There are two main limits in film - the contrast resolution (depending on noise) and the spatial resolution. For NDT films I will try to see where are the limits. As "measurement instrument" I am using two prototypes of the KowoScan film scanner, one with 27µm pixel size and the second one with 46µm (which is now the KowoScan X50).
I start with the film density and measure which densities the film digitizer can still resolve and/or which density the film can still deliver.
The scans of the four films with the BAM5 weld seam were done with one calibration of the digitizer (prototype of the KowoScan X50) within half an hour.
My research was greatly assisted by Dr. Zscherpel from BAM, who exposed the various films of the BAM5 weld seam using sometimes exotic exposure times and conditions. For evaluation I used the KowoScan software, which displays the density out of the gray values in the lower right corner and offers a high sophisticated filter for showing even the smallest details in an image (called Quality Control function in the KowoScan software).
The first film is the well know BAM5 weld (8mm steel), which was exposed on December 12, 1997 at the BAM as reference with a nominal density of D=4.5.

Beside the weld the density is D=4.49 - which is close to the nominal value of D=4.5 (see mouse pointer in the image).

To see the details I used the filter of the KowoScan software

A lot of fine crack can be seen in the image. The white lines are from the FE13 wire set.

There is a nose at the edge and we can see there a small indication (red circle).

In the year 2005, I took a sabbatical at BAM with the task of comparing the new DDA technology (in 2005) with classic film and measuring whether the DDA technology is equivalent.
The reference was the film from above, which had to be beaten.
In April 2005 I took an image with a prototype of a PerkinElmer XRD1620 AM3 using a Microfocus tube (210kV, 0.25mA) with magnification 4 integration of 4000 frames with 1s each (more than one hour exposure time). The software for image capture and for display now is IMAGE 3500 from YXLON. A similar area as from above is cut out:

Without a filter the images show no big differences.

Using the Special2 filter in IMAGE 3500 (similar to the KowoScan filter) we get a very clear and detailed view of the seam and the cracks:


If we now concentrate on the nose we could see small indications - we assumed porosity. The cracks show fine ramifications.

You even can see the movement of the welder during welding with the circular arc-shaped structures in the image.
The porosity and these structures were never been seen on a film - not on a light box and not with a film digitizer.
Since then the reference has changed from the film to the DDA.

The first film, which was exposed specifically for this study in November 2021, is a D2 film that was exposed without lead foils, and the distance between focal spot and film was increased to 1700 mm to rule out unsharpness due to geometry or the lead foils. The film was exposed for 90 minutes with 120 kV and 7 mA to a density of D=3.2.

At the mouse pointer we measure a density of D=3.1.

To see the details the KowoScan filter is applied

Now we can see the fine crack much clearer.
Magnifying to the nose we can see now the porosity

and also some fine ramifications.

The film digitization standard is designed for a density up to D=4.5. To test the limits of the film digitizer, the next film was exposed to a density of about D=5.2. With 120kV and 7mA, the film was exposed for 135 minutes at a distance of 900mm, using 3mm lead foils.

With the KowoScan software we measure a density of D=5.3 at the mouse pointer.

To see the details we applied the KowoScan filter again:

In the filtered image we see nearly the same as with the D=3.2 film.

To see if the unsharpness due to the geometry and lead foils will spoil the details we cut out the region with the nose again:

I can nearly see the same porosity and ramifications as in the D=3.2 film image.

But the message is:
This film scanner can operate at a density D=5.2 and offer the same image quality as with a much lower density. There must be enough light to penetrate the areas with density D=5.3 (which is a factor of 200,000 less light than for density D=0)!

The next film was exposed "over night" with a nominative density of D=6.2. That's blacker than black.
To safely look at the sun, you need a special film with an optical density of at least D=5.0 for visual observation. The film with a density of D=6.2 has a fifteen times higher density and you may have problems to see the sun through this film .

The digitizes film image looks similar to the other two films with "lower" densities.

The density at the mouse pointer is D=6.46 - but this is beside the weld at 8mm flat steel.

To see some details I used the KowoScan filter

and the image of the weld itself shows all the cracks very clear.


Noise dominates the image at the edge of the weld seam, with the result that the porosities are no longer visible. However, the fine branches of the cracks are all present.
I am not quite sure if the noise is "produced" by the digitizer or if this is in the film itself. Please have in your mind that we are talking about densities >D=6!

But the message here is:
The film digitizer can scan films with densities higher than D=6.0 or - in a nutshell - there is no density limit with this digitizer.