Frustrated with T2i colours / tonal range

[Ron German]

I have to admit I was expecting more from T2i colours, dynamic range.
I was aware about the (strong) aliasing / moire limitations and not so good detail resolution.
But I`ve used the camera about two months and, despite have tried many recipes (picture styles), all I got are dark tones, crushed blacks (didn`t like the super / extra flats), very restricted tonal colour range.
And when I watch some Panasonic GH13 images, I can not understand (to my eyes) how it can show so much more information regarding tonal range / dynamic range than T2i, being both restricted to 4:2:0.
Best
Ron

[Jon Fairhurst]

What software are you using for decoding/transcoding/editing. That can make a difference.

For instance, Quicktime is used by many NLEs as the default decoder. And Quicktime is known to mess up the gamma of the Canon MOV files. Before QT 7.6, Quicktime would literally clip the blacks and whites.

Personally, I use Cineform NeoScene, and it handles the colors well. It's possible that a different workflow would improve your results.

[Matthias Krause]

In my experience using Neoscene makes a world of difference when it comes to the tonal and the dynamic range of my 550D files...

[Ian Holb]

How are you playing back these files? Through HDMI to your TV monitor or on computer? Some low end video cards do not properly play back the HD in full quality. Is your TV calibrated?

[Joel Peregrine]

Hi Ron,

What lenses are you using with the T2i? I use a mix of 40 year old lenses with adapters and new Canon L series and notice a very difference in the contrast, saturation and definition. The newer lenses are not more accurate necessarily (I actually notice they tend to enhance the image rather than reproducing it faithfully) but they do have a greater dynamic range, reduced halos and much better contrast and saturation. The attached picture was taken in as close to identical conditions as possible with a Vivitar 24mm f2.0 and a Canon 24-105 f4L. Both cameras were set at f4, 1/50 and 800 ASA. The left is the Vivitar. UV filters also have an effect, especially cheaper ones. They rob the image of contrast and saturation and, at least in my experience, create a cooler image tint.

Shameless plug - the lens in the image is currently for sale: http://is.gd/e2xMR

[Jon Fairhurst]

Ian makes a good point. Video over HDMI can set black at 0 or at 16 (in the eight bit world.) The camera delivers full range. If your monitor expects 16-235, your whites and blacks will clip.

The bottom line is that the tonal range can be harmed if the signal isn't handled properly from end to end. Once, we establish a good workflow, then we can make an honest critique of the dynamic range.

Regarding colors, that generally comes down to the white balance. You can offset the WB in the menus, and you can also offset the colors with a colored card or filter.

The Zacuto Shootout gives a solid review of the dynamic range of these cameras. One note: they shoot at ISOs of 160, 320, 640, etc. These ISOs give the lowest noise in the blacks, but they don't have the highest dynamic range. Use 100, 200, 400, etc for the highest dynamic range.

A simple approach is to use 100, 200, 320, 640, and 1250 as your ISOs. When it's bright, you would be using 100 and 200 for dynamic range, and when it's dark you would be using 320, 640, and 1250 for low noise in the blacks.

[Perrone Ford]

Interesting:

Here is a photo I took 2 weeks ago with my T2i.

Perrone Ford's Photos | Facebook

This was recorded with my normal "flat" picture style. Lens was a 1972 Nikon 300mm F4.5 wide open.

Here is the same shot after color correction. Note, I did not TOUCH saturation. Just brought the blacks and whites into line:

Perrone Ford's Photos | Facebook

I did change the gamma somewhat which pushed midtones down a bit darker. But other than that, no manipulation was done to the image.

Quite frankly, I am thrilled with the dynamic range of the camera when shot well. And I am well pleased with the colors and tonality given the limitations of the codec.

[Ron German]

Thank you all
I`m using Vegas 9.0.e.
I transcode to both Neoscene and Sony MXF. The resulting clips are similar to my eyes - and I tried many times to notice some diference comparing them).
I`ve used the kit Canon 18-55mm and a new Tamron 28-75mm 2.8.
I`m playing back through computer monitor and 37 Sansung LCD TV display (DVD copies).
Jon, I have a T2i, so I don`t have 125, 160, 1250, etc ISO option.
Regarding GH13 and T2i dynamic range comparation, I see a clear difference in average examples on Vimeu clips.
Best
Ron

[Perrone Ford]

Originally Posted by Ron German

Regarding GH13 and T2i dynamic range comparation, I see a clear difference in average examples on Vimeu clips.
Best
Ron

Comparing video quality based on Vimeo clips is like comparing cars by playing with Matchbox cars. The Canon cameras have found their way onto broadcast television in the US and the UK, and feature films (like Ironman2 in small roles). No one is going to confuse these cameras for a RED, Alexa, or Genesis, but they do an amazing job for the price.

[Ron German]

OK Perrone, but I´m not interested in what position Canon HDSLR cameras have in the marketplace.
I`m just talking about what I`m most frustrated with, regarding tonal colour / dynamic range of my T2i.
And with my 31 years experience dealing with 16mm / 35mm motion picture film and broadcast tv, I think I can evaluate the diference of (average) GH13 X T2i clips, even displayed on Vimeu.
Best
Ron

[Perrone Ford]

Alright,

What is the dynamic range of the T2i (or other Canon) and what is the dynamic range of the GH1? I'm somewhat confused because you say you have 31 years of experience in motion pictures and broadcast TV and yet issue this statement:

"And when I watch some Panasonic GH13 images, I can not understand (to my eyes) how it can show so much more information regarding tonal range / dynamic range than T2i, being both restricted to 4:2:0."

The answer is fundamentally the codec. The GH1 recording at 17Mbps was rather unremarkable. The GH1 recording at 75-100Mbps is excellent. However, its the same glass, same sensor, same path through the camera. The only thing that's changed is the codec bitrate. The dynamic range hasn't changed one bit from the base camera because the sensor is exactly the same.

Perhaps you'd be happier with that camera. If so, best of luck to you. It's a fine system.

[David W. Jones]

As well as a number of other cameras, I also own both the GH1 and a T2i.
I'm just not seeing what you are seeing Ron.
I cut the T2i, GH1, and Red together all the time without the issues you describe.

[Ron German]

Thank you Perrone
I understand what you are saying (about GH13 codec diference), but I guess it`s not only the sensor size what defines the dynamic range. Isn`t the camera`s internal signal processing (besides codec) important?

David, I´m glad you don`t have problems cutting images from the 3 cameras.
By the way, what picture style are you using with your T2i?

Regards
Ron

[Perrone Ford]

Originally Posted by Ron German

Thank you Perrone
I understand what you are saying (about GH13 codec diference), but I guess it`s not only the sensor size what defines the dynamic range. Isn`t the camera`s internal signal processing (besides codec) important?

Dynamic range isn't tied to sensor size. The dynamic range is simply the ability of a chip to record bright white to deep black. A sensor could have ten steps between brilliant white and jet black, but if it could do that without blowing out or crushing, it would have very high dynamic range.

What you would LIKE to have is a camera with both high dynamic range, and a codec that has enough granularity to make a lovely, smooth transition between them. This is why having 14bit linear recording or 12bit log recording is so favored. LOTS of fine steps between black and white. But even having that ability to record finely, doesn't say anything about the dynamic range of the camera.

And we haven't even mentioned color subsampling.

[Jon Fairhurst]

I think a larger pixel can gather more photons, so in theory a larger sensor (with more area per pixel) can deliver more dynamic range.

That said, Perrone is exactly right. In practice, the dynamic range of a camera depends on the whole system, including sensor technology, size, and the rest of the signal chain. All it takes is one bottleneck to reduce the dynamic range - including a bottleneck in post.

One trick to regain some DR is to apply noise reduction in a 16-bit or higher post process. Though you might start with 8-bit video, the values created by noise reduction can now be in between the 8-bit bins. NR can change a steppy or blocky area into a smooth area. Go directly from NR to grading in the 16-bit or higher domain and you can scale those smooth areas as needed to fit your color scheme. After grading, you can output to 8 or 10 bits.

Many people complain about 8-bit video, but often the problem is 8-bit post processing, not the source video.