Page translated by Claude — switch to Italian to read the original article.

Light pollution: the Nemesis of every urban and suburban amateur astronomer.

Whether you are a visual observer or a photographer, one of the greatest obstacles to your passion is undoubtedly the rampant phenomenon of light pollution.

It does not affect only large cities, but also smaller urban centers, so much so that it can be stated without fear of contradiction that, in Italy, there is no longer an unspoiled sky.

And if you live, as in my case, in the Po Valley, then the situation becomes desperate: I live in Lombardy, a handful of km from the center of Brescia, under a perpetually illuminated sky.

For this reason the major manufacturers of astronomy filters have been offering, for some years now, several models of filters to counter light pollution: the latest addition in this field is the L-eNhance filter by Optolong.

Thanks to the new collaboration with Astrottica of Legnago, I was finally able to get my hands on one of these filters to put it to the test under the bright sky of Lombardy.

Operating principle

Before moving on to the actual review, it is worth spending a few words to describe the operating principle of this innovative filter.

Most light pollution filters are made so as to block (or reduce) certain wavelengths (typical of street lighting) and let the rest of the light spectrum through.

The basic philosophy of these filters is to remove only what is harmful while leaving everything else unaltered; they are therefore “broadband” filters.

The advantage of this approach is that the filter lets through a lot of light from most astronomical sources while altering their color balance only slightly.

Unfortunately, in recent years street lighting based on the use of LED technology has been spreading more and more; LEDs have a substantially continuous emission spectrum, so the classic anti-light-pollution filters are losing effectiveness because they are “too permissive”.

The L-eNhance filter tackles the problem from the opposite side: just like the narrowband filters commonly used with monochrome sensors, the filter blocks everything except the emission lines typical of astronomical objects.

In astrophysics, in fact, there are mainly 2 emission mechanisms:

Black-body emission from stars: it is a continuous emission that follows a well-defined curve depending on the temperature of the star.

Emission from ionized gas: it is a discrete emission that depends only on the type and ionization state of the gas.

It is precisely this type of emission that the L-eNhance filter targets.

To tell the truth, in my opinion, this filter is not a simple anti-light-pollution filter, it is much more: a true multi-narrowband filter designed for color sensors (although it might also have some use with monochrome sensors)..

As can be seen from the spectrum declared by the manufacturer, the filter lets through a narrow band corresponding to the Hα emission line at 656 nm and a slightly wider band that includes both the doublet of lines of doubly ionized oxygen (OIII) and the Hβ line of ionized hydrogen.

These lines make up the vast majority of the emissions of astrophotographic interest.

The filter is therefore comparable to narrowband filters and, as we will see, it shares their merits and flaws.

Given these premises, it is clear that with this filter it is possible to effectively image only a well-defined class of astronomical objects, the emission nebulae:

• HII regions
• Planetary nebulae
• Supernova remnants

It is instead totally ineffective on objects that have continuous-emission components:

• Galaxies
• Star clusters
• Reflection nebulae

On these objects the filter will give no advantage, on the contrary, it will be harmful as it will substantially alter their color balance.

Description of the filter

The filter under test is the 2" model mounted in a cell; on first analysis the filter appears well made, protected by a small box that is easy to open and with a thick layer of foam to protect the filter.

To the naked eye the filter appears with the mirror-like surface typical of narrowband filters, with a greenish-blue color in transmission reminiscent of the color of a visual OIII filter.

Unfortunately I own neither a spectroscope nor a spectrophotometer and was unable to verify the actual shape of the passband, but with a simple prism and a providential ray of sunlight I was able to verify, as can be seen in the following images, the double passband.

The simple prism used to analyze the filter
The solar spectrum obtained with the prism	The spectrum obtained by inserting the filter

Obviously this test is absolutely qualitative and does not allow the actual bandwidth of the filter to be estimated.

The test data

I wanted to carry out the test in realistic city-sky conditions.

Date of the test 20 September 2019.

I did the test literally from the courtyard of my house in Castenedolo, a few km from Brescia.

At the time of imaging the sky had a brightness of about 18.5 Mag/arcsec².

The camera used is a Baader-modified CANON 350D

The telescope a William Optics 110 FLT with a Type IV reducer/flattener that brings it to F/5.8

Mount 10Micron GM1000 HPS.

The shots were acquired using Astrophotography Tools, processed and developed with PixInsight 1.8.

The exposures without the filter were taken first, those with the filter were taken afterward; in the meantime the last-quarter Moon rose, making the test even more critical.

The object chosen for the test is the Pacman nebula NGC 281, chosen because it has abundant emission in both HII and OIII.

All the images were taken at 800 ISO (close to unity gain for the 350D) with an exposure time of 120"; the exposure time was chosen on the basis of the full-light image to avoid overly overexposing the sky background.

The data sets consist of

6 shots without the filter to be used as a reference

32 shots with the filter to quantitatively verify the improvement in image quality and to do a trial processing.

The RAW files were calibrated with 20 DARKs, 20 FLATs and 20 flat-DARKs.
to avoid introducing color casts due to the color of the flatbox, I preventively neutralized the individual RGB components of the master flat.

In this way the calibrated image will have exactly the same color composition as the original image.

The first impression

A very first impression, qualitative but significant, came already from the analysis of the single shots: the following images are two shots, one with the filter and the other without, obtained simply by developing the RAW files with the camera’s standard settings as one would do with any daytime image: no additional processing of any kind was applied.

Single 120" full-light shot Click for the enlarged version Normalized median value of the sky background on the calibrated image: 0.0073623*	Single 120" shot with L-eNhance filter Click for the enlarged version Normalized median value of the sky background on the calibrated image: 0.0004850*
*PixInsight represents intensity values by normalizing them to 1. This means that absolute black is equal to 0 and absolute white is equal to 1. since the image dynamic range is 16-bit, and therefore with a maximum of 65535, 0.0073623 is equivalent to 482 ADU.

Already this first example demonstrates the effectiveness of the filter in cutting down light pollution on this type of object: it is evident that the sky background in the unfiltered image, with the same exposure time and telescope, is about 15 times brighter than in the filtered image.

This corresponds to a considerable reduction of light pollution.

If, as the filter promises, the intensity of the nebulosity remained unaltered, this would translate into a considerable increase in the signal-to-noise ratio, as I will try to verify with the quantitative analysis.

The details of the analysis

Beyond any subjective opinion on the appearance of the images, it is important to define an objective method to understand whether and how much the image quality improves through the use of the filter.

The basic rule that every astrophotographer should ALWAYS follow is: The Signal-to-Noise ratio is always right.

The objective of the quantitative analysis was therefore to try to provide an objective assessment of S/N.

As mentioned above, the data analysis was carried out with PixInsight 1.8 using the Statistics process on a few suitably chosen previews on the image.

To estimate the S/N ratio it is necessary to measure the net signal of the nebula and compare it with the “noise” in the same area.

A good estimate of the signal is given by the “mean” function calculated within an area of the image; what matters is the net signal of the nebula cleaned of the contribution of the sky background.

The noise can instead be estimated thanks to the standard deviation in the same area of the image.

To do this, in PixInsight, two small previews are created on the image: one on the sky background, the other on the nebula.

The important thing is to choose the two previews so that they do not contain stars that would distort both the mean value and, above all, the standard deviation.

Thanks to Statistics it is possible to measure mean and standard deviation in the previews.

the estimate of the Signal-to-Noise ratio is given by the formula

S/N = (mean(Nebula)-mean(Sky))/StdDeviation(Nebula)

The higher the S/N, the better the image quality.

The processing was applied to the integration of 6 frames

Here is the result of the analysis carried out on the sum of 6 frames, respectively without the filter and with the filter.

The measurements were taken on the brightest part, near the center.

Once again the measurements refer to the value normalized to 1

As can be seen, the increase in the Signal-to-Noise ratio between the two images is appreciable (especially in the red, where the nebula has most of its emission); there is therefore an evident increase in image quality.

In reality, however, I would have expected something more, given the 15-fold reduction in the sky background value.

What leaves me most perplexed is the intensity value of the nebula between the image with the filter and the one without: given the declared shape of the filter’s passband, I would have expected the net intensity of the nebula to vary little (the filter should be substantially transparent to the Hα, Hβ and OIII wavelengths); instead, in the red it goes from 6.77×10^-3 without the filter to 3.50×10^-3 with the filter, a reduction of about 48% of the filtered image compared to the unfiltered one.

With the little data at my disposal, however, I will not venture into possible explanations of this behavior, which nevertheless deserves further investigation.

What is interesting, instead, is the direct comparison between the filtered image (on the right) and the unfiltered one (on the left), visible in the following image.

As can be seen, the filter, thanks to its narrow passband, considerably increases the contrast between stars and nebula, allowing it to be detached from the sky background.

It also significantly reduces the light-pollution gradient, making the processing phase decidedly easier; from this point of view the filter does exactly what it promises.

Trial processing

As a final test I processed the sum of 38 shots of 120 seconds for a total of about one hour and a quarter of exposure.

The processing carried out was actually rather simple:

Manual white balance, histogram transformation, curve adjustment and denoise.

The result obtained is decidedly interesting considering the imaging location and the camera (the Canon 350D), now almost 15 years old.

Conclusions

The L-eNhance filter proved to be decidedly effective in cutting down light pollution, even if perhaps it fell a little short of my expectations.

The quantitative analysis, in fact, showed a clear improvement in the S/N ratio under otherwise equal conditions; however, a clear decrease in the brightness of the nebula also emerged, something that, given the design of the filter, should not happen.

Before venturing interpretations of this phenomenon I reserve the right to do further tests.

The advantage of this filter is undoubtedly that of significantly increasing the contrast between stars and nebulosity, making it easier to extract details.

The main disadvantage, instead, consists in the marked change in the color of the stars, which is difficult to manage; but such behavior is entirely expected from the design of the filter and I would therefore not define it as a real “defect”.

Another negative point, linked to the design of the filter, is the severe limitation on the type of objects that can be imaged, being limited to emission nebulae.

This fact is only implicitly hinted at on the technical data sheet on the manufacturer’s website: it would have been nice, instead, to find an explicit mention of the type of objects for which the filter is recommended.

In fact, although it is clear, observing the shape of the passband, not all astrophotographers have the astrophysics knowledge needed to draw these conclusions.

Stay tuned, more tests on this interesting product in the future.