Optolong L-eXtreme Filter Comparison

by Jim Thompson, P.Eng

Test Report – August 25th, 2020

Introduction:
The light pollution filter market continues to grow, with new offerings being released every
couple of months. One of the latest filters to be released is the new L-eXtreme filter from the
Chinese company Yulong Optics Co. Ltd. under their Optolong brand. I have a fair amount of
experience with Optolong brand filters, all of which started in 2015 with a request directly from
Optolong for me to test some of their filters for them*. I was impressed by Optolong’s filters
then and I am still impressed today. Because of my history of testing Optolong filters I have a
library of their products available to which I can compare any new offerings. This is exactly
what I have summarized in this report: a comparison between the new L-eXtreme filter and
Optolong’s other filter models. Figure 1 shows the spectral response of the new L-eXtreme filter
that was released in North America just this past June. The filter is a refinement of their popular
L-eNhance filter, having even narrower pass bands around O-III and Hα.

Figure 1 Spectral Response of the New L-eXtreme Filter

* You can read the test report here:

http://karmalimbo.com/aro/reports/Optolong_preliminary%20filter%20test%20report_26Aug2015.pdf

Objective:
As indicated below in Table 1, Optolong has a large variety of filters available. The table lists
only the filters that I have on-hand for my testing. Optolong also sells a variety of narrowband
filters (O-III, Hβ, Hα, SII, etc.), as well as some specialty filters. I have chosen to limit my
comparison to the filters that pass multiple nebula emission wavelengths since they are of the
most interest to one-shot colour (OSC) users. The retail prices of these offerings vary widely,
which raises the question: “Is the cost of a filter justified by its performance?” That is the
objective of the testing summarized in this report, to test samples of the filters listed and compare

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 1 of 12

them to each other in terms of relative performance. Of special interest is the L-eXtreme filter,
and whether its performance is better than the other filter models by a large enough margin to
justify its cost.

Retail Price [USD]**

Filter Name
1.25” 2”
L-Pro $149 $199
CLS (non-CCD) $55 $85
UHC $55 $85
UHC + IR cut $55 + $44 $85 + $65
L-eNhance $179 $229
L-eXtreme $239 $309
** Prices quoted from AgenaAstro.com

Table 1 Summary of Optolong Filters Considered During Test

Method:
Testing consisted of data collection in the following manner:

 Spectral transmissivity data, from near-UV to near-IR, measured using an Ocean

Optics USB4000 spectrometer; and
 Image data, collected using a William Optics FLT98 triplet apochromatic
refractor and a ZWO ASI-294MC Pro OSC camera.

The spectrometer data was collected in my basement workshop with the USB4000 and a broad
spectrum light source. To collect the data I recorded two back-to-back scans from each filter and
calculated the average. In the event that the data varied by more than 0.1% between back-to-
back scans, I rejected the data set and repeated the whole measurement again.

The image data was collected from my backyard in central Ottawa where the naked eye limiting
magnitude (NELM) due to light pollution is +2.9 on average, which translates to Bortle 9+. I
don’t have a filter wheel, so to switch filter configurations I had to remove the camera from the
focuser, and swap the filter manually. Each time I changed filters I would refocus on a
conveniently located bright star using a Bahtinov mask. Images were collected on a single
evening, August 12th, 2020. A single deepsky target was used, the Eastern Veil Nebula
(NGC6992/95), which was located near the zenith for the duration of the image captures.

Results – Spectrum Measurements:

Figures 2 and 3 present plots of the measured spectral responses for each of the filters under test.
The CLS, L-Pro, and UHC filters all have relatively wide pass bands around Hβ/O-III and Hα.
The L-eNhance and L-eXtreme filters on the other hand have significantly narrower pass bands,
with the L-eXtreme having the narrowest pass bands of all the Optolong filters tested. From the
measured spectral response data I extracted the filter characteristics summarized in Table 2.

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 2 of 12

 Figure 2 Measured Spectral Response of Optolong Filters – 1

Hbeta/O-III Pass Band Halpha Pass Band

Mean
Hbeta O-IIIA O-IIIB Halpha N-II S-II Off-
Filter FWHM FWHM
(486.1) (495.9) (500.7) (656.3) (658.4) (672.4) Band
Blocking
L-Pro 81.2nm 96.4% 95.6% 94.9% 49.4nm 92.7% 91.5% 95.8% OD 1.3
CLS (non-CCD) 62.8nm 98.4% 94.5% 94.1% 110.7nm 94.7% 95.6% 99.8% OD 2.2
UHC 47.5nm 95.4% 97.6% 96.8% 77.1nm 99.3% 99.6% 98.4% OD 2.4
UHC + IR Cut 47.1nm 93.2% 95.7% 94.1% 63.3nm 98.3% 98.6% 98.1% OD 2.5
L-eNhance 26.7nm 95.9% 98.1% 97.0% 11.8nm 90.1% 90.7% 1.5% OD 2.5
L-eXtreme 8.1nm 0.3% 60.7% 89.9% 7.8nm 79.5% 42.8% 0.1% OD 2.8

Table 2 Measured Filter Performance Summary

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 3 of 12

 Figure 3 Measured Spectral Response of Optolong Filters - 2

Knowing the measured spectral response of the sample filters also allowed me to predict the
theoretical relative performance of each filter on different kinds of deepsky object, under
different sky conditions. To do this I used the method I developed back in 2012 which uses the
spectral response of the filter and sensor combined with the spectral emission from the deepsky
object and background sky to estimate the apparent luminance observed. If interested you can
read more about the method at the following link:

http://karmalimbo.com/aro/reports/paper_MethodForEvaluatingFilters-part1.pdf

To help visualize the results of this analysis I have plotted the predicted % increase in contrast
for each filter versus the filter’s % Luminous Transmissivity (%LT). %LT is a measure of how
much light gets through the filter in the wave band being observed, which varies depending on
whether the observer is a human or a camera. Figure 4 shows the resulting plot corresponding to
filter performance when using a monochrome CMOS camera under heavily light polluted skies
complete with local LED street lights (i.e. my backyard).

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 4 of 12

 Figure 4 Predicted Filter Performance: Monochrome CMOS, LP w/LED (NELM+2.9)

For emission type nebulae, there is an obvious trend to the predictions: as the pass bands get
narrower (ie. %LT is lower), the contrast increase gets larger. For broadband targets like
galaxies and reflection nebulae there does not seem to be any sort of trend with %LT; if anything
the contrast increase reduces with decreasing %LT. The average expected increase in contrast is
also significantly less for broadband targets than for emission nebulae. The values plotted in
Figure 4 are also provided below in Table 3 for reference. Note that these are theoretical values,
and may not reflect what a user will experience with their setup. The relative performance of
one filter to another should however be representative.

Bright Dim
Nebulae Nebulae Reflection
Filter %LT (O-III rich) (Hα rich) Galaxy Nebulae
L-Pro 36.7 116.9 108.1 -8.4 -0.8
CLS 38.1 255.8 242.5 63.8 22.0
UHC 34.5 368.7 345.2 92.6 27.8
UHC + IR Cut 23.0 417.8 396.8 45.4 24.6
L-eNhance 8.9 973.1 851.8 21.9 33.8
L-eXtreme 3.8 1596.2 1279.8 16.6 6.1
Table 3 Predicted Filter Performance : Monochrome CMOS, LP w/LED (NELM+2.9)
- % Increase In Object Contrast

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 5 of 12

Results – Imaging:
The images collected using the different filters under test are presented below in Figures 5 to 11.
Included is a reference image taken with no filter at all. All seven images were taken using the
same total exposure time of 5 minutes, but the sub-exposure time was varied as required to avoid
clipping of the histogram. All other camera settings (white balance, gain, binning) were left
unchanged between the different filter configurations. The no-filter image was captured using 5
second sub-exposures as that was the longest I could go without saturating the image with light
pollution. For the image captures using the L-Pro, CLS, and UHC filters, I was able to use sub-
exposure times of 10 seconds. These filters removed some of the contribution to the scene
brightness resulting from the light pollution, thus allowing for a longer exposure time without
saturating the image. For the image captures made using the L-eNhance and L-eXtreme filters, I
had to use sub-exposure times of 30 seconds in order to not clip the image at the dark end of the
histogram. These two filters removed so much of the light pollution that the resulting signal was
too low below 30 seconds exposure to not clip data, especially in the case of the red channel.
Using the histograms from my raw captured images, combined with the sub-exposure times, I
pulled out the impact of each filter on relative exposure for each colour channel. The results are
summarized in Table 4.

Sub- Exposure Relative To No Filter

Exposure RGGB
Filter (s) R G B Avg %LT*
None 5 100% 100% 100% 100% 100%
L-Pro 10 33.2% 40.3% 49.4% 40.8% 36.7%
CLS 10 22.7% 25.4% 34.9% 27.1% 38.1%
UHC 10 20.0% 21.2% 27.9% 22.6% 34.5%
UHC + IR Cut 10 13.5% 17.7% 24.0% 18.2% 23.0%
L-eNhance 30 4.4% 8.7% 10.3% 8.0% 8.9%
L-eXtreme 30 3.6% 4.4% 5.3% 4.4% 3.8%
* For generic monochrome CMOS camera
Table 4 Measured Relative Exposure By Colour Channel

One of the challenges of this test was applying white balancing and levels adjustments to all the
collected images in a way that was repeatable, and that did not diminish or over-emphasize the
performance of one filter relative to another. I accomplished this by separating each raw image
into its three colour channels, and doing some initial analysis of each channel’s histogram.
Using the histogram data from the image with the maximum contrast, the L-eXtreme image, I set
black point, mid point, and white point values to apply to all the images so that the end result
was the same amount of histogram stretching as well as a matching white balance. I
accomplished this using the freeware software AstroImageJ, working from 16-bit per channel
FITS files captured directly out of Sharpcap. The resulting output images are what is presented
in Figures 5 to 11.

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 6 of 12

 Figure 5 No Filter – White Balancing + Levels Adjustment Applied

Figure 6 L-Pro – White Balancing + Levels Adjustment Applied

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 7 of 12

 Figure 7 CLS (non-CCD) – White Balancing + Levels Adjustment Applied

Figure 8 UHC – White Balancing + Levels Adjustment Applied

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 8 of 12

 Figure 9 UHC + IR Cut – White Balancing + Levels Adjustment Applied

Figure 10 L-eNhance – White Balancing + Levels Adjustment Applied

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 9 of 12

 Figure 11 L-eXtreme – White Balancing + Levels Adjustment Applied

The adjusted images confirm visually the filter performance predictions presented in Figure 4.
Using the L-Pro, CLS, or UHC filter all resulted in a noticeable improvement in the contrast of
the nebula. As predicted, the UHC provides a slightly better contrast than the CLS, which in turn
provides a slightly better contrast than the L-Pro filter. One interesting observation was that the
L-Pro filter image had a non-uniform colour cast that is not observed with the other filters. This
may be related to the non-uniform appearance of my L-Pro filter sample’s coatings that I
observed originally back in 2015 when I first tested these filters. Adding the IR Cut filter to the
UHC does not increase the contrast of the nebula significantly, however it does greatly reduce
the brightness of stars in the image, making the nebula easier to see. Using the L-eNhance filter
resulted in a large improvement in the nebula’s contrast compared with the first four filter
configurations. One peculiarity of the L-eNhance image however is that the image still has a
turquoise hew after being white balanced in the same way as the other filters’ images. I don’t
know if this is a specific issue when using the filter with the ASI294 camera, but I have heard
reports of similar issues with white balancing this filter on other cameras. Finally, using the L-
eXtreme filter resulted in another significant increase in image contrast over the L-eNhance
filter. The visibility of stars was also greatly reduced, making the nebula easier to see. I also
noted that the white balance was much better than with the L-eNhance filter. I have since used
the L-eXtreme filter for some live EAA observing, and can confirm that this new filter is much
easier to get a nice white balance than the L-eNhance filter.

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 10 of 12

Results - Angle Sensitivity:
The final thing I confirmed from my testing was how sensitive each filter’s performance is to the
f-ratio of one’s optics. I did not bother measuring the angle sensitivity for the L-Pro, CLS, and
UHC filters because their pass bands are so wide – too wide really to be affected in any
significant way by f-ratio. I did however measure the spectrum of the L-eNhance and L-
eXtreme filters for a range of f-ratios. The results are summarized in Figure 12.

/6 /5 /4 /3 /2

100

80
% Transmission

60

40
Legend
L-eNhance - O-III
20 L-eNhance - Halpha
L-eXtreme - O-III
L-eXtreme - Halpha

0
0 2 4 6 8 10 12 14
Light Angle from Perpendicular (deg)
Figure 12 Impact of Angle On Filter Performance

As expected, the narrower pass bands of the L-eXtreme filter make it more susceptible to losses
in performance with progressively faster optics. The images I collected were captured at f/6.3,
so some of the effects of f-ratio should have been apparent. Since the object I was observing did
not fill the field of view (FOV), the images presented above are cropped and so don’t really
show any noticeable gradient in performance from center to edge using the L-eXtreme filter. On
an object that fills the FOV the impact would be more visible. Even with the supposed reduction
in performance due to f-ratio, the L-eXtreme filter still produced an image of superior contrast
compared to all the other filters tested.

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 11 of 12

Conclusions:
The data collected during this test confirms yet again that a filter with narrow pass bands
produces an image with superior contrast on emission-type nebulae. Some other more specific
conclusions drawn from this test are:

 When compared to other available filters with the same %LT, the Optolong L-Pro
filter provides inferior performance under light polluted skies. This observation is
true for all filters of this type (i.e. Multi-band), regardless of brand.
 In terms of cost per performance the Optolong UHC filter provides a very good
value. The filter performs very similar to the Astronomik UHC but at a fraction
of the cost.
 There is a significant improvement in contrast on emission-type nebulae resulting
from the use of the Optolong L-eNhance filter. There is however an issue with
white balancing that was not encountered with the other filter models.
 Of the six filter configurations tested, the Optolong L-eXtreme filter provided the
largest increase in contrast. The increase in contrast relative to the L-eNhance
filter was observed to be significant. The filter was also easier to use than the L-
eNhance due to it being easier to white balance.

If you have any questions, please feel free to contact me.

Cheers!

Jim Thompson
[email protected]

© Abbey Road Observatory, aka Jim Thompson, Aug 2020 Page 12 of 12