Digital latitude tests

[Stephen Williams]

Hi,

 

This may be of interest to some peole here.

 

http://www.cinematography.net/hdcamtests/HDcurves.htm

 

Stephen

[Keith Mottram]

Hi,

 

This may be of interest to some peole here.

 

http://www.cinematography.net/hdcamtests/HDcurves.htm

 

Stephen

i'm not very good at reading graphs- does the D20 have the largest range or am i mistaken? i am somewhat confused, can you give me a breakdown Stephen?

 

cheers,

 

keith

[Stephen Williams]

i'm not very good at reading graphs- does the D20 have the largest range or am i mistaken? i am somewhat confused, can you give me a breakdown Stephen?

 

cheers,

 

keith

 

Hi Keith,

 

D20 would seem to be a clear winner.

 

JJ did post on CML that they had made vast improvements since David Stumps tests. Time will tell.

 

Stephen

[Francesco Bonomo]

D20 would seem to be a clear winner.

 

So it seems, I wish I could run a few test with it (and the Viper, too), but there's still no D20 in Italy yet. Of course, Latitutde is not the only thing we should look for in a HD camera, but this chart is very interesting indeed. I'm not surprised by the F900 curve, but it's curious to see a new camera, Red, so close to the F900 latitude (at least in the straight line region).

 

Isn't it a little bit ironic seeing the D20, a camera advertised as being targeted at the TV market, having the best latitude? :D

 

Thanks for posting this, Stephen

[Brian Drysdale]

Hi Keith,

 

D20 would seem to be a clear winner.

 

JJ did post on CML that they had made vast improvements since David Stumps tests. Time will tell.

 

Stephen

 

The Canon held up surprisingly well.

[Stephen Williams]

So it seems, I wish I could run a few test with it (and the Viper, too), but there's still no D20 in Italy yet. Of course, Latitutde is not the only thing we should look for in a HD camera, but this chart is very interesting indeed. I'm not surprised by the F900 curve, but it's curious to see a new camera, Red, so close to the F900 latitude (at least in the straight line region).

 

Isn't it a little bit ironic seeing the D20, a camera advertised as being targeted at the TV market, having the best latitude? :D

 

Thanks for posting this, Stephen

 

Hi Francesco,

 

I've been shooting with the Viper as it's the best I have available in Switzerland.

 

Arri are realistic that the D20 should be targeted at TV, it's still no where near the latitude or Dynmic range of film. Having a horse in both races keep's their feet firmly on the ground.

 

Just my 2c

 

Stephen

[Francesco Bonomo]

Arri are realistic that the D20 should be targeted at TV, it's still no where near the latitude or Dynmic range of film. Having a horse in both races keep's their feet firmly on the ground.

 

Stephen, I didn't mean to compare film's latitude to the D20 (I'd never dare to say something like that, yet!). What I meant is that a camera supposedly aimed at the tv market shows to have a better latitude curve than cameras that have been marketed as good replacement for film.

[George Lekovic]

Interestingly,

 

RED seems to have flunked this test, as curves show mere 7-8 stops of latitude. Or am I misreading the chart?

 

 

George.

[Stephen Williams]

Interestingly,

 

RED seems to have flunked this test, as curves show mere 7-8 stops of latitude. Or am I misreading the chart?

George.

 

Hi George,

 

Funny thing was, I said I was dissapointed with the images I saw at IBC and got flamed. I wanted to test the camera myself but was refused!

 

BTW Jim, I don't think there is anything wrong with your sensor

 

Stephen

[Graeme Nattress]

Well, you've got 6 cameras, tested at 6 different times, with different procedures. That chart is about as useful as a chocolate teapot. Not least it's comparing finished, in the field cameras, when RED is still in heavy R&D and does not yet exist as a fully finished camera. Indeed, the tests that David did with us did allow us to change some sensor variables and for us to achieve better results. Without a proper test methodology, such tests, when combined together like that, reflect more on the procedure itself, rather than the devices under test.

 

Graeme

[Stephen Williams]

Well, you've got 6 cameras, tested at 6 different times, with different procedures. That chart is about as useful as a chocolate teapot. Not least it's comparing finished, in the field cameras, when RED is still in heavy R&D and does not yet exist as a fully finished camera. Indeed, the tests that David did with us did allow us to change some sensor variables and for us to achieve better results. Without a proper test methodology, such tests, when combined together like that, reflect more on the procedure itself, rather than the devices under test.

 

Graeme

 

Graeme,

 

"without a proper test methodology." With all due respect stopping down a lens 1 stop at a time is hardly rocket science, and very little to go wrong!

 

Geoff makes it clear that the F900 would do better with one of the Digital Praxis gamma curves but if that is how Sony deliver the camera that is their fault. The Viper would have done far better had a magenta filter been added, again GVG's problem.

 

I understand the RED sensor is now showing over 11 stops of Dynamic Range, I look forward to seeing the new tests.

 

Stephen

[Graeme Nattress]

Stephen - Exactly. Geoff and David haven't looked at the RAW camera data, but "processed" data. I don't know what they did with my software, or what settings were used to generate the images they used. As you say, other cameras would have performed better under different conditions and I agree.

 

We have indeed improved the sensor's performance since David used it.

 

Graeme

[Phil Rhodes]

Hi,

 

If these tests were not carried out under matched conditions the results are extremely suspect.

 

Phil

[Graeme Nattress]

Yes Phil! That's totally right. And without any specified repeatable metric for acceptable shadow noise, where do you place black clip??

 

Graeme

[Stephen Williams]

Hi,

 

If these tests were not carried out under matched conditions the results are extremely suspect.

 

Phil

 

Phil,

 

Unless sombody cheats isn't 1/2 a stop 1/2 a stop? The centre point could be wrong, but the range would still show up.

 

Stephen

[Graeme Nattress]

It all depends on what the image processing did in camera to the highlights and what you define as when the noise swamps shadow detail. That can be very subjective unless you adopt a better testing methodology.

 

Graeme

[Phil Rhodes]

Hi,

 

Further to previous discussion about black clip, yes, depending on how much noise you are willing to put up with the JVC GY-HD101 has about a ten and a half stop range. Only it doesn't, obviously. Yes I have tried it.

 

Which is, incidentally, about the most accurate way of doing it. Putting your favourite step chart in front of a camera, waving a light meter about and saying "Right, that's obviously an F4" is fraught with difficulty concerning specularity and the inverse square law. But really, the reason for the suspicion is a little scientific principle which states that experiements must be repeatable. We don't know the circumstances under which these tests were done, and considering they were completely different, they're unlikely to be repeatable anyway.

 

Phil

[Gavin Greenwalt]

If all cameras shoot a Macbeth chart, finding a consistant noise level should be an easy prospect sinc eyou could just tell shake or combustion or fusion to analyze a neutral patch for grain characteristics and then find two exposures that give you the same "grain gain" setting.

Edited December 12, 2006 by Gavin Greenwalt

[Stephen Williams]

If all cameras shoot a Macbeth chart, finding a consistant noise level should be an easy prospect sinc eyou could just tell shake or combustion or fusion to analyze a neutral patch for grain characteristics and then find two exposures that give you the same "grain gain" setting.

 

Hi Gavin,

 

If your on CML Future cameras, you will know by know that is what Geoff Boyle did.

 

Stephen

 

Well, you've got 6 cameras, tested at 6 different times, with different procedures. That chart is about as useful as a chocolate teapot. Not least it's comparing finished, in the field cameras, when RED is still in heavy R&D and does not yet exist as a fully finished camera. Indeed, the tests that David did with us did allow us to change some sensor variables and for us to achieve better results. Without a proper test methodology, such tests, when combined together like that, reflect more on the procedure itself, rather than the devices under test.

 

Graeme

 

Hi Graeme,

 

Great news, Geoff Boyle is going to test all the cameras he can get, side by side in January, I hope Red is included, with the promised 11+ stops it should kick ass!

 

Best wishes,

 

Stephen

[Gavin Greenwalt]

Hi Gavin,

 

If your on CML Future cameras, you will know by know that is what Geoff Boyle did.

 

Stephen

 

aha see I knew it sounded like a good idea. ;)

 

Has he posted more specific testing procedures? Last I had seen it was just a simple 'Here is a graph.' description attached. I'll drop by future and take a look.

 

- Gavin

[Emanuel A Guedes]

Here is another (real) one (coming from Graeme himself):

 

http://www.dvxuser.com/V6/showpost.php?p=7...amp;postcount=1

 

[EDIT -- LINK now fixed]

Edited December 14, 2006 by Mr. Emanuel A. Guedes

[Emanuel A Guedes]

Methodology summed up (t)here:

 

http://www.dvxuser.com/V6/showpost.php?p=7...mp;postcount=18

 

http://www.dvxuser.com/V6/showpost.php?p=7...amp;postcount=7

 

http://www.dvxuser.com/V6/showpost.php?p=7...mp;postcount=12

 

http://www.dvxuser.com/V6/showpost.php?p=7...mp;postcount=16

[Emanuel A Guedes]

And as last (but not least) update:

 

http://www.dvxuser.com/V6/showpost.php?p=7...mp;postcount=25

[Keith Mottram]

It all depends on what the image processing did in camera to the highlights and what you define as when the noise swamps shadow detail. That can be very subjective unless you adopt a better testing methodology.

 

Graeme

 

Graeme,

 

Your test shows a larger range from your camera, using the same test do you think that a 950 or viper would yield a larger range and therefor consequently would you expect the d20 to have further stops of range than the red? basically do you think red is the only camera that was short changed on the cml test? and on a side note thats relevelent to me currently do you know when you'll be updating your final touch plugs for universal?

 

cheers,

 

keith

[Phil Rhodes]

Hi,

 

Speculatively, one would expect that the Red CCD would do pretty well due to its size.

 

However, since they're (nonsensically) refusing to enter into discussion about who made it and how, it's difficult to evaluate the thing on absolute terms.

 

Phil