Up next in 10
Sony a7R3 vs Canon 5D Mark IV | Part 2: Dynamic Range | 4K
39K views
Jul 25, 2023
Sony a7R3 vs Canon 5D Mark IV | Part 2: Dynamic Range | Sony's new a7R3 (a7r III) is reputed to have great dynamic range. Can the Canon 5D Mark IV keep up? Find out in Part 2 of the a7R3 vs 5DIV showdown. Visit the Image Gallery: http://bit.ly/Sonya7R3ig Purchase the a7R III: B&H Photo: https://bhpho.to/2D6ibNO | Amazon: http://amzn.to/2CNxOvH | Amazon.ca: http://amzn.to/2CNYz36 | Amazon UK: http://amzn.to/2CRGvWM Purchase the 5D Mark IV: B&H Photo: http://bhpho.to/2cLng3e | Amazon: http://amzn.to/2mAJ7MX Strap of Choice: Peak Design Slide Lite: B&H Photo: https://bhpho.to/2BsubH2 | Amazon: http://amzn.to/2GcxINh | Amazon Canada http://amzn.to/2Fa5Yr7 | Amazon UK: http://amzn.to/2n6ALxC Want to support this channel? Use these affiliate links to shop at: B&H Photo: http://bhpho.to/1TA0Xge Amazon: https://www.amazon.com/shop/dustinabbott Ebay: http://bit.ly/DustineBay Make a donation via Paypal: paypal.me/dustinTWI Get a discount off all Skylum Editing Software (Luminar, Aurora HDR, AirMagic) by using code DUSTINHDR at checkout: http://bit.ly/LuminarDLA Become a Patron: https://www.patreon.com/dustinabbott | Check me out on: Personal Website: http://dustinabbott.net/ | Sign up for my Newsletter: http://bit.ly/1RHvUNp | Instagram: http://bit.ly/DLAinsta | Google+: http://bit.ly/24PjMzv | Facebook: http://on.fb.me/1nuUUeH | Twitter: http://bit.ly/1RyYxIH | Flickr: http://bit.ly/1UcnC0B | 500px: http://bit.ly/1Sy2Ngu My filming setup: Sony a7R III: B&H Photo: https://bhpho.to/2D6ibNO or Amazon: http://amzn.to/2CNxOvH | or | Sony a9 @ B&H Photo: https://bhpho.to/2HyWIyt or Amazon: https://amzn.to/2s1vYE0 Tamron 28-75mm F2.8 RXD @B&H Photo: https://bhpho.to/2FA00la or Amazon https://amzn.to/2G2kaEr
View Video Transcript
0:00
Hi, I'm Dustin Abbott, and I'm here today to give you part two of this comparison series looking at camera and specifically sensor performance between the tried and true Canon 5D Mark 4 and the new Sony A7R3
0:22
And so in today's episode, we're going to be taking a look at dynamic range. And of course, this was an area
0:29
where, for example, when the A7R Mark 2 was introduced and still being compared to Canon cameras of a previous generation
0:38
where it kind of blew Canon out of the water. The 5D Mark 4 was a pretty huge step in the right direction for Canon
0:46
although we found that when we did the comparison series to the D850
0:51
that while it is close and roughly similar at, you know, less extreme under-exposing or over-exposed areas that in the absolute kind of four-stop and beyond
1:07
highlight or shadow pushes, the DA50 is still superior. We'll find out how the A7R3 does in this metric as well
1:17
So in this episode today, we're going to take a look both at what happens when you under-expose
1:22
images by one, two, three, and four stops. And then also what happens when you over-expoise images. And then also what happens when you over-expoise
1:27
expose them by the same number of stops. And we'll see how successfully you can recover shadow
1:32
information and then how successfully you can recover highlight information. And that should give you a
1:37
good idea of how this is going to play out in real world shooting situations when you're trying to
1:43
either, you know, drop shadows and areas maybe in your foreground if you're a landscape shooter or
1:50
recovering sky information if you are trying to recover on the other end. And so let's jump in and
1:57
let's see how these two cameras compare. Okay, so first let's take a look at this is base ISO, ISO 100, and this is our base exposure
2:07
130th of the second, F4, this is using the Zeiss-Milvis 135 millimeter F2 lens
2:14
And so what we're going to do now is that we're going to see how the camera handles under-exposing the image here
2:22
And so here we'll take a quick look at our shadows at base ISO. So, of course, everything is very, very smooth
2:28
Looks fantastic. So now what we have done is we have under-exposed one-stop on the left side
2:35
and so you can see, you know, looks a little bit dim. And on the right side, all I've done is I've dialed back in one stop of exposure
2:43
And so this is, you know, kind of a shadow recovery with one-stop
2:48
And so let's say you under-exposed by one-stop, and so you didn't quite nail your shot
2:53
what kind of penalty would you be paying for that? And so as we look here, we see that the answer is that you would be paying very little
3:04
You know, there's a little, the shadows are not quite as inky black here
3:08
but you can see that we've managed to raise up shadow information
3:13
and we haven't really added noise anywhere at all. Everything still looks really nice and clean
3:20
Looking over here, everything looks good. you've restored lost information and you've done it really without much of any penalty
3:27
If we look at the colors here, everything is still looking pretty clean and so no real issues there
3:35
So what happens now if we under-exposed by two stops instead and then we bring those two stops back
3:41
And so just take a look at this globally. Then we'll compare the one-stop under-exposure and the two-stop under-exposure recovery
3:50
And so you've missed by two stops. What's that going to? to cost you in terms of shadow recovery. Well, if we look comparing the left to the right
3:58
we can see that there is the finest bit more of shadow noise that is there. It's mostly just
4:05
in the form of what you might call grain, pixel noise, not any kind of discoloration. And for example
4:11
if we look here inside, the actual, you know, inside the camera itself, we're not really seeing
4:18
any really additional noise. Let's take a look over here on this side
4:22
and you know everything still looks nice and dark and clean. No real extra noise information introduced
4:31
Looking over here at color information, you might see the finest bit more of, you know
4:38
some color noise there, but it is still so mild as to be, you know
4:43
only detectable if you're putting the two images side by side. All right, so what happens if you've missed an under-exposed by three stops
4:50
As you can see, our image is starting to get really, really dim, but here we've dialed back in three stops of exposure and looking at the image globally, it still holds up remarkably well
5:01
So if we compare our two-stop recovery on the left with our three-stop recovery on the right, what are we going to find here
5:08
We're looking at this area here, we can see that yeah, there is a little bit more pixel noise that has been introduced
5:16
You can also see a bit more in this color here but you know really all things considered it still looking pretty clean there and are actually kind of inky level We take a peek inside the camera here
5:28
And, you know, if anything, it doesn't really look any different to me
5:33
Very similar on that side. And so looking over here as well, our, you know
5:38
our three-stop recovery hasn't really cost us anymore than what our two-stop recovery has
5:44
And in some ways, I would say that there is less difference between the two and the three-stop result than there was necessarily between the one and the two-stop recovery
5:53
Okay, so let's say you really blew it. You've undis-exposed here by four stops
5:58
There's not a whole lot of information left here. Now on the right, we've dialed that exposure back in, and once again, look at the image globally
6:06
There's really not a huge penalty here. I'm not really seeing any kind of color banding on a global level that's been introduced
6:13
colors still look pretty good and so pretty impressive there. So if we can compare our three-stop
6:20
recovery on the left to our four-stop recovery on the right, we can see that yes, we have added
6:26
in a little bit more noise in terms of the roughness of the pixels. I can also see just a little
6:32
bit of a color banding right at that transition there looking into the actual black area
6:39
you can see that the pixel, you know, kind of roughness is a little bit more obvious
6:45
And if we come down here and take a look, you can see that, you know, the recovery has produced a little bit rougher
6:50
not quite a smooth result with these four stops. And you can also see a little bit in this area that there is some discoloration
7:00
However, it's still not all that extreme. And you can see it show up in some of the transition areas
7:06
But, you know, really all things considered, we've still got a pretty strong result here for a four-stop recovery, looking into our color information here
7:17
Yes, the pixels themselves look a little bit rougher, but at the same time, the color rendition is held up quite well
7:25
So what happens if we compare our recovered four-stop under-exposure result with our baseline image on the left
7:33
Well, obviously we would expect if we look into the out of focus area that the left is going to be smoother, and it certainly is
7:40
There's not really any apparent pixel information there, while there is some roughness here
7:45
But if we kind of scroll through this, I'm not really seeing any kind of strong color banding in the shadows, and so that certainly is impressive
7:54
If we are to look into this area of the image, you know, you can see a bit of a difference, but at the same time, it's not strong
8:03
strongly pronounced here and looking down into this area of the image here
8:08
You know, there's minor, minor variances in our focus, but our results really are not highly different
8:15
The biggest difference I think you're going to see is just in the basic roughness of the pixels
8:20
There is the, you know, the most minute amount of color banding that is introduced at areas of transition
8:27
But all, you know, all in all, I'm really quite impressed by how good that result looks
8:33
for having been recovered so much. Now, how do these results compare to the canon
8:37
And so, looking at our base level, we would expect that, you know, all settings being equal
8:43
that these images are going to look, you know, pretty similar other than what we have discussed in another episode
8:49
We've talked about color science, and so the canon color is a little bit different
8:54
than the Sony color, but we're looking more at just the, you know, the appearance of noise
8:59
And as we can see, you know, both of these cameras look very, very, very
9:03
very good at base ISO, no apparent noise, none that we would expect
9:07
Now, if we compare the one-stop recovery here, we can see that neither one of these cameras
9:14
has introduced any kind of color banding in that area, looking at our, you know, just kind of
9:20
area that should be, you know, consistent color. We see that it is on both of them
9:26
I'm not expecting to see any kind of real impact here. And of course, I'm not seeing one
9:31
If we look in this area here, just, you know, the color information and that transition zone that we looked out before
9:37
Both of them look nice and clean. Now, in our two-stop recovery here, we can see that still both of them look pretty much similar
9:45
They're nice and even and the kind of area that should be completely black back there
9:51
Looking at this transition area, of course, we didn't see any fringing on the Sony as we walk through this
9:57
We're not seeing any on the cannon as well, looking at the amount of, you know
10:01
kind of the pixel roughness and you know how colors are holding up they look at this stage they look
10:07
roughly equal to me in this area here we're not really seeing any color banding on either one of them
10:14
and the cannon is holding up nicely with the two-stop transition here without any kind of real
10:21
impact for recovering that shadow information now if we compare a three-stop result what we are seeing
10:28
is it's very faint but we can see a little bit of noise patterning in the area that
10:34
should be a consistently smooth and it it just more obvious on the Canon than it is on the Sony the Sony at this stage I think is handling that a little bit better Now in terms of the roughness of the pixels I don think of course there more pixels on the Sony than there is on the Canon
10:55
But at the same time, I am not seeing any extra roughness of pixels
11:01
Both of them look roughly equal, which in some ways is, you know, very close to a win for Sony
11:07
just because there are more pixels there. Both of them are still, you know
11:11
holding up nice and smooth as far as that. But we have seen that up in this area here there is
11:18
which tells us that in the image overall, there is just a little bit of some color banding
11:23
that's taking place. But at the same time, you know, kudos to both companies
11:28
you know, a three-stop recovery looks pretty great. Okay, now finally with a four-stop recovery
11:33
And this is where I think that really for Canon shooters, three stops of under exposure, not to say it's your absolute limit, but that's where it's after
11:42
that you really start to pay a price. Now, as you can see, there are some, you know, vertical
11:46
striations here that aren't really apparent in the Sony result and definitely some brighter
11:51
magenta coloration that's showing up. And that also shows up here where you can see some
11:57
magenta patterning. There is a little bit of green patterning in the Sony, which, you know
12:03
reminds me a little bit of the result from the Nikon D850
12:07
in terms of those biases. But if we look inside here, the Sony is much more neutral
12:11
whereas you can see some magenta bleed through that's taking place on the cannon
12:17
And so it's at the four-stop range where I think that Canon starts to take a hit here
12:25
And so it's not to say that it's not usable, it's just it's not as usable maybe as what the Sony result is
12:33
And so as a result, you know, it's just going to impact
12:37
colors a little bit too that may be a little bit less consistent with the cannon now as they were
12:43
with the on the Sony result okay now we're going to do is we're going to push things the opposite
12:47
direction of course dynamic range is not just shadow recovery it's highlight recovery and so here
12:52
we have overexposed by one stop and then on the right side we have drawn that back down so in this
12:59
case we're mostly looking at areas where you know you could have hot spots you can lose
13:04
information due to being blown out. And so it's going to affect the bright areas here the most
13:09
And so here, for example, we have seen that we're able to pull that back around some of this
13:14
color area here. We've pulling that hot spot or and there's not really any true hot spots that
13:20
are here. We're able to bring that down without any kind of loss of information at this point
13:25
Now, in this case, we've overexposed by two stops. We've brought it back down on the right
13:30
And so we can see looking globally at the image that it's held up
13:34
just fine. So if we compare our one-stop recovery on the left and our two-stop recovery on the right
13:39
we can see that it's handled this just fine. You know, for example, this area here, it's still
13:45
recovered that fine. If we look in kind of the transition here around the letters and this area
13:52
that gets completely washed out, we can see that that is held up fine. Also, looking at the front
13:58
fascia of the camera body itself, we can also see here that we haven't lost any information
14:04
information yet. Now if we overexposed by three stops, things start to get a little bit
14:08
rougher to recover. However, if we look at the image globally, there are a few hot spots here
14:14
but the image as a whole doesn't look too bad. But if we look at kind of a few of our key spots
14:21
we're going to see, for example, this area hasn't fully recovered. We can also see that this
14:26
area here hasn't fully recovered. There's a bit of an issue there as well. So if we compare our two-stop
14:31
recovery and our three-stop recovery, we can see how that there are a few areas where we have
14:37
started to lose some information and this has held up okay. You can see just the faintest outline
14:43
from recovery, but here you can see that we've definitely got a pretty big hot spot that has emerged
14:49
Now finally, if we examine the four-stop over-exposure, we can see that we have exposed beyond our
14:55
ability to adequately recover. It doesn't look natural. It doesn't look close any longer. We've
15:01
completely lost the front fascia of the camera, completely blown out all the detail there
15:06
And on our case here, we've just seen everything has pretty much become a mess there
15:11
And so there's your, you know, your upper limits of what's possible. So if you can compare Sony and Canon with a one-stop over exposure and recovery
15:20
we can see that both cameras have done fine here looking at them globally
15:26
And certainly, you know, for example, in our spot here, everything is held up fine
15:30
if we look at the face here. There's no kind of weird edges, no hot spot that has remained
15:36
So both of them doing fine at a one-stop recovery. Now, at a two-stop recovery, we are still, you know, looking fairly similar here
15:46
Both of them have recovered are, you know, kind of blown out areas, areas that are prone
15:51
And if we look here, the cannon has shows maybe just the slightest bit less of an outline here
15:59
The hotspot I think is a little bit more obvious on the Sony than it is on the Canon which frankly is surprising as the Sony is rated a little bit higher when it comes to the dynamic range
16:13
Now at our three-stop recovery, I think that here's where the edge for the Sony starts to come
16:19
in. For example, if you look in this area here, the Sony, while it's starting to lose information
16:23
too, the Canon is losing more. And so that's the beginning of the kind of
16:30
true erosion here. And also if we look over at this area here, both of them are, you know
16:36
quite blown out at this point. So in the extremes, they're kind of equal. But if we look, you know
16:44
for example, at an area like these hotspots, the Sony is hanging on to just a little bit more
16:50
than what the cannon is, though not by a huge margin. Now with either of these, if you over-exposed
16:56
by four stops, I mean, you're done. I mean, the image is just not going to be recoverable
17:02
You're not going to recover sky that is at that kind of extreme. That gives you the practical limits
17:08
of the dynamic range. And so as you can see, the Sony in some areas is a little bit better
17:14
but it really shows up more at the extremes. So in many ways, the results that I saw here
17:21
remind me a lot of when I did the comparison with the Nikon D850
17:26
And in a lot of ways, I think that the D850 and the A7R3 deliver roughly similar results
17:34
And, of course, without comparing them side by side, it would be hard for me to know which gives any kind of edge
17:40
But certainly these two cameras, I believe, are right now the best of the current bunch when it comes to the dynamic range
17:48
Now, we saw with the Canon 5D Mark 4 that it's not as good at the extremes
17:54
and on the under exposure end, basically somewhere between three stops of under exposure and four stops
18:02
the Canon starts to fade when compared to the Sony A7R3. And while in our first episode, when we looked at color science
18:10
we determined that Canon does have an advantage when it comes to the color science there
18:17
But of course, when you're pushing at the extremes, you may get a little bit better result from the Sony in that regard
18:22
Now, when we looked at over exposure, both of the cameras didn't last as long
18:28
And a four shot or four stop of under exposure results are still very, very clean with the A7R3
18:36
But already by three stops of overexposure, you're starting to see some ragged edges start to appear
18:44
And of course, by four stops, there is a slight advantage for the Sony, but it's really a moot point
18:49
because both of them are unusable by four stops of overexposure. And so I would say that you can very successfully recover four stops of under exposure on the A7R3
19:01
but your limit is somewhere between two and three stops of overexposure
19:06
before you start to develop hotspots. And with the cannon, you can really successfully recover more like three stops of under exposure
19:15
before you're going to start to see some color banding issues. And again, somewhere between two and three stops of overexposure before you basically start to fall apart
19:27
Now, the reality for a lot of shooters is that unless you are specifically shooting very, very, you know, high contrast scenes that really need a lot of dynamic range
19:39
the reality is that for most shooters who actually know how to shoot and get shots reasonably right, you're going to have enough latitude, I think
19:47
for recovering from both of these cameras. And after having shot with the 5D Mark 4 for a year and a half at this point
19:57
I found a huge advantage over previous generation Canon cameras when it comes to just kind of general recovery
20:04
And I also find that I resort already to shooting HDR-type images
20:09
or bracketing exposures. I resort to that far less often because, frankly, it's really not all that necessary
20:14
So that's even a little bit more true when you're talking. about the Sony A7R3, which really is a pretty fantastic landscape camera for example
20:24
I'm Dustin Abbott and stay tuned as we continue on with this comparison series and in future episodes we'll be covering ISO performance from both of these cameras
20:35
We'll look at resolution and how that additional 12 megapixels of information on the A7R3, what kind of advantages that gives it
20:43
And so stay tuned for these and future episodes. future episodes as I continue my ongoing coverage of the Sony A7R3. If you haven't already, of course
20:54
you can become a subscriber here, and I will be sharing some of these raw files for my patrons
21:02
And so if you would like to join and become a supportive there at Patreon, you certainly can
21:08
have access to those things. And I want to say thank you to those of you who are already supporting this channel in that way. And of course, thanks for watching today
21:16
Have a great day
#Reference