CAMERA REVIEW SITES
Dave  Etchells’   http://www.imaging-resource.com/DIGCAM01.HTM  (Includes side-by-side comparisons)
Steve's DigiCams, http://www.steves-digicams.com/hardware_reviews.html (in Florida)
Phil Askey’s  http://www.dpreview.com/reviews/ (in England)
http://www.megapixel.net/html/issueindex.php?lang=en (Possibly in France)

ARTICLES
Jump to: AF Test Targets
Jump to: Suggested Extended-range Arrangement
Jump to: CMOS vs. CCD vs. MP
Jump to: SENSOR SIZES

To use this chart, frame the test picture exactly between the top an bottom tick marks (200 mm).  Read the number where all the lines closest together can be seen and multiply by 100.  This will give you the LW/PH (Line Widths/Picture Height) resolution and is the method used in all the review web sites above.

An example (2 MB) file of how well the Fuji S602 can resolve this chart is at: http://www.gpsinformation.org/jack/photo-test/pics/fuji_602iso.jpg  Here the resolution is read out from gradually converging lines (usually nine).  It appears that the Fuji resolved about 1250 LW/PH.

NOTE:  This chart can no longer test the current high-resolution digital cameras.  See below for an alternate solution.

                                      Resolution Results, LW/PH -Assending Resolution

                Zoom    MPx        Horiz.       Vert.    Lines/mm equivalent
Kodak DC260     (3x)    1.6          660         720           15
Nikon 990       (3x)    3.3         1000         NA            21
Nikon 5000      (3X)    5.0         1250        1250           26
Nikon 5700      (8x)    5.0         1300        1300           27
Fuji S602       (6x)    6.0         1350        1350           28
Canon SD300     (3x)    4.0         1350        1350           28
Canon Pro1      (7x)    8.0         1650        1600           33
Nikon 8800     (10x)    8.0         1650        1650           34
Canon SD1100IS  (3x)    8.0         2100        2100           44
Canon S95       (4x)   10.0         2200        2200           46
Canon G-12      (5x)   10.0         2206        2206           46
Canon 300HS     (5x)   12.0         2250        2250           47
Canon A3300IS   (5x)   16.0         2400        2400           50
Canon T3i       (3x)   18.0         2476        2476           52

Estimated best film resolution      2400        2400           50

TEST TARGETS –AIR FORCE METHOD

Download a high-resolution copy of this target already sized to print on 4x6 photo paper (HERE).
The specifications for the USAF targets can be found at:  http://www.efg2.com/Lab/ImageProcessing/TestTargets/#USAF1951

Here the terminology is different from the ISO chart, only lines are counted.  A test pattern that is 1 line per mm on the USAF chart would be called 2 "line widths" per mm on the ISO chart.  Just remember that one line per mm on the USAF charts equals two ISO line widths per mm.

The printed part of the small targets are only 64x64mm and have "pairs" of  three horizontal and three vertical lines, 1 line-per-mm (l/mm) apart at Group “0,1”).   Each successive Group has twice the lines per mm as the previous Group.  The Groups are divided into six pairs, each pair being 1.12246 (the sixth root of 2.0) more lines per mm than the previous pair.  The chart requires a supplied table to read the resolution, but could be calculated with arithmetic, if one didn't have the table.

The instructions are to measure the distance from the target to the lens in mm, divide by the Focal Length and subtract 1.0.  This is the working Magnification.  Determine the lines per mm resolution visible on the target and multiply by M.  For 35 mm film cameras, this is the end of it.

However, with a digital camera, the Focal Length is based on a chip smaller than a 35 mm frame (which is 24 x 36 mm).  Thus one must divide the distance to the target by the DIGITAL lens Focal Length (not the "35 mm equivalent") and then multiply by the resolution read out.

NOTE: The Focal Length of a digital test shot can be read out from most modern PC .jpg viewers "Properties", and Nikon also supplies a file that has this information in it.

CONVERTING AIR-FORCE CHART READINGS TO PIMA/ISO 'LW/PH' READINGS
Analogy to USAF film "lines-per-mm"
The PIMA/ISO test target being 200 mm high, has 1 line per mm at "4" (400 line widths per 200 mm Picture Height or 200 Line Widths per 200 mm).  Thus, divide the LW/PH results by 2.0 and divide again by 24 to get the l/mm equivalent.  The PIMA/ISO target goes up to 2,000 LW/PH (1,000 L/PH) -or 42 l/mm (for 24 mm film frames).  This would say that the  PIMA/ISO chart cannot resolve the better film lenses of 50 l/mm resolution.

Suggested Extended-range Arrangement:

......
Download this color chart already sized to print on 5x7 photo paper (HERE).

NOTE:  The ISO chart is no longer able to test the resolution of the newer cameras with 12MP and higher, since it can only test up to 2,000 LW/PH.  However there is a simple solution.  Just paste the chart on a larger board and move the 200 mm top and bottom tick marks to 400 mm apart  Then multiply the resolved readings by 200.  This larger arrangement allows one to include four AF test targets and four color charts.  The AF targets have been cut into 4" squares.

Measuring the AF-target resolutions:
Since the AF targets have one line per mm pairs at 0,1 (2 Line Widths per mm), and the Picture Height is 400 mm, then Group 0,1 resolves 800 LW/PH and 1,1 resolves 1,600 LW/PH.  Here are the higher resolutions in LW/PH:  1,2 =  1,800; 1.3 = 2,015; 1,4 =  2,260; 1,5 =  2,540; 1,6 = 2,850.

Dave's New Imaging Resources MULTI-TARGET:
Figuring out just how high the test target really is:  Since the 60 mm ruler in a 3,000 pixel-high test shot is 231 pixels long, then  60 mm per 231 pixels x 3,000 pixels per Picture Height  = 780 mm per Picture Height (PH).  And since the AF target 0,1 is 2 LW (above), then Group 0,1 is 1,560 LW/PH.  Here are other resolutions:
0,1 = 1,560; 0,2 = 1,751; 0,3 = 1,965; 0,4 = 2,206; 0,5 = 2,476; and 0,6 = 2,780.

CMOS vs. CCD vs. MP:
The Canon 1100 IS uses an 8MP CCD sensor while the Canon 300 HS (HighSensitivity) uses a 12MP CMOS sensor, and the Canon A3300, 16MP CCD..  A low-light photo was taken with the 1100 IS at f 2.8, 1/6 Sec., and ISO 200.  The same scene photographed by the 300 HS was taken at  f 2.8, 1/20th Sec., and ISO 800.  The ISO value was four times as high and the shutter speed, 3 1/3 times faster.

As for resolution,  a 12MP camera resolution should be 22% more than an 8MP camera, but the CMOS pictures taken with the CMOS sensor don't seem to be as sharp as with the CCD sensors. Contrast may be a problem for the CMOS sensors.  The CMOS sensors score well with high-contrast test targets, but in the 'real world' have trouble resolving subtle shade differences.

The below crops are of the actual pixels magnified x2.  For an easier comparison,  the 8MP and 12MP shots were increased in size, so that the bricks are equal to the size of the 16MP camera bricks.

THUMBNAIL

SD 1100IS 8MP  ° 300HS 12MP  ° A3300 16MP

As compared to the SD 1100IS, the 12 and 16MP cameras both show a loss of delicate shading, and (surprisingly) show no increase in detail.  Apparently, the only advantage of the 16MP camera over the 8MP camear is its 5x zoom instead of 3x.

SENSOR SIZES:
For a description of Sensor sizes, see:  http://www.dpreview.com/news/2002/10/7/sensorsizes
Also, for a visual graphic of relative sensor sizes see (HERE).  This file had been modified so that, if you print it, the sensors are their actual size.
The "1.5"  CMOS are 14.0mm x 18.7mm
The "4/3"  CCDs are 13.5mm x 18mm
The "1.0"  CCDs are 9.6mm x 12.8mm
The "1/1.7" CCDs are 5.7mm x 7.6mm
The "1/1.8" CCDs are 5.319mm  x 7.176mm.
The "1/2.3" CCDs used in some 12 and 16MP cameras are only 4.55 mm x 6.17mm, resulting in inferrior performance.
The larger CCD, the lower the ratio -and would be 1.0 if the CCD were to be the same height as a film frame.  The Canon EOS-1Ds have a CCD of this size.

WHEN WILL DIGITAL EQUAL FILM -A Personal Analysis:
If one looks  at: http://www.imaging-resource.com/DIGCAM01.HTM  Dave will report the LW/PH (Line Widths/Picture Height) resolution of cameras from 1.3 to 21 MP.  It's interesting that from 1.3 to 5.2 MP, resolution has only doubled -which is mathematically correct.  It will take a 21 MP camera to double the resolution again.

Determining film's comparable resolution of LW/PH is hard come by, but this site attempted to do it:
http://www.crystalcanyons.net/Pages/TechNotes/FilmVsCP5000.shtm
Here John Hart has come up with 75 lines/mm on the film (under very special conditions), which would be equivalent to 3,600 LW/PH (24 mm film 'height' x 75 x 2 = 3,600).

So, if normal print processing of film is more like 2,400 LW/PH then a 14 MP camera should do it.   There is already the Kodak DCS Pro 14n which has 14 MP.  The price for such a camera has come from $4,600 to about $1,000, but still hard come by for most people!  NOTE: 2,400 LW/PH is a higher resolution than the ISO chart can resolve -which is only 2,000 LW/PH.  However, see (HERE) for a solution.

Now comes the hard part of finding an affordable printer that can print this resolution. See: Printer Resolution Test Procedures (HERE).

See Fujifilm Finepix S602 Gallery  for some representative test shots with this camera.  These pics are in the 6 MegaPixel/Normal mode (about 1 MB per pic) on this page.  Tests of the same objects in the TIFF mode (17 MB per pic) and 6 MegaPixel/Fine mode (2 MB per pic) revealed the no increase in quality up to 800% viewing before a quality difference could be seen.

Jack Yeazel