Panasonic ZS50/TZ70


In this Section:

The Panasonic ZS50/TZ70 compared with the ZS40/TZ60

Testing of the Resolution and Noise of the ZS50 compared with the ZS40

Test Images

Test Results


The Panasonic ZS50/TZ70 compared with the ZS40/TZ60

The ZS50/TZ70 is similar in many respects to the previous model, the ZS40/TZ60, with both having the same 24-720mm EFL lens and a 1/2.3” sensor.  Since the lens is the same on both models, they have the same widest available apertures at different focal length settings, as well as the same macro and telemacro characteristics, as described earlier for the ZS40 here.

The major change on the ZS50 is that its sensor has 12MP (4000x 3000 pixels) while the ZS40 has 18MP (4896 x 3672 pixels).  This means that the size of the pixels is larger for the ZS50, being 1.52µm wide for the ZS50 compared with 1.24µm for the ZS40.  In area terms this equates to 2.31 sq µm for the ZS50 and 1.53 sq µm for the ZS40, so the pixels of the ZS50 are 50% larger in area than those of the ZS40.

All other factors being equal, the larger pixels for the ZS50 would mean less noise in the image for the same ISO setting.  However, fewer pixels also means, other factors being equal, a lower resolution for the ZS50.  It was seen from the previous testing that in general for 1/2.3” sensors, the centre resolution in LPH (lines per picture height) at or near full WA and base ISO is about equal to the image height in pixels divided by 1.5.  For the ZS40 this was 2400 LPH, and for the ZS50 it would be predicted to be 3000/1.5 = 2000 LPH.  There is however an interaction between resolution and noise, because higher noise will tend to lower the resolution, other factors being equal.  With regard to diffraction, the greater iresolution which results from the increased number of pixels on the sensor of the TZ60 is offset to some extent by an increase in the effect of diffraction.

Testing of the Resolution and Noise of the ZS50 compared with the ZS40

I measured the resolution as described previously here.  To the chart board I also added a copy of the black stepped triplet lines chart developed more recently by Dr J. C. Brown as described here.  For the measurement of noise, I included a greyscale chart on the chart board.  A view of the chart board is shown below:

Chart board used for the tests

I used mid-grey, the middle band on the greyscale chart which extends across both the upper and lower scales, for the noise measurements.  The tool provided in PSE allows direct readings from the luminosity histogram of a selection of the mid-grey band in the test image.  The standard deviation (SD) of the luminosity represents the amount of noise in the image.

For assessing the noise in the RAW files, so as to provide equal and minimum processing I used an early version of Photo Ninja which did not have inbuilt profiles for either the ZS40 or the ZS50.  That enabled the RAW files to be demosaiced only with minimum generic colour correction.  I found it necessary to apply that minimum colour correction because the uncorrected demosaiced image had a very strong green cast which would interfere with the noise measurements.

Under the same lighting conditions I compared the resolution and noise of both models at 40mm EFL and the resolution at 400mm EFL indoors using a tripod with diffused light from two daylight (5500º K) lamps at 110 lux on the chart board.  I also conducted tests to compare the two models for resolution at their maximum EFL of 720mm.  The 720mm tests were conducted outdoors to obtain the longer camera to chart distances that were required.

The resolution tests require the camera to chart distance to be adjusted so that the scale line on the chart is equal to 400 pixels in the image.  That means that since the two models have different pixel counts, those distances differ.  So as to provide direct IQ comparisons between the models, I also took additional images at base ISO using the same tripod position (camera to chart distance) for both models.  I also took images after adjusting the FL of the ZS50 to give about the same framing as the ZS40 while keeping the distance the same.  In addition, I also took images with both models set to the same FL and adjusted the distance of each to give the same framing for both.

On the following page I have provided the full sized test images, as JPGs and RAW files, which can be downloaded from the links.

Continue to Test Images

Continue to Results

Return to Home Page