xa20 dynamic range

[Bruce Watson]

Anyone give me an idea of the dynamic range capabilities of the XA20? I'm not talking about it's low light capability.

Any of you wedding shooters willing to tell me how well it does pulling detail from a black tux and a white dress at the same time? In direct sunlight?

How much better than, say, an xa10? How much worse than, say, a GH3?

[Bruce Watson]

No one? Really?

[Don Palomaki]

Have not had the XA20 long enough to assess the question you ask with respect to a XA10.
In any case that will require some shooting in a controlled environment.

I do not have access to a GH3,

[Don Palomaki]

This will not answer your question, but it provides some information from a quick test I did this AM.

Target was a Macbeth ColorChecker chart illuminated by a 50W PAR20 halogen flood lamp from about six feet
The cameras were an XA10 and XA20, set in AWB mode, 60i, manual exposure, composed so the chart filled about 95 percent of the vertical frame. Same camera position.
Exposure was 1/60, f/2.8 with 0 dB gain for both.
Also shot in a unlit room with lens cap on for true black.

Looked at the overall IRE levels.

Loaded the AVCHD file into Edius Pro 7. and saved snaps from the video on the timeline to JPG files.

Loaded the JPG into Photoshop and looked at histograms of the black and white squares.

Results:
True Black was black - 0 IRE, 16 in 8-bit digital JPG file for both camcorders.

XA10
White square (optical density 0.5) was digital 148 about 58 IRE
Black square (optical density 1.50)was digital 35, about 8 IRE
A reflections from a metal tool near the chart peaked at about 103 IRE

XA20
White square was digital 157 about 64 IRE
Black square was digital 38, about 9 IRE
A reflections from a metal tool near the chart peaked at about 105 IRE

This indicates that the XA20 is somewhat more sensitive than the XA10.
There is detail capability in the image below the Macbeth black level
There is detail capability in the image above the Macbeth white level

This test did not address high contract outdoor situations; e.g. a black tux in a deep shadow vs. a white gown in full midday sun.

Dynamic range is limited the the characteristics of the sensor including the noise floor and saturation point.

Latitude of the image you produce is also limited by the bit depth of the file format, the noise of the electronics, the gamma of the image processing, and so on.

[Jeff Harper]

In an offhand way I can tell you this: the XA20 dynamic range is a tad or two better than the XA10. The GH2 is much, much better, of course. You will never see anything close to the dynamic range of the GH2 or GH3 with the XA20.

Don's numbers showing a somewhat better DR with the XA20 vs XA10 seem to be close to my observations. I don't know what his number mean, exactly, but the differences between the cams sounds about right. All in all not much of an improvement over the XA10, but I'll take whatever I can get.

The lack of dynamic range is what I most miss from my GH2s.

[Don Palomaki]

It goes to the issues of the internal gamma, knee and black press/stretch setting and how they ultimatedly map the sensor output to the available 8-bit video levels. While these are adjustable on camcorders like the XH- and XF-series, we get what Canon programmed with the XA. FWIW: Canon literature says it was improved with the XA20 over the XA10.

Perhaps someone can run tests to demonstrate/evaluate deep shadow and full sun performance under controlled conditions. A chance to take advantage of October's bright blue weather!

[Don Palomaki]

A bit more in dynamic range.

This test was in mid morning direct sun and adjacent shade with a Macbeth chart.
Exposure set to 0DB gain, shutter and aperture to just put the bright white square in full zebra , (measured later to be 101 IRE). The chart was also shot in adjacent full shade at the same exposure. (1/250, f/4.4, ND1/8. 0dB gain)

The brightness ratio between sun and shade was about 4 stops as measured with an incident light meter.

In summary on my edit station and monitor I could see shadow differences between the two blackest squares on the Macbeth chart in the shade at an exposure that produced 100 IRE on the white square in the sun. As I understand it this translates to roughly 9 stops - chart plus shadow.

Snips from the two Macbeth charts attached for viewing pleasure. (What you see on your computer screen will depend on how your monitors and graphic card are setup.)