http://www.clarkvision.com/articles/evaluation-canon-7d
R. N. Clark Email contact (is encoded to prevent spam):
has the following form: username@qwest.net where
username is rnclark
This page shows an analysis of noise, dynamic range, and full well capacity of a Canon 7D camera.
Procedures for performing this analysis are described in: Procedures for Evaluating Digital Camera Noise, Dynamic Range, and Full Well Capacities; Canon 1D Mark II Analysis
The lowest possible noise from a system detecting light is the noise due to Poisson statistics from the random rate of the arrival of photons. This is called photon statistics, or photon noise. Noise from the electronics will add to the photon noise. Noise in Canon 7D images is limited by photon statistics at high signal levels and by electronic noise from reading the sensor (called readout noise) and noise from the downstream electronics at very low signal levels. In the case of high signal levels, a system that is photon statistics limited enables us to directly measure how many photons the sensor captures, and by increasing the exposure, we can determine how many photons are required to saturate the sensor. That is called the full well capacity, or simply, maximum signal capacity. With data on the lowest noise to the highest signal, we can then determine the dynamic range of the sensor.
Results: Canon 7D sensor analysis
Table 1
-------------------------------------------------
Maximum Measured
ISO Gain Read Noise signal Dynamic range
e/DN (electrons) (electrons) stops
100 2.15 12.1 24800 11.0
200 1.23 7.6 16300 11.1
400 0.61 4.9 8150 10.7
800 0.31 3.8 4080 10.1
1600 0.15 3.0 2040 9.4
3200 0.077 2.7 1020 8.6
6400 0.038 ~2.7 510 7.6
Pixel pitch: 4.3 microns.
18 megapixels.
S/N on 18% gray card, ISO 100 = 67.
Sensor Full Well Capacity at lowest ISO: 24,800 electrons.
Sensor dynamic range = 24800/2.7 = 9,185 = 13.2 stops.
ISO at unity gain (scaled to 12 bit) = 984 (14-bit unity gain = ISO 246).
Low Light sensitivity Factor: 364.
Apparent Image Quality, AIQ = 60.
-------------------------------------------------
Table 2: ISO 100 Sensor Data and Analysis
Offset= 2048
Model gain = 2.15 e/DN
Model read noise = 12.2 electrons
Observed Observed Observed
min max mean 2-img std noise S/N signal ISO relative S/N S/N
file (DN) (DN) (DN) (DN) (DN) (electrons) exposure model obs/model
_MG_0557 11535.00 11535.00 11535.00 0.00 sensor saturated
_MG_0558 11536.00 11536.00 11536.00 0.00 sensor saturated
_MG_0559 9951.00 10993.00 10477.52 99.15 70.11 149.45 22334.6 100 1.000000 149.6 1.00
_MG_0560 10031.00 11139.00 10573.92 99.15 70.11 150.82 22747.5 100 1.000000 150.3 0.99
_MG_0561 6150.00 7018.00 6582.11 79.91 56.51 116.49 13569.3 100 0.645162 118.3 0.98
_MG_0562 6248.00 7102.00 6669.38 79.91 56.51 118.03 13931.4 100 0.645162 119.1 0.98
_MG_0563 3702.00 4415.00 4081.06 62.58 44.25 92.22 8504.7 100 0.384616 92.9 0.99
_MG_0564 3747.00 4377.00 4087.87 62.58 44.25 92.38 8533.2 100 0.384616 93.0 0.99
_MG_0565 2264.00 2781.00 2566.34 49.42 34.95 73.43 5392.7 100 0.250000 73.3 1.00
_MG_0566 2281.00 2830.00 2573.59 49.42 34.95 73.64 5423.2 100 0.250000 73.4 1.00
_MG_0567 1351.00 1752.00 1582.42 38.75 27.40 57.76 3336.0 100 0.162602 57.1 1.01
_MG_0568 1359.00 1743.00 1580.54 38.75 27.40 57.69 3328.1 100 0.162602 57.1 1.01
_MG_0569 852.00 1140.00 996.19 30.57 21.62 46.09 2123.9 100 0.096618 44.8 1.03
_MG_0570 849.00 1097.00 980.93 30.57 21.62 45.38 2059.3 100 0.096618 44.4 1.04
_MG_0571 512.00 710.00 620.47 24.43 17.27 35.92 1290.5 100 0.062500 34.6 1.04
_MG_0572 503.00 711.00 618.07 24.43 17.27 35.78 1280.5 100 0.062500 34.6 1.04
_MG_0573 320.00 457.00 390.52 19.96 14.12 27.66 765.3 100 0.040486 26.7 1.04
_MG_0574 312.00 461.00 394.69 19.96 14.12 27.96 781.8 100 0.040486 26.9 1.03
_MG_0575 188.00 301.00 243.33 16.56 11.71 20.78 431.8 100 0.024096 20.2 1.03
_MG_0576 189.00 316.00 254.92 16.56 11.71 21.77 474.0 100 0.024096 20.8 1.00
_MG_0577 112.00 198.00 156.03 13.92 9.84 15.86 251.4 100 0.015624 15.2 1.04
_MG_0578 114.00 210.00 156.43 13.92 9.84 15.90 252.7 100 0.015624 15.3 1.04
_MG_0579 66.00 142.00 101.54 12.53 8.86 11.46 131.3 100 0.010132 11.4 1.01
_MG_0580 61.00 136.00 98.79 12.53 8.86 11.15 124.3 100 0.010132 11.2 1.03
_MG_0581 30.00 105.00 64.58 10.84 7.67 8.42 70.9 100 0.006024 8.2 1.03
_MG_0582 31.00 101.00 64.56 10.84 7.67 8.42 70.9 100 0.006024 8.2 1.03
_MG_0583 13.00 75.00 41.63 9.91 7.01 5.94 35.3 100 0.003906 5.8 1.02
_MG_0584 11.00 77.00 41.87 9.91 7.01 5.98 35.7 100 0.003906 5.8 1.02
_MG_0585 -4.00 63.00 25.78 9.13 6.46 3.99 15.9 100 0.002532 3.9 1.03
_MG_0586 -8.00 56.00 23.48 9.13 6.46 3.64 13.2 100 0.002532 3.6 1.12
_MG_0587 -11.00 49.00 17.60 8.96 6.33 2.78 7.7 100 0.001506 2.8 1.00
_MG_0588 -15.00 55.00 17.53 8.96 6.33 2.77 7.7 100 0.001506 2.8 1.01
_MG_0589 -14.00 44.00 11.67 8.56 6.06 1.93 3.7 100 0.000976 1.9 1.01
_MG_0590 -17.00 41.00 9.91 8.56 6.06 1.64 2.7 100 0.000976 1.6 1.18
_MG_0591 -23.00 35.00 6.38 8.40 5.94 1.07 1.2 100 0.000634 1.1 1.00
_MG_0592 -27.00 35.00 6.62 8.40 5.94 1.11 1.2 100 0.000634 1.1 0.96
_MG_0593 -24.00 32.00 3.97 8.25 5.83 0.68 0.5 100 0.000376 0.7 1.00
_MG_0594 -25.00 39.00 4.17 8.25 5.83 0.71 0.5 100 0.000376 0.7 0.95
_MG_0595 -25.00 43.00 2.30 8.17 5.78 0.40 0.2 100 0.000244 0.4 1.00
_MG_0596 -25.00 30.00 2.72 8.17 5.78 0.47 0.2 100 0.000244 0.5 0.85
Table 3: ISO 200 Sensor Data and Analysis
Offset= 2048
Model gain = 1.23 e/DN
Model read noise = 7.6 electrons
Observed Observed Observed
min max mean 2-img std noise S/N signal ISO relative S/N S/N
file (DN) (DN) (DN) (DN) (DN) (electrons) exposure model obs/model
_MG_0597 13255.00 13255.00 13255.00 0.00 sensor saturated
_MG_0598 13255.00 13255.00 13255.00 0.00 sensor saturated
_MG_0599 13255.00 13255.00 13255.00 0.00 sensor saturated
_MG_0600 13255.00 13255.00 13255.00 0.00 sensor saturated
_MG_0601 12278.00 13257.00 13088.49 146.62 103.67 126.25 15938.0 200 1.000001 126.7 1.00
_MG_0602 12607.00 13257.00 13241.76 146.62 103.67 127.72 16313.5 200 1.000001 127.4 0.99
_MG_0603 7480.00 8918.00 8235.19 125.57 88.79 92.75 8602.8 200 0.596155 100.4 0.92
_MG_0604 7434.00 8797.00 8195.07 125.57 88.79 92.30 8519.2 200 0.596155 100.1 0.93
_MG_0605 4505.00 5541.00 5030.62 97.71 69.09 72.81 5302.0 200 0.387500 78.3 0.93
_MG_0606 4569.00 5560.00 5111.64 97.71 69.09 73.99 5474.1 200 0.387500 78.9 0.92
_MG_0607 2752.00 3497.00 3164.79 76.26 53.93 58.69 3444.1 200 0.252033 61.9 0.95
_MG_0608 2775.00 3484.00 3178.18 76.26 53.93 58.93 3473.3 200 0.252033 62.1 0.95
_MG_0609 1682.00 2203.00 1949.83 59.27 41.91 46.52 2164.4 200 0.149758 48.4 0.96
_MG_0610 1677.00 2240.00 1975.01 59.27 41.91 47.12 2220.7 200 0.149758 48.7 0.96
_MG_0611 1023.00 1414.00 1232.22 46.87 33.14 37.18 1382.4 200 0.096875 38.2 0.97
_MG_0612 1022.00 1425.00 1239.68 46.87 33.14 37.41 1399.2 200 0.096875 38.3 0.97
_MG_0613 634.00 919.00 786.45 38.00 26.87 29.27 856.4 200 0.062753 30.2 0.97
_MG_0614 636.00 907.00 785.94 38.00 26.87 29.25 855.3 200 0.062753 30.2 0.97
_MG_0615 379.00 602.00 481.95 30.15 21.32 22.60 511.0 200 0.037349 23.2 0.97
_MG_0616 390.00 620.00 502.41 30.15 21.32 23.56 555.3 200 0.037349 23.8 0.95
_MG_0617 228.00 390.00 308.63 24.25 17.15 18.00 324.0 200 0.024217 18.2 0.99
_MG_0618 236.00 386.00 309.91 24.25 17.15 18.08 326.7 200 0.024217 18.2 0.99
_MG_0619 130.00 262.00 189.72 19.88 14.06 13.49 182.1 200 0.015705 13.7 0.99
_MG_0620 134.00 268.00 202.95 19.88 14.06 14.44 208.4 200 0.015705 14.2 0.95
_MG_0621 63.00 186.00 125.11 17.05 12.06 10.38 107.7 200 0.009337 10.6 0.98
_MG_0622 75.00 187.00 125.21 17.05 12.06 10.39 107.9 200 0.009337 10.6 0.98
_MG_0623 23.00 141.00 82.84 14.56 10.30 8.05 64.7 200 0.006054 8.1 1.00
_MG_0624 23.00 150.00 76.85 14.56 10.30 7.46 55.7 200 0.006054 7.7 1.05
_MG_0625 -12.00 122.00 50.34 12.47 8.82 5.71 32.6 200 0.003925 5.7 1.01
_MG_0626 -6.00 113.00 45.77 12.47 8.82 5.19 27.0 200 0.003925 5.3 1.08
_MG_0627 -22.00 81.00 30.17 10.92 7.72 3.91 15.3 200 0.002334 3.8 1.03
_MG_0628 -22.00 80.00 28.61 10.92 7.72 3.70 13.7 200 0.002334 3.6 1.07
_MG_0629 -36.00 71.00 19.05 10.36 7.33 2.60 6.8 200 0.001513 2.6 1.00
_MG_0630 -39.00 85.00 20.79 10.36 7.33 2.84 8.1 200 0.001513 2.8 0.93
_MG_0631 -34.00 63.00 14.53 10.01 7.08 2.05 4.2 200 0.000983 2.1 1.00
_MG_0632 -33.00 63.00 14.56 10.01 7.08 2.06 4.2 200 0.000983 2.1 1.00
_MG_0633 -45.00 67.00 9.85 9.55 6.75 1.46 2.1 200 0.000583 1.4 1.01
_MG_0634 -44.00 57.00 8.66 9.55 6.75 1.28 1.6 200 0.000583 1.3 1.13
_MG_0635 -39.00 62.00 5.87 9.29 6.57 0.89 0.8 200 0.000378 0.9 1.00
_MG_0636 -41.00 52.00 6.28 9.29 6.57 0.96 0.9 200 0.000378 1.0 0.94
Table 4: Read Noise Data and Analysis
file min max mean 2-img std noise electrons gain(e/DN) ISO
_MG_0427 2022.00 2088.00 2047.13 7.96 5.63 12.102 2.150 100
_MG_0428 2021.00 2078.00 2047.82 7.96 5.63 12.102 2.150 100
_MG_0429 2004.00 2110.00 2048.31 8.74 6.18 7.598 1.230 200
_MG_0430 2001.00 2100.00 2047.84 8.74 6.18 7.598 1.230 200
_MG_0431 1954.00 2147.00 2047.19 11.25 7.95 4.852 0.610 400
_MG_0432 1957.00 2157.00 2047.57 11.25 7.95 4.852 0.610 400
_MG_0433 1879.00 2298.00 2047.00 17.16 12.13 3.762 0.310 800
_MG_0434 1864.00 2253.00 2046.92 17.16 12.13 3.762 0.310 800
_MG_0435 1736.00 2477.00 2046.68 28.68 20.28 3.042 0.150 1600
_MG_0436 1744.00 2451.00 2047.99 28.68 20.28 3.042 0.150 1600
_MG_0437 1531.00 2637.00 2047.52 49.94 35.31 2.719 0.077 3200
_MG_0438 1379.00 2611.00 2046.91 49.94 35.31 2.719 0.077 3200
Data acquired by Peter A. Hawrylyshyn, M.D. November, 2009
Analysis by R. N. Clark December 31. 2009
Conclusions
The data shown here for the Canon 7D shows that the camera is operating at near perfect levels for the sensor. This means that for high signals, noise is dominated by photon statistics. The model assumes photon noise and a noise floor given by the read noise value. The model predicts the observed data to within a few pecent, proving the camera is photon noise limited over most of its range. Noise at low signal levels is very good, among the lowest cameras measured. There is evidence that Canon has improved the fill factor as claimed in press reports, but on a pixel level, this is offset by the smaller pixel area. To achieve higher signal-to-noise ratio images, a higher quantum efficiency sensor with a larger full well would be needed. Without increasing the full well capacity, the maximum signal-to-noise ratio would not improve. Improvements in quantum efficiency of about 2 to 3 are possible (Reference 2), but currently that would require much more expensive detectors, such as back-side illuminated CCD or CMOD sensors. Note some DSLRs have maximum full well capacities on the order of 50,000 electrons (e.g. Canon 20D) to 80,000 electrons (e.g. Canon 1D Mark II), so the Canon 7D is on the low end of full well capacity, as is expected from its small pixel size. For comparison with a larger pixel DSLR, compare to the Tables and graphs in Digital Sensor Performance Summary.
References
1) CCD Gain. http://spiff.rit.edu/classes/phys559/lectures/gain/gain.html
4) http://www.photomet.com/library_enc_fwcapacity.shtml
Notes:
DN is "Data Number." That is the number in the file for each pixel. I'm quoting the luminance level (although red, green and blue are almost the same in the cases I cited).
16-bit signed integer: -32768 to +32767
16-bit unsigned integer: 0 to 65535
Photoshop uses signed integers, but the 16-bit tiff is unsigned integer (correctly read by ImagesPlus).
Home Page: ClarkVision.com
Back to: Digital Imaging Information index on this site: http://www.clarkvision.com/articles
First published December 31, 2009.
Last updated December 31, 2009.