Canon RF-S18-150mm F3.5-6.3 IS STM

[shinksma]

I wanted to post an interesting tidbit / lens comparison in the thread for this lens, but it doesn't seem to exist. Perhaps because it is "just" a kit lens, and get overlooked by some of the more serious glass now available in the RF series.

If anyone wants to post some typical images, please do. I'll post some when I get a chance. But first, the reason for my posting:

We just transitioned to the R7 from APS-C DSLRs - we have two of the R7 bodies, so we can both enjoy the new tech as we go out on trips or photo shoots. I wanted to compare how a lens we had previously used on our T3i/T2i, the Tamron 16-300mm, did on the new R7 bodies (via EF-R adapter, of course), compared to "something similar" in the RF range, since the 16-300mm was a great all-round travel lens if you weren't planning on publishing in National Geographic (it is a bit soft, even on the old T3i).

One of the R7s we bought came with the RF-S 18-150mm. I think it is a pretty good starter lens, compared to the EF-S 18-55mm I got with the T3i, for example, or the EF-S 55-250mm that was also often a second lens in the T3i kit.

I just did my little experiment with the RF-S 18-150 on one R7, and the Tamron (EF-S) 16-300 on the other R7. I don't have a huge number of images posted here, because they aren't very interesting. I've just included two that show a comparison I discuss at the end.

The extra 2mm at the short end, and the 150mm at the long end, of the Tamron 16-300 really seem to help in capturing the desired shot. But:

At full image size on my desktop (which has a pretty good 4K-ish monitor) I can barely see a difference - the Tamron is a hair softer in overall focus at various ranges. If I pixel peep, I can see the softness in the Tamron. If I crop one of the photos from the RF-S 18-150, taken at 150mm, to simulate the same image I get on the Tamron 16-300 at 300mm, the RF-S 18-150 is actually still a touch clearer at full image size on my monitor. So the extra 150mm on the Tamron is not making up for the better optics of the RF-S 18-150. Back when I was using the Tamron on a T3i, the softness was slightly lost on the less capable sensor and less-accurate AF. Also, as expected with high-ratio telephoto lens, to get the "simulated" 300mm, I did not have to crop the 150mm image in half (50% size), more like 55%. Which probably means the Tamron at full extension is really about 275mm, not truly 300mm.

That being said, the extra 2mm on the short end could potentially be useful, if you can't foot zoom back to get the extra wideness. However, that's where carrying a phone with good camera can grab that "touristy" shot if I don't want to go all artistic with a wide-angle lens on the R7. I still have my Canon EF-S 10-22 and Tokina 12-28mm F/4 (intended for EF-S mount), both of which are "good" wide-angle lenses, IMHO.

In conclusion, the RF-S 18-150mm is a pretty good walk-around non-birding lens that can take a fair bit of cropping and still give reasonable results compared to an older Tamron lens with wider zoom range.

Here are the two shots, one from the RF-S 18-150mm at 150mm, cropped to match the subject size of the other image, taken with the Tamron 16-300 at 300mm. Images resized to 1600px for posting here, with mild sharpening in resize because it seems to match what I see on-screen better. No Post-processing done, except exposure level dropped on the Tamron image, because it came out a bit brighter. Camera settings identical, except exposure changed for zoom level in Av mode (except the R7 IBIS is disabled on the Tamron, I guess).

 

[paddler4]

Images at this size are so downsized that they really can't tell you all that much about lens quality. Even mediocre lenses often produce images that look sharp when downsized this much.

You can look here for a test comparing the two lenses. It will give you side by side comparisons for specific apertures and focal lengths. I'd look particularly at the corners, as sharpness often drops markedly between the center and corners. I don't know which 18-55 you had, so you may need to change the selection for that one.

 

[shinksma]

Images at this size are so downsized that they really can't tell you all that much about lens quality. Even mediocre lenses often produce images that look sharp when downsized this much.

You can look here for a test comparing the two lenses. It will give you side by side comparisons for specific apertures and focal lengths. I'd look particularly at the corners, as sharpness often drops markedly between the center and corners. I don't know which 18-55 you had, so you may need to change the selection for that one.

Yeah, that website is the first place I go to see how well a lens performs in general against other lenses that I know or am considering. However, there are limitations when doing the comparisons - they don't go back and retest all the older lenses on the newer bodies, so I'll never find that EF-S 18-55 of any type tested on an R7, for example.

Plus, the-digital-picture website is doing that test regimen under very specific and repeatable conditions, which is nice for "lab test results". But I also like to see how it does with everyday shooting.

I did consider posting 100% pixel-for-pixel crops (which is what that website does), as that does generally show a lens' limitations, especially at the edges/corners. Maybe I'll take some time to do that later on today.

 

[paddler4]

However, there are limitations when doing the comparisons - they don't go back and retest all the older lenses on the newer bodies, so I'll never find that EF-S 18-55 of any type tested on an R7, for example.

The performance of a given lens should be comparable across bodies that have sensors of the same size and similar pixel counts, unless one body has an AA filter and the other doesn't. Leaving aside AA filters, there is nothing in the optical path between the rear of the lens and the sensor that should affect performance. This assumes you are shooting raw. If you are shooting JPEG, all bets are off because cameras differ in terms of the processing that cameras do to produce JPEGs, and there is no way to separate those effects from lens characteristics when viewing a single image. In the case of Canon bodies, it also matters which picture style you are using to create the JPEGs.

The nice thing about controlled conditions is that it removes extraneous factors that might degrade the image, e.g., camera motion. A test under controlled conditions is an upper bound. That is, under real life circumstances, you won't do better, but you might do worse, for reasons unrelated to the lens itself.

On the other hand, pixel peeping, whether with controlled tests or field tests, doesn't tell you what the results will be like for a given output format. I print a lot, and one of the key variables is the size of the print. As a general rule, the smaller the print, the less a lack of sharpness in the lens matters. I print up to 17 x 22 (roughly A2), so deficiencies in the lens matter more to me than to someone who only prints small or views images on a screen.

So my approach to this is:

1. Look at controlled tests (or do them myself) to get a good idea of the best I can expect, and then
2. Filter that to ignore stuff that won't matter for the work I do.

I never use online images for this.

I've exhibited a fair amount, and it seems like the people viewing my images rarely notice flaws that I see in them.

 

[shinksma]

The performance of a given lens should be comparable across bodies that have sensors of the same size and similar pixel counts

I agree. And if I take the pixel count and FF vs APS-C format differences into account, I can rationalize what I'm seeing. But even taking the same lens and comparing it across the multiple bodies that may have been tested in the past, it's annoying (IMHO) to look at the test images and scale what that really means. e.g. the Canon EF-S 18-55mm f/3.5-5.6 IS II on the 7DII vs the 50D - the images appear differently due to the different pixel density on the sensors, so you have to assess one vs the other as purely a result of the pixel counts, and convince yourself that is the equivalent image quality from that lens (because it should be). Doing that when comparing an older lens on an older body with a lower pixel count, vs a newer lens on a newer body with a much higher pixel count (e.g. the R7) it can be a bit misleading.

e.g. the Canon RF-S 18-45mm F4.5-6.3 IS STM on an R7 vs the Canon EF-S 18-55mm f/3.5-5.6 IS STM on a 60D.

So while I value the availability of the tool, for me it is most useful when comparing different lenses on the same body (or at least identical sensor density/size), because then it becomes a true "blink comparator" tool. That's why I did the experiments above - the same body, and, while not a controlled lab test like the linked website, my test demonstrates to my satisfaction that there are indeed differences, but as you pointed out, when published to the web at 1600 px size, you really can't tell in this case - which is what a lot of people end up doing with their photos. Not everyone prints to 18"x24". My wife and I do - the house is full of large and not-so-large prints on various media.

 

[mcoren]

The RF-S 18-150 is a fantastic lens for the price! I loved the EF-M version, and I was happy that Canon chose this as one of the first RF-S lenses, and the R7 kit lens. That made the R7 kit (versus body only) a no-brainer for me.

I agree that an extra couple of mm at the wide end would help a lot.

This lens really produces great IQ for an all-in-one travel lens. With the EF-M version, I've shot straight into the afternoon sun with almost no flare.

As for its durability, you can read my experience here.

 

[shinksma]

As for its durability, you can read my experience here.

Ah yes, I remember reading the cat and gravity tag team vs the lens and camera combo. Ouch!

 

[mcoren]

Ah yes, I remember reading the cat and gravity tag team vs the lens and camera combo. Ouch!

Yes, gravity. My old nemesis.

I'm waiting to see how the rumors of new "higher-end" Canon RF-S lenses pan out, but I wouldn't have any misgivings about getting another 18-150 if I needed the FL range in a single package.

Brevity by Dan Thompson for July 28, 2005 | GoComics.com

May I present Sir Isaac Newton. He claims to have discovered something big.

www.gocomics.com

 

[Mk1Racer]

I saw one review of this lens, where they compared it to an RF 24-105 f/4 L IS, on a R7 body. The general conclusion was that the RF-S 18-150 was pretty comparable to the RF 24-104 f/4 L, at similar focal lengths. It seems that the 24-105 f/4 L IS really shines on full-frame cameras, but the RF-S 18-150 gets comparable results, while being wider and longer, on an APS-C camera. It certainly seems to be better quality than they old EF-S kit lenses. We know Canon knows how to make good lenses designed for APS-C sensors, by virtue of the EF-S 17-55 f/2.8 IS USM. The rumor is, there should be more RF-S lenses to coincide w/ the launch of the R7 II.