"...complete sharp picture of night sky or landscape keeping F 1.4 or 1.8."
Why f1.4? It may not be necessary or even desired. Find the brightest star and focus on it. Use MF of course and recompose as needed.
Focusing on the night sky is a bit tricky because stars are not bright enough to use auto-focus.
The focus indicator on the lens wont be accurate enough. You'll need to carefully adjust focus and I highly recommend that even when you're convinced you have achieved accurate focus, test this by taking a short shot, downloading the image and inspect the results. Loads of astrophotographers have looked at the image on the rear-lcd ... belived they nailed focus, got home to the computer, and were disappointed. Take your time here ... it pays off.
One nice thing about astrophotography is if *anything* in space is in accurate focus... then *everything* is in accurate focus. You can take advantage of this if the area you want to captures lacks bright enough stars for focus... just re-point the camera to an area that *does* have bright stars, switch to manual focus and focus the lens, then move the camera back to the area you want to catpure ... being careful *not* to touch the focus ring (some people use rubber bands to hold the focus ring once they nail focus.)
There are two ways to focus...
Method #1 -- Using "live-view"
Switch to "live view" mode. Switch to manual focus mode. Manually turn the focus ring to approximately the "infinity" point (just as a starting point).
Canon supports "exposure simulation in live-view" ... meaning if you use live-view and you set the shutter speed to 1/4000... everything will look dark (probably black). But if you adjust shutter speed to 30 seconds, it'll brighten up (even though you aren't actively taking a shot... it does this to "simulate" what you might get if you did take the shot). SO... to take advantage, go crazy... set the ISO to your camera's max, set f-stop to wide-open, set shutter exposure time to maximum (30 seconds), and now the live-view screen should have lots of stars (if the camera focus is nowhere near close then the stars will blur to the point that you see nothing. This is why I suggest manually rotating the focus ring to the "infinity" mark ... because at least it'll be close enough that you'll see something.
Center a bright star (and it's better to use a star ... not a planet).
Adjust focus as well as possible.
Now use the live-view zoom controls to zoom to the 10x level.
Re-adjust focus ... trying to make the star as close to a pin-point as you possibly can.
Adjust the shutter speed to 5 seconds and take a sample shot. If you do not have a remote shutter release then use the camera's self-timer to take the shot. This is because the act of "touching" the camera to press the shutter button will trigger a vibration in the tripod and you need to wait a few seconds to let that settle down before you take the shot (or you'll have elongated stars caused by the vibration ... not the focus.)
Having this... import the shot and inspect. I know it's a hassle to do this in the field... but this hassle is rewarded (I know this from experienc. That tiny 3" lcd isn't nearly large enough to really know if the image is sharp.
Once you are convinced you have good focus (repeat as many test shots as needed until you nail focus) re-point the camera to the area of sky you really want to shoot (being careful not to touch the focus ring) ... also return your ISO and shutter speeds to sane settings (cranking everything to max was just to make it easier to focus) and start capturing images.
Method #2 - Using a Bahtinov Focusing Mask
This is a better method... but it requires you buy or build extra equipment. A "Bahtinov" focusing mask is a card that has slots cut in it at specific angles. These angles cause point of light to create diffraction spikes which are orthogonal to the direction of the slots.
You can do a Google search for "Bahtinov Mask" and see loads of examples.
Most of these are made for telescopes... but there are models made for camera lenses.
When you attach to the lens and point at a very bright star ... if you're "close" to accurate focus you'll see diffraction spikes. If instead you see a blurry image that resembles the mask, then you're far away from focus... adjust focus until you start to see a point with spikes of light (diffraction spikes).
The idea of the mask is very simple...
You'll see three spikes. It will look like an "X" shaped pair of spikes with a third straight spike "|" that crosses through the "X". If the three spikes converge at a common center point... you've nailed focus (bang-on accurate). If they don't converge (the "|" is off to one side of the "X" then you are out of focus and need to slowly adjust until all three spikes are converged.
But some tips on the mask. Since the mask is usually solid black with slots cut in it... it is cutting out more than 50% of the light and quite frankly the stars are already pretty dim. The spikes may be very hard to see (if you increase the exposure simulation in live-view then it's easier to see... or take a test exposure and you'll really see big spikes).
So there are a couple of companies that make a focus mask out of a clear resin material and they cut grooves in it. This creates spikes withouth cutting out all that light... and it makes it easier to see.
Lonely Speck (lonelyspeck.com) makes a mask called the "SharpStar 2". You'll need a filter holder (the same filter holder that would be used to hold Cokin filters or Lee Filters, etc.) -- because it's not a round threaded filter.
Farpoint Astro (farpointastro.com) makes focus mask inserts. This is the actual mask... without the threaded filter. You use an existing clear filter (such as a UV filter) and attach the mask to your filter, then thread the filter onto the camera lens. (they make them in different diameters).
I have many Bahtinov focus masks ... sized to work with my various telescopes and lenses. It is my favorite way to be sure that I've nailed focus.
One last tip... just don't forget to remove the mask after you focus. I've forgotten this step... started an imaging run... and as I start to inspect my data (as the images come in from the camera) I realize the mask is still on the lens. (fortunately I noticed before I wasted too much time.)
Keep in mind that no lens is perfect and you may get some "coma" or bloat issues out of bright stars near the edges of the frame when shooting at very low focal ratios ... but things quickly sharpen up if you just stop down slightly.
The really amazing shots are done using a "tracking" head or tracking mount. These take away the time limits on how long you can expose before you see elongated stars because the head is rotating the camera to match the movement of the stars. The Earth spins west to east (making an illusion that we are still while the sky moves east to west) on it's polar axis. If the tracking head is tilted so that rotation axis is parallel to Earth's polar axis (there's an alignment aid that comes with the device) then the tracker will rotate east to west as the Earth spins west to east ... and at EXACTLY the same rate (cancelling the Earth's rotation so the camera's view remains perfectly still).
This also allows you to stop down:
The image above is a single 8-minute long shot using a Canon EF 135mm f/2 lens ... stopped down to f/10. But it's on a tacking head and you can see I have tack-sharp stars with no elongation. f/10 is also causing diffraction spikes to appear around the brightest stars.
That was actually a test shot I did to verify that the tracker was accurately "polar aligned".
Here's the finished shot ... but this was done at f/2 using the same lens. It's an HDR built by capturing a series of shots that were 2 minutes, 1 minute, 15 seconds, and 3 seconds (about 8x of each). These are then stacked and combined ONLY with images of identical duration to create sub-masters (for each duration). The sub-masters (four of them) were then combined via HDR to produce the result:
You can see that here: https://www.flickr.com/photos/thevirtualtim/24856906857/in/dateposted-public/lightbox/
Three stars of Orion's belt (Orion is rotated on it's side and this is just the lower region -- the stars are blue and somewhat bloated) are not pin-point... this is not a focus issue. This is "coma". Notice each star has a couple of little wings or filaments that come off the sides ... but they're all orthogonal to the center of the frame. That's caused by lens optics because these images were captured at "wide open" (f/2 for this lens).
Had I stopped down even slightly ... say to f/4 ... that would bave been greatly reduced for a better result. But f/4 would require 4x more exposure time ... so I'd have needed to shoot 8 minute, 4 minute, 2 minute, and 15 second exposures (and since I captured 8 of each, this would have required about 2 hours to capture all the data (I captured all my data in about 30 minutes). Longer is better ... if you have the time.
Also, I only captured 8 of each and I normally recommend getting at least 10 of each. This is because stacking software can use "sigma clipping" as a stacking method instead of "averaging" as the stacking method. (sigma clipping is MUCH better at noise reduction.)
To capture the landscape at night (I don't normally do landscapes) you would switch off the tracking mode (if using a tracking head). Then nudge the camera down slightly (this raises the horizon line in the camera frame), re-focus to the landscape as needed (don't worry about the stars) and capture the shot.
Assuming you've stacked the "stars", you use Photoshop (or whatever tool you prefer) to merge the sharp sky with the horizon.
One last tip... if your lens is a little softer in the corners than in the center, don't position a star *exactly* in the center of the frame to focus. Place the star about 1/3 to 1/2 way toward the corner of the frame. This improves the "average" focus for stars across the frame. (the corners become much sharper at the expense of the center become fractionally softer ... but the overall look is better because everything is much closer to focus.) Stopping down slightly (which means you must take longer exposures) will certainly help.
12/05/2023: New firmware updates are available.
09/26/2023: New firmware updates are available.
08/18/2023: Canon EOS R5 C training series is released.
07/31/2023: New firmware updates are available.
05/18/2023: New firmware updates are available.
03/30/2023: New firmware updates are available.