Shoot-out: TV-60is vs Canon 400mm/5.6L
By Samir Kharusi and Hilmi Kindy
It so happened that between the two of us we owned these OTAs, not very different in aperture and focal length, each well reputed in its own niche. The TV-60is is a small, premium imaging scope that has received glowing reviews in astronomy circles, for its sharp stellar images and good contrast. The Canon 400mm/5.6L is considered amongst birders as a great bird-in-flight lens, light enough for hand holding and has fast autofocus. The question we had was, in these days of DSLR astrophotography, how do they compare as astro-imaging OTAs? Will the super-sharp TeleVue soundly trounce the camera lens in delivering tiny, non-distended stars? Is the fast autofocus of the camera lens even good enough for astroimaging? Of course, reality invariably turns out more complex than initial expectations. Read on.
The TV-60is costs $1695 inclusive of a dedicated field flattener, but without an optional digital read-out for focusing. That's still quite a high price for a 60mm aperture, even for a premium imaging refractor, but you do get a beautiful example of sensuous craftsmanship, 360mm focal length, f6. The Canon 400mm/5.6L costs $1100 (71mm aperture) and has the usual, good L-build quality cherished by Canon users. But placed next to the TeleVue, the camera lens feels like a cheapo. Even the weights of these OTAs seem to corroborate; the TeleVue weighs twice as much, 2.76kg, as the lens, 1.37kg. Anyone thinking of using the TV-60is as a super finder scope should take into consideration the weight and ponder carefully the implications as regards rebalancing his main OTA. On the other hand, if you are a lover of wide visual fields, stick in a 2" diagonal together with, say, a 41mm Panoptic (widest field in a 2" barrel) and you have a FoV so wide that many binoculars have a hard time matching. Of course, you can also use it as a birding scope on an appropriately heavy tripod. On the other hand the Canon lens is light enough to use a camera with hand-held, though the focal length is such that you do require very high shutter speeds. It does not have Canon's IS (Image Stabilization).
We started off by putting the OTAs onto a normal photo tripod and photographing a far off building, basically to verify that the TV-60is is nicely orthogonal to the sensors of a Canon 40D (a 1.6x crop or APS-sized DSLR) and a 1Ds (full 35mm format). Upon reviewing the frames on a PC monitor it quickly became apparent that even at the modest focal lengths of the OTAs, there was too much heat shimmer to compare meaningfully these two high performance OTAs. So we had to await night. The OTAs were tandem-mounted onto an AP1200.
Focusing: the 40D Live-View is simply wow! for focusing the TeleVue. We did try to better it by running ImagesPlus in focus mode, but it became immediately obvious that variations in seeing, between frame downloads, was too high to actually better the Live-View. Consequently our conclusion was that using the focus mode in ImagesPlus, with the seeing we had, was simply an exercise in avoidable frustration. The 40D Live-View is that good. Newer versions of ImagesPlus are going to incorporate Live-View. Similarly one-click autofocus with the lens also matched the best we could achieve with the focus mode in ImagesPlus. With the TV-60is we simply used Live-View to focus on Capella and applied its focus locks. We then slewed to the Double Cluster for the exposures. With the camera lens we autofocused on Capella and continued onto the Double Cluster. For both OTAs focus was retained precisely, even when we switched bodies from the 40D to the 1Ds. Kudos to Canon (and Hutech?) for excellent focus calibration between the Hutech modded 40D and the unmodded 1Ds. Perhaps sheer luck, but that was that. Just for extra reassurance, when we were using the 1Ds, the camera lens was autofocused on Capella in the 1Ds itself.
What we are presenting below are the best frames out of multiple attempts. First the TV-60is, ISO 800, 30sec exposure at f6.0, then the camera lens, 25sec at f5.6, both in the non-modded 1Ds. Normal nonlinear conversion was used to generate the Jpegs from Raw (i.e. the usual gamma stretch is implied, but no additional contrast stretching was required to bring out the differences). The inset crops are all at 100%, unsharpened, neither in-camera nor in post processing.
The full frames may be accessed by clicking on the two images above (NB 1.2+ megabytes each!). We notice straightaway that the TV-60is is remarkably sharp centrally, in fact it seems to be sharper than the 8.8micron pixel pitch of the 1Ds can resolve. Note the block-like appearance of some of the stars. We also notice that its coverage does not extend gracefully to the peripheries of the 35mm-format frame, despite having a dedicated field flattener. The camera lens is more even in its delivery, but the stars are less tight dead center. The TeleVue is connected via a T-adapter, 36mm ID, and this falls short of the 43mm diagonal of the 35mm format. Note the vignetted corners. Actually there is a lot less vignetting than what one might expect from simple geometry. The field flattener at the rear likely helps in spreading the illumination. On the APS-sized frame one notes a slight asymmetry and it would have been nice for perfectionists if the TV-60is had some means of adjusting orthogonality, like its bigger brothers the 127is and the 101is. On the large pixels of the 1Ds we could not see great improvement by closing down the Canon lens slightly, from f5.6 to f6.3. But you may judge for yourself from the 100% central crops below if the improvement is worth chasing, taking the consequent light loss and the probable introduction of diffraction spikes on very bright stars into account. There are no bright stars in the target field used. Also please note that no sharpening has been employed for any of the 100% crops of the star fields. The same degree of improvement also takes place over the rest of the frame. You can access the full frame at f6.3 by simply clicking on the image below:
Our conclusion is that the lens may be used wide open for astrophotography. Even closing down the lens further, to f8, does not make a major difference, on the basis of an earlier daytime test.
Evenness of Illumination
Keeping in mind the small ID of its T-Ring, the TV-60is has remarkably even illumination of its image field, except for the extreme corners of the 35mm format. Shot off many daytime sky flats, converted them in ImagesPlus (Linear conversion), stacked them and took the diagonal line profiles:
Note that the TeleVue is extremely even along both diagonals, right into the corners where the T-Ring imposes a sharp drop. Rather unexpectedly, the Canon lens is hot-spotting in the center! Not noticeable in daytime usage, but certainly flat-fielding is required for astro use. Conclusions so far? The TV-60is out-resolves the 8.8micron pixel pitch of the 1Ds on-axis, and it has very even illumination. The use of a T-Ring for coupling perhaps implies that it was never meant to cover a full 35mm format. It does a very good job of covering an APS-sized sensor, however, both in star shapes and in illumination, but "just". From our results we would expect that with enough pixel peeping one is bound to find that centration will likely be slightly imperfect, and one edge or corner is bound to show slightly distended stars. So, do not pixel peep and be happy! Some slight cropping and one gets an almost perfect star-field on APS-sized sensors. The Canon lens does not have such a stellar performance on-axis but it does perform very evenly right across a full 35mm format. The one-click autofocus on a bright star is totally reliable and the lens may be used wide open for astrophotography, but do include flats in your calibration.
For the Die-Hard Pixel Peeper
The newer cameras come with smaller and ever smaller pixels. Lenses and OTAs already had a hard time meeting the demands of 8.8micron pixels employed in the foregoing tests, but along comes the current generation of DSLRs with a 5.7micron pixel pitch, like the 40D. It's like we are suddenly using a 1.5x tele-extender full time! So if you are a die-hard pixel peeper, enjoy below scrutinizing these two OTAs with a Canon 40D. First, daytime tests early morning before the afore-mentioned heat shimmer got in the way. The building was focused upon when it was framed dead-center, and then the camera was simply slewed to place it on the 40D frame edges, no refocus, no change in exposure. Moderate sharpening was employed, as one would do for normal daytime photography, e.g. birding. Focusing was by Live-View for the TeleVue, center-spot autofocus for the Canon. The upper thumbnails demonstrate how the placement was accomplished:
And below are 100% crops of the same star field as previously, but shot with the Hutech modded 40D, UV/IR Blocker over the sensor, unsharpened:
And the conclusion is? With enough scrutiny one can always find deficiencies. The 5.7micron pixel pitch of the 40D is not out-resolved by either OTA and some colour fringes are starting to appear. Hard sharpening makes them into coloured rings. Whether these are camera artifacts or lens deficiencies, it's hard to differentiate. Even the tracking of these 30sec and 25sec exposures becomes an issue. OK, enough pixel peeping for now J
Here is an example image using the Canon 400mm/5.6L on a Hutech Canon 40D, autofocused on a bright star, 36x3min sub-exposures at ISO 1600 and f5.6, unguided on a little Kenko Skymemo mount:
Of course, in giving an example like the above image, a lot depends on what the post-processing-to-taste has done to the stack. A version of the above, with more aggressive processing, sharpening, saturation, etc., may be viewed here..
For more stuff return to Samir's Home
To send an e-mail just click on: