I compare a Canon EOS R filter-modified by AstroGear against Canon’s own factory-modified Ra to see which camera performs best for astrophotography.
Today’s DSLR and mirrorless cameras work great for astrophotography. Their main shortfall is their lack of sensitivity to the deep-red wavelength emitted by hydrogen atoms, the primary source of light from nebulas along the Milky Way.
All stock cameras have a filter in front of their sensors that blocks infrared (IR) light that digital sensors would otherwise readily pick up, flooding the image with unwanted light. In addition, most lenses don’t focus IR properly.
The problem is in blocking IR most sensor filters also cut into the red end of the visible spectrum, attenuating the wanted Hydrogen-alpha (H𝛂) wavelength at 656 nanometers. While red nebulas do get recorded, they can appear pale and colorless, lacking the full extent and detail we’d like to pick up.
The solution is to modify the camera, which entails disassembling it to remove the manufacturer’s IR-cutoff filter and replace it with one of two choices:
— A clear glass filter of similar thickness. Called a “Full Spectrum conversion,” this mod is for those who might want to shoot in the infrared for special-effect landscapes or beating light pollution under urban skies.
— Replace the stock filter with another IR-cutoff filter with a steeper cutoff that still blocks IR but transmits much more H𝛂. This is usually referred to as an “Astro conversion.” It’s the modification I tested here.
(For more on modification options see AstroGear’s webpage here, and Nico Carver’s excellent comparison review here on his Nebula Photos YouTube channel.)
The AstroGear Camera
Several companies will either modify a camera you send them, or sell you a camera body already modified with your choice of a Full Spectrum or Astro mod. One company, AstroGear.net based in California, offered to send me one of their modified cameras for testing. (There is absolutely no relation between AstroGear.net and AstroGearToday.com; the similar names are just a coincidence!)
I chose to have Daniel Amado, owner of AstroGear, send me a used Canon EOS R, Canon’s first mirrorless camera that he modified with his Astro conversion. The modification retained both the dust removal mechanism and, as explained below, the anti-alias filter.
The test camera proved to be in “like new” condition with a low shutter count, making the camera much cheaper than a brand new unit, a great option when buying a camera for astrophotography.
While AstroGear offers many models of cameras, I chose to test the modified Canon R as it allowed me to compare it directly to the Canon Ra, a specialized astro “a” version of the 30-megapixel R that Canon sold for two years between 2019 and 2021. For the Ra, Canon factory-installed their own IR-cutoff filter with extended H𝛂 sensitivity. I have taken thousands of images with my Ra since 2019 and love it. I was curious to see if the AstroGear-modded model would match or outperform the Ra. In some key ways it did the latter.
Like all independently modified cameras, the AstroGear R produces an overly red or pink image when shooting with Auto White Balance or normal daylight color temperature of 5200 to 5400 Kelvin. This can be corrected in one of three ways:
- As I show below, shoot an image of a white or grey card and use that image as the basis for a Custom White Balance that applies a color correction that is stored in-camera.
- Shoot through a cyan-tinted filter on the lens to re-balance the color temperature back to a more normal tone.
- Simply apply the needed color correction in raw processing, using the White Balance eyedropper tool (present in most raw processing programs) to sample a neutral area of the image. This is what I did for the test images.
The advantage of using a factory-modified camera like the Ra, however, is that Canon has built a custom white balance into the camera’s firmware. Images look quite normal out of camera even using Auto White Balance, making the camera easy to use for all types of photography in all forms of lighting. To preserve that factory-set color balance when updating the camera firmware, only the version made for the Ra (not R) should be used.
While the Ra’s images can be used with minimal color correction, any third-party modified camera like the AstroGear R that I tested will always require major color corrections, especially for nightscapes and normal daytime images.
Even deep-sky images require neutralizing the overly reddish tint. Nevertheless, I found the AstroGear camera’s images could be corrected back to neutral quite easily when processing raw files.
I took same-night test shots with both cameras, and also with an unmodified 45-megapixel Canon R5 (reviewed here at AGT) as a control, shooting both widefield Milky Way images and close-ups of Orion’s nebula-rich Belt and Sword. A pair of the widefield images is shown below under Halos.
For the telescopic close-ups, I shot with a William Optics 51mm RedCat astrograph, using the three cameras to shoot unfiltered images and images through an IDAS NBZ dual narrowband filter I reviewed here at AGT.
As one would expect, in the unfiltered images (shown above) the two modified cameras picked up more nebulosity than the stock Canon R5, though the R5 performed admirably, recording a surprising amount of red nebulosity.
However, the AstroGear R and Canon Ra showed more of the faint nebulosity around the Horsehead Nebula and more of the pink regions in the Orion Nebula. The example images here are single frames only, lightly processed for neutral color balance in the background sky, equal brightness, and identically boosted contrast.
I’m hard pressed to see a difference between the AstroGear R and the Canon Ra. Both show similar levels of nebulosity. Where the difference lies is in their image brightness. The Ra produced darker images, requiring an exposure boost of 0.65 stops in this example to match the AstroGear R (as indicated by the identical histograms).
Shooting through the IDAS NBZ filter revealed more nebulosity, as expected when using a filter that isolates just the green Oxygen III and red H𝛂 emission lines of nebulas. The stock R5 again performed surprisingly well but the modified cameras did better, again with no marked difference between them.
However, in this case the exposure level of the Ra’s image had to be boosted by one full stop to match the image from the AstroGear R. The IR-cutoff filter Canon installed in the Ra doesn’t have the high light transmission of the Optolong UV/IR-cutoff filter AstroGear installed in my test camera.
That was a surprise. I was expecting to see differences in red sensitivity (which I didn’t) but not in overall sensitivity. The high efficiency of the AstroGear-installed filter means exposures can be shorter, which is especially helpful when shooting through narrowband filters.
An area where I expected to see a difference – and did – was in the presence of halos around bright objects. The Ra is known for showing prominent halos though, as shown in my tests, mainly around reddish objects like Betelgeuse, Aldebaran, and Mars (shown here). But I have seen them around Jupiter and the Moon, and only faintly around Sirius.
These single widefield images with a Canon RF28-70mm lens at 40mm show the difference. The Ra exhibits a sharply defined bright halo on Mars plus a much fainter, larger halo. Residual IR light may be reflecting off the Canon IR-cutoff filter’s surfaces, a suspicion borne out because adding an additional UV/IR-cutoff filter (I use an Astronomik filter) into the light path reduces the Ra’s halos.
By comparison, the Optolong UV/IR-cutoff filter AstroGear installs does a good job on its own suppressing halos, though it doesn’t entirely eliminate them. A large, faint halo remained, revealed when boosting contrast, as I did here. (In these shots, the diffraction spikes are from stopping the lens down to f/2.8 and not a product of the camera or filter.)
My test shots of Orion with the RedCat included the dreaded star Alnitak, the bright blue Belt star near the Horsehead Nebula that is notorious for inducing filter halos. As I show above in extreme blowups of the RedCat images of Orion, I saw no halos in my exposures with the stock R5, the AstroGear R or Canon Ra. Claims that the Ra is unusable because of rampant halos are overblown. But the AstroGear mod does perform better, though its Optolong filter isn’t completely immune to halo artifacts.
The closeup views above and below reveal essentially no difference between the two modified cameras for star sharpness and image quality. My test camera retained the Canon anti-alias filter (often called a low-pass filter), another filter that is placed in front of the sensors of most DSLR and mirrorless cameras. This AA filter smooths jagged pixelation and suppresses moiré banding.
Removing the AA filter is also an option for those seeking the highest resolution, but I would caution against it. Cameras I have owned and tested that lack an anti-alias filter have showed oddly colored green stars due, I think, to improper de-Bayering of the color information from the sensor when stars occupy only one or a few pixels.
Exceptions Daniel told me about are the Canon R5 and R6 where removal of their anti-alias filters is required in order to have camera lenses focus normally to infinity with his modifications. The reason is that the R5 and R6’s IR-cutoff filters are only 0.5mm thick, half that of the R’s filter, while the replacement Optolong filter is 1mm thick. If the anti-alias filter were left in place, lenses would not focus properly. Any filter between the optics and sensor shifts focus position, and the thicker the filter, the greater the focus shift. So, the total thickness of filters going in has to match what comes out.
As it was, the AstroGear R camera I tested focused normally, both manually and under auto-focus, with my Canon RF lenses. Star image quality and their natural colors matched what I saw in the Canon Ra and R5’s images. (The green stars on the filtered images above with the NBZ filter are an inevitable byproduct of shooting through a narrowband filter that transmits only green and red wavelengths.)
Of course, stars will look good only if sharply focused. The Canon Ra offers a unique 30x zoom view in its Live View focus mode, making it easy to nail focus. Modifying a stock Canon R still leaves you with its maximum 10x zoom when focusing, a downgrade when you’ve been used to the precision of 30x.
Fearing I might brick the camera, I’ve not tried applying a Canon Ra firmware update to the modded Canon R to see if that adds the 30x zoom and re-balances the color temperature. And I’ve not found anyone who has risked performing such a cross-camera firmware transplant.
I’ve long preferred using modified DSLRs, and now exclusively mirrorless cameras, as they are versatile, being useful for nightscapes, wide-field Milky Way images, and telescopic deep-sky shots. Even movies. They are easier to power and operate in the field than dedicated cooled CMOS astro-cameras. And a modified full-frame mirrorless camera can be purchased for much less than a full-frame cooled CMOS camera. While cooled CMOS cameras with smaller sensors do cost less, I prefer the wide field a full-frame camera provides.
However, as with Canon’s previous “a” models – the cropped-frame 20Da (from 2005) and 60Da (2012), and Nikon’s full-frame D810a (2015) – the Ra (2019) was offered for only two years before being discontinued. Units appear only rarely on the used market.
So, if you are considering acquiring a modified camera, having one you already own modified, or buying a modified camera body from one of the small companies that perform the service is likely the way to go. Based on the performance of the modified camera I tested, I can recommend AstroGear.net as a fine choice for either option. I was pleased enough with the test camera sent on loan that I bought it. That’s my endorsement!
Excellent deep red sensitivity
Higher light transmission than Canon-modified EOS Ra
Greatly reduced halos compared to the Ra
AstroGear offers modified used cameras for reduced prices
Requires major color balancing in processing
Slight residual halo from Optolong UV/IR-cutoff filter
Lacks 30x focus zoom of Canon Ra
Modification of a user-supplied full-frame camera is typically $349 depending on the camera model and filter choices.
Complete modified Canon R: $1,799 (camera is used)
Editor’s note: AstroGear is not associated with AstroGear Today.