Noise test shoot. No lighting and tweak in post

toke lahti said:
Saru,
did you have same amount of light when you shot with dvx as with hvx?
If you did, dvx seems to be at least one stop faster.

Yeah by off reading gray card at the same time under the same lighting conditions, there seems to be 2 stops difference, DVX being faster.

And for the noise, I looked at both cams footage on 23" Cinema display and because DVX gets more magnified to fill entire screen, DVX seems much noisier. But at the native pixel size, HVX's noise are more apparent. Overall, HVX noise are big.
 
Alex Leith said:
Thank you Bgundu for sharing your findings.


No Problem. Just to be clear, I'm not complaining about this camera. Anyone who had the DVX will definitely see the merits of the HVX. The discussions about noise is always the same in the DSLR world too. Everyone compares all the specs between cameras for bragging rights. After the dust settles people begin to finally show some beautiful images.
 
Barry_Green said:
Be careful with that advice, as that depends entirely on how the brightness/contrast in your particular LCD is set up.

If you want to know whether you're really blowing out your highlights, set your zebras on 105% and USE THEM. Check for blowouts. If you're not seeing zebras, you're not blowing out, regardless of whether it "looks" blown-out on the LCD or not. If you are seeing zebras, you're in danger of blowing out.


That seems like a good idea to set the zebras at 105%.

My reply to adjust the brightness contrast of the lcd was more to insure the lcd matches the recorded footage.
I am not really having the issues of low light noise (probably more to do with shooting mainly outside), but having more of an issue with weak color, that looks like SD desaturated.

To me,my lcd looks normal but the exposure is actually a bit too low.

Question: is it just mine, or does anyone else have an lcd that does not display the recorded footage the same brightness contrast?

(This is not about blowouts, it is about lcd matching recorded footage when set to center or default settings)

I can only get the lcd to match the footage by turning down the lcd brightness.
 
Last edited:
you should NEVER be judging exposure or color on the lcd - it is NOT accurate - you need a production monitor or at least something that will display in the ballpark
 
If you have lcd with internal LUT and it is calibrated and profiled with accuracy, it shows color as accurately as any crt.
Crt's tend to hide some problems, so they are not the best choise for critical viewing.
And most of the planet will be watching hd from lcd.
 
Saru,
could you do one more test.
This time with constant gamma (like cine-d) and go through all matrixes?
Or with constant f-stop with all gammas and matrixes?
 
bgundu said:
- Matrix set to Normal will yield less chroma noise. Cine-matrix has the same noise but contains more color. I prefer having more of the lumance type of noise.
Saru's first tests show that even with normal matrix only cine-d gamma gives quite block free shadows.
So I have to thank Panasonic for giving me lots of gammas and matrixes, but so far it seems that only what I'm ever going to use is cine-d/normal and normal/normal...
 
I am slated to shoot a short film in April with the HVX. The story takes place at night so there will be lots of darkness. I also want to shoot at f2.8 on roughly a 20mm to milk as much DoF out of the camera as I can. With all this discussion of noise/blocks and the dark story and the (almost) wide open f-stop I'll be shooting at, can anyone offer any tips?

Seems like my highlights should only be roughly 2.5 or 3 stops higher than a key that is exposed at a f4 then increase the contrast in post. I have yet to do tests with a person and chart.
 
Last edited:
Saru, is the DVX in progressive or interlaced mode when it is 2 stops faster than the HVX? Is the HVX slower in progressive than in interlaced, like the DVX?
 
My advice is do not worry about the noise - period - worry about good shots with great exposure and it will look great - too much worrying about the noise will take your creativeness into too much technical.
 
Can anyone tell yet how many stops you have with out noise. For instance if I set a spot meter at ASA320 and the highlight is f11 and the deepest shadow is f2.2 are there detail in both and is the picture clean?
 
Back
Top