying this once again Graeme.

I know you probably should not answer this right now, but on the new DRAGON sensor coming
out late this year/early next year, how many stops above mid grey will ISO 800 give you?

My dream is that it is 10 stops, but 8 would be nice.

:)
 
Evin Grant said:
I never use the RAW check, the histogram and other exposure tools are much more useful.

Adrian, you need to get a grip dude. Your subtle allusions to "more refined" cameras are not helping the discussion, more importantly they're inaccurate.
Click to expand...


Sorry, but you are assuming the worst. I am not "alluding" to anything. I am simply stating that all current digital motion cameras use the same underlying CCD and CMOS sensor technology. Some cameras ARE more refined than others: like the EPIC, ALEXA, and F65 have a much more refined image than the 5D or my old Sony Handycam.
 
You must log in or register to reply here.
Back
Top