Is the future FF35?

[Tom Lowe]

Hi Daniel,

My only reason to shoot FF35 would be for better quality.

Oh, only a little thing like image quality? Who needs that! :wink:

You also get better sensitivity, I think. The 7D got smoked by the 5D2 in recent ISO 3200 and 6400 tests, even though the 7D supposedly has some new, post-5D2 noise-reduction technology built in to its sensors.

 

[Daniel Browning]

You also get better sensitivity, I think.

Stephen is talking about using them at the same DOF and same light. In order to do that, you have to use up the entire sensitivity advantage just to get the same DOF. There is none left over for superior noise, dynamic range, etc. But at least they can reach parity.

The 7D got smoked by the 5D2 in recent ISO 3200 and 6400 tests

Those tests used thinner DOF or more light for the 5D2, which is fine for some situations (daylight exteriors, zen master focus pullers, etc.), but not all.

...even though the 7D supposedly has some new, post-5D2 noise-reduction technology built in to its sensors.

That's just JPEG engine changes. There is, however, a bias offset imbalance between the 8-channel readout that causes unfortunate demosaic artifacts in the shadows at low ISO. It affects 9 out of 10 7D cameras. Some 5D2 cameras had a similar problem, except it was a gain imbalance. I hope Canon will soon release a firmware update to fix the problem.

Aside from that problem, I can say that the 7D technology (performance "per area") did catch up to the 5D2. On a per-area basis, they have equal performance. Specifically, 1.7 e- read noise on 7D adds in quadrature to match the 2.6 e- on 5D2. The 50D was not up to that level.

The 7D and 5D2 are a case study in what I'm talking about: same performance when used at the same DOF.

 

[Tom Lowe]

Yes, I understood what Stephen was talking about. I mentioned the speed as a general statement. For many situations, like nature shooting, the gain in ISO/ASA ratings will be significant, because the shooters are not concerned about lighting sets or pulling focus on actors.

But my point was that, Yes, the gains in image quality are very real shooting on a FF35 sensor vs APS-C. There can be no doubt about that. And the benefits are strikingly clear, even downsampled to 1080p.

 

[Stephen Williams]

But my point was that, Yes, the gains in image quality are very real shooting on a FF35 sensor vs APS-C. There can be no doubt about that. And the benefits are strikingly clear, even downsampled to 1080p.

Hi Tom,

I could also point out there are very real advantages to shooting 35mm film for motion picture V Digital at this point in time. It's also an option that can be done today.
The quality gains you talk about are in still photography, it's not possible to empirically test what you are saying when applied to motion picture with current technology.
From what I have seen in the Cinema Benjamin Button, looked pretty good, shot on old 2/3" sensor technology, not sure if your theory will hold up in practice.

Stephen

 

[Tom Lowe]

Hi Tom,

I could also point out there are very real advantages to shooting 35mm film for motion picture V Digital at this point in time. It's also an option that can be done today.
The quality gains you talk about are in still photography, it's not possible to empirically test what you are saying when applied to motion picture with current technology.
From what I have seen in the Cinema Benjamin Button, looked pretty good, shot on old 2/3" sensor technology, not sure if your theory will hold up in practice.

Stephen

Stephen, there is no firewall left between stills and motion. These things are coming together. In terms of empirical tests, I can only go off what my own eyes tell me. You can try it yourself. Go rent a 5D2 and a 7D, then shoot a side-by-side 240-frame RAW timelapse sequence on each in low light, maybe a sunset. Downsample and render your tests at 1080p and see if you can spot the differences in image quality.

I agree that Button was a great film, and I have said many times that 2/3" can most certainly be used for feature film work. I'm just saying, if you can choose higher quality over less quality, I will go with the higher quality almost every time, unless there is some specific need for greater DOF, etc.

 

[Stephen Williams]

Stephen, there is no firewall left between stills and motion. These things are coming together. In terms of empirical tests, I can only go off what my own eyes tell me. You can try it yourself. Go rent a 5D2 and a 7D, then shoot a side-by-side 240-frame RAW timelapse sequence on each in low light, maybe a sunset. Downsample and render your tests at 1080p and see if you can spot the differences in image quality.
.

Hi Tom,

Compare a time lapse shot with film S35 against the 5D, the 5D is nowhere close for dynamic range or color palette. The markets I work in people want the best.

Stephen

 

[Tom Lowe]

Hi Tom,

Compare a time lapse shot with film S35 against the 5D, the 5D is nowhere close for dynamic range or color palette. The markets I work in people want the best.

Stephen

Hehe. Anytime you want to compare your S35 film timelapse to a 5D2 RAW timelapse shot by me, let's do it. Just bring me on to one of your big mega-million-dollar television commercials, Stephen, and we will get this done! :sifone:

 

[Sanjin Jukic]

Hehe. Anytime you want to compare your S35 film timelapse to a 5D2 RAW timelapse shot by me, let's do it. Just bring me on to one of your big mega-million-dollar television commercials, Stephen, and we will get this done! :sifone:

Very good point, Tom!

 

[Daniel Browning]

Hehe. Anytime you want to compare your S35 film timelapse to a 5D2 RAW timelapse shot by me, let's do it. Just bring me on to one of your big mega-million-dollar television commercials, Stephen, and we will get this done! :sifone:

Don't bring a knife to a gun fight. Low ISO film will walk all over the 5D2 when you get over 8 stops of dynamic range because of the pattern noise in the 5D2. You have to stick to low dynamic range or high ISO stuff if you want the 5D2 to shine.

The D3X, on the other hand, has 13.5 stops of dynamic range with no pattern noise: it can match any print film.

 

[Tom Lowe]

yeah, but DR isn't everything. and there are things that can be done to RAW frames in post to help compensate for that. i'm not disputing film's advantage in DR, though.

 

[Sanjin Jukic]

The D3X, on the other hand, has 13.5 stops of dynamic range with no pattern noise: it can match any print film.

Amazing "13.5 stops of dynamic range with no pattern noise: it can match any print film."

RED should look to achieve this DR...

Shot on Nikon D3x.

LINK>>>

 

[Tom Lowe]

I have a hard time believing that the D3x blows the 1DsM3 and 5D2 out of the water when it comes to DR. It's the same basic technology.

 

[Daniel Browning]

FWIW, I don't think Nikon's web examples show the DR very well at all. (The reds are blown, probably due to sRGB out-of-gamut conversion.)

RED should look to achieve this DR...

Agreed. I'm sure it's a high priority.

 

[Daniel Browning]

I have a hard time believing that the D3x blows the 1DsM3 and 5D2 out of the water when it comes to DR. It's the same basic technology.

Hobbiest rockets and the Saturn V are the same basic technology, but only one of them goes to the moon.

The 1Ds3 has 4 readout channels and off-sensor ADC.
The D3X has 6000 column parallel ADC on the sensor itself.

They do have the same basic quantum efficiency, which means they gather close to the same amount of light. But where they greatly differ is in read noise (shadow noise).

Firstly, it does not have pattern noise. The 5D2 has random noise that limits the dynamic range to 11.5 stops. But you'll never get close to using that much, because there is another problem you run into first: pattern noise. Lines running through the deep shadows. That kind of noise will limit you to 8-10 stops, depending on the color balance of the light.

Of course, where the 5D2 shines is at high ISO. There the read noise is lower than the D3X.

Here is an example of the pattern noise I'm talking about:

(Click image for full size.)

It's actually one shot from a multi-exposure HDR sequence I did shortly after I got the 5D2. But I wouldn't have to shoot multi-exposures if the 5D2 had enough dynamic range to start with. (And anyway, multi-exposure is not an option if you have movement in the frame.)

It was developed for ~11 stops of dynamic range during raw conversion, but you don't have to go that far to see the pattern noise.

The 1Ds3, on the other hand, seems to have much better electronics. The pattern noise is reduced enough that you can use 11 stops of dynamic range. But the random noise still limits you. The D3X random noise is low enough that you can get over 13 stops.

 

[Tom Lowe]

This is the first I have heard about the D3x being such a marvelous camera. Maybe I will have to rent one for a gig and find out!

 

[Daniel Browning]

This is the first I have heard about the D3x being such a marvelous camera. Maybe I will have to rent one for a gig and find out!

Well, it had better be marvelous for the price they put on it. Even so:

• It can't touch a 5D2 in low light.
• There is no Nikon equivalent for my favorite Canon lenses, like the 24mm f/1.4 II, 35mm f/1.4, 70-200 f/4 L IS, and more. (Though they do have some of their own unique offerings like the 200-400 f/4 VR and the excellent 14-24 f/2.8, if you're into slow zooms :wink5.
• Nikon has some issues for starry nightscapes: if you disable in-camera dark frame subtraction ("long exposure noise reduction") so you can do 360-degree shutter, then Nikon will enable a different kind of noise reduction that smudges hot pixels into the background. Unfortunately, it tends to think that stars are hot pixels.
• No video.

 

[Andrew Walker]

Well, it had better be marvelous for the price they put on it. Even so:

• It can't touch a 5D2 in low light.
• There is no Nikon equivalent for my favorite Canon lenses, like the 24mm f/1.4 II, 35mm f/1.4, 70-200 f/4 L IS, and more. (Though they do have some of their own unique offerings like the 200-400 f/4 VR and the excellent 14-24 f/2.8, if you're into slow zooms :wink5.
• Nikon has some issues for starry nightscapes: if you disable in-camera dark frame subtraction ("long exposure noise reduction") so you can do 360-degree shutter, then Nikon will enable a different kind of noise reduction that smudges hot pixels into the background. Unfortunately, it tends to think that stars are hot pixels.
• No video.

This is a great reason why not to get the D3X...well at least for myself and I think Tom. I mainly use my 5D2 for night timelapse. Either cityscapes or star timelapse. I very rarely use my 5D2 for just taking a couple stills of something random. For me its primarily a timelapse camera. The shots you can get with the 5D2 at night are amazing.

 

[Daniel Browning]

This is a great reason why not to get the D3X...well at least for myself and I think Tom. I mainly use my 5D2 for night timelapse. Either cityscapes or star timelapse.

Me too. The 5D2 is the best thing going for stars and any low light stuff right now, and that's a big part of why I own it. Unlike Nikon, it also has electronic first curtain ("Silent mode 2") which is great when your camera support is delicate enough that the focal plane shutter itself can introduce sufficient camera shake to affect the image. For me that means sharper macro and astrophotos. (My timelapses are usually wide angle, and I don't even need MLU.)

Of course, the list of things I dislike about the camera is quite long. Canon always seems to strike that delicate balance between just enough features to entice me away from the competition, and just enough flaws to breed contempt.

 

[Stephen Williams]

yeah, but DR isn't everything. and there are things that can be done to RAW frames in post to help compensate for that. i'm not disputing film's advantage in DR, though.

So for a sunrise timelapse there is no point in doing the side by side test you suggested.

 

[Tom Lowe]

So for a sunrise timelapse there is no point in doing the side by side test you suggested.

Well let's find out. :emote_popcorn:

There are advantages to digital other than DR. And again, there are things that can be done in post at the RAW stage that simply cannot be done on film.

Dragging a chemical strip through a huge camera presents other problems. There used to be an old-school guy who posted here who described in detail all the problems and horrors that can arise while shooting timelapse on chemical film, including lab problems, exposure problems, flicker, scratches, etc. I'll be standing by with my 5D2 whenever you're ready. :ihih: