GH3 The 'usable' or real dynamic range of the DSLR's / GH3 in video mode... Anyone??

Stab

Well-known member
I'm just really curious. How many stops of dynamic range does the GH3 have? And what about other often used camera's, like the 5D M3, the FS 700, C100, 7D?

DXO just published that the GH3 has 12.4 stops of dynamic range. I assume this applies to the stills-mode only? What about video mode?
Is there a loss of dynamic range? And how does it work?

Anyone has hard figures? :)
 
Can´t say, I guess all newly released HDSLRs offer around the same DR. At least I don´t get exited when some benchmarks put the GH3 over the 5DIII, just because it scored some percentage more on a test chart.
More important seems to be how each camera interprets the data from the sensor in video mode and squeezes a scene's DR into its codec. The GH3 still needs a flat profile like the Technicolor one. As always early adopters pay the price, I guess end of the year we´ll have stable hacks, fingers crossed.
For everything else, there´s the BMC ;)
 
I never shot with a GH2, but in the GH3 you can dial all settings down to -5. Was the Cinema profile that flat?
 
I never shot with a GH2, but in the GH3 you can dial all settings down to -5. Was the Cinema profile that flat?

No.

Its possible to get the GH3 to shoot too flat, I rarely experienced that with the GH2, -5 across the board is too much for most subjects. I think that finding the balance between "baked-in" and "gradeable" requires a lot more testing with the GH3, particularly in the NR and Sharpening categories.
 
I have a feeling this has been asked many times before. I'd doubt very much if the dr of a gh3 is within half a stop of a BMCC though
 
Was the Cinema profile that flat?

That depends on how you use the term flat. I mean a flat exposure, no curve to it. I don't mean "lacking contrast" (which seems to be a fairly popular definition). All the other modes boosted chroma, or midtones, etc. Cinema at 0 is the most flat on the GH2, and properly exposes from 0 to 100 IRE (assuming the scene has such a range).
 
There's a difference between latitude and useable latitude, in a test, yeah sure you might see a little more over the preious models, but in a real world shooting enviornment, I don't see any improvement over the gh2. To my eyes, it looks as if the gh2 holds on to the colors better, but that might be because of a higher contrast, IDK. But this was a cool little test just to see the slight improvements over the prior. Of course the ergonomics are much better along with the menus build quality, ect.... But image quality, barely any.
 
I'm planning to do a DR test with my GH3 and my hacked GH2. I'm looking for feedback on my test setup. Here it is:

1. Get a large sheet of ND 0.3 (1-stop).
2. Cut it up into strips of varying lengths.
3. Stack the strips in such a way as there are steps from 1-stop, 2-stop, 3-stop, ..., up to 12-stop.
4. Put an even light behind the stacked strips.
5. Expose the brightest patch on the strip at 100% luminosity.
6. Use a waveform monitor to count how many stops of DR the footage captured.

Any thoughts on this approach?
 
That approach will not evaluate tonality or colour shifts (you might actually see colour shifts)

There is DR and 'usable DR'

ie if your highlights are wonky or shadows noisy then thr DR is not usable

Why not just point a light at a wall and do some brackets (of aperture)

S
 
I think of latitude as flexibility to fix

S

Answer to the original question...

I pretty much assume that the current DSLRs, for video have about 8 stops, that's from 'black' to 'white', which means theres about 6 stops of 'grey' to use for tonailities in the scene. (Color... we don't need no stinking colors...).

On the topic of 'latitude'...


Latitude has to be determined in the context of the output image. In still photography a paper print had about 7-8 stops of from 'black to white' density. (Most print papers were more 'contrasty' than the corresponding negative...).
To be sure there were grades of paper and some adjustment could be made... as well as adjustment in the developing of the negative...

Which leads to 'latitude'... for a film negative, latitude meant how many stops over or under an exposure could be, and with development changes or paper changes, a 'adequate print' could be made.

So, in a scene that had 12 stops of range from lowest light to highest light a choice had to be made... record the image on the negative, then develop or print to conform the densities recorded in the negative, to the output print.

From reading about movie film, there is very little conceptual difference in making a 'print' from a negative. Movie projection print film had about a 8 or so stop range from 'clear', that is film base + fog to 'max effective density'(*).

*While print material may have shown a denser value for a given 'dark' area using a densitometer, after a certain density, the darkness of the theater ( or in fact 'lightness'...) would impact the contrast ratio, as well as the human ability to determine such a change in density...

In the case of modern still cameras and still images, the 'raw' format usually allows for higher dynamic ranges, because the data has not been compressed, and one can effectively speak of 10 bit or 12 bit intensity values.

Higher end moving picture digital cameras may now allow for larger values of intensity digitiziation, but as it is, from what I can tell, most DSLR's that ordinary mortals by are still in the 8 bit digitization range, and as such, can only accommodate a 8 'stop' range... more or less. There are tricks that can allow for accommodating a larger dynamic range, but those tricks have a cost, and they are usually seen when the image is manipulated to 'stretch' or 'match' or 'grade'...

I think the advice of 'shoot DSLR's as if one were shooting on slide film', for 'video' still holds. Slide film was always very unforgiving of 'over exposure', and some 'underexposures' could be saved, but within limits.
 
Back
Top