24-240 / 24-105 7.1 / 24mm IS comparison

tkbslc wrote:

I tried again as someone mentioned maybe RAW gives better results. But when I viewed the JPEGs, I found the 24-105 seems to be a lot closer today. It could be field curvature? I made sure to focus at the exact spot where these crops were taken. Or maybe the lens just didn't focus as crisply yesterday.

These are DLO JPEGs at 24mm f8. 24-105 on the right. (color difference is from cloud movement between photos)

Top corner, 100% crop from 4K monitor. Light slightly changed while changing lenses.

Tree looks in front of the focal plane to me. House fence railing excellent by comparison.

I have to keep things in perspective as these aren't professional lenses, but from my own experience of having an EF 24-105 with soft corners at 24mm, printing either of these wouldn't be good enough - for me. I hate areas that are just a little soft! (As opposed to intentionally OOF).

RexTremendae wrote:

tkbslc wrote:

RexTremendae wrote:

The RF 24-240 is excellent as you can see from these images. My two issues with it are pretty minor, but still a problem. First, the zoom ring is stiff and jerky, forget about doing smooth zooms in videos. Second, the video servo AF tends to hunt on the EOS R/RP. All in all the AF tends to miss a little too much to reach "pro" level for photo or video. Optically, it's quite good as you noticed. I would classify it as ridiculously good for the price and range. It's actually prime like from 35mm to 105mm IMHO. If it had a nice smooth zoom and dual AF motors like the RF 70-200 I don't know if most people would ever need another lens. Canon should do an L version of this lens with a larger aperture and better build quality and AF.

I can agree it's quite good in the middle. Comparing with my 70-200 F4 L IS the 24-240 there is really no difference in sharpness from about 70-135mm. I actually think it is better than the 24-105L from 85-105.

My main issues are the massive corrections it requires. My preferred software is ACDSee, but there are no RF lens profiles. So the 24-240 is basically unusable with that software as a result. Also I find that with lower contrast targets it gives up at the long end really quickly. Have to manual focus or zoom out to acquire focus.

It's also got pretty weak corners at 240mm, but I often centrally frame at that focal length. If I was doing landscapes, I'd probably stop in the mid 100s for corner sharpness.

I always stop down to f/7.1 at the long end. It's sharp at f/6.3 but tends to just miss focus, maybe it's the extra DoF but f/7.1 works most of the time for static subjects.

If they made an upgraded version of this lens, I think rather than a larger aperture, I'd want it to be better corrected and cover the whole sensor at wide angle. Nikon's 24-200 and Tamron's 28-200 both seem better optically, so maybe cutting a little off the telephoto is a good idea.

Are they better optically? You tell me:

Maybe I just got a good copy but I think the image quality is more than good enough. This was shot by my girlfriend from our car.

Great image! But for some of us, it's the corner performance at 24mm that is important.

My EF24-105 is a quite decent performer from about 35-80mm, so pairs well with my 16-35, but I would so like an all round lens that does a good job from 24mm. [Cheap option for me appears to be the EF 24-70/4, more expensive is going RF and the 24-105/4. You get what you pay for...]

tkbslc wrote:

highdesertmesa wrote:

tkbslc wrote:

highdesertmesa wrote:

jwilliams wrote:

highdesertmesa wrote:

FWIW, the 24-105 STM is much better corrected using RAW and doing so manually in Capture One — it corrects to where it looks better than the 24-240 in the corners at 24mm. Canon didn’t do enough peripheral sharpening correction in their processing, and Lightroom/Adobe simply mimics what Canon does in-camera. On the 24-240, the distortion correction is too extreme for extra sharpening in the corners to make any difference. My guess is that given the resources available during the pandemic, Canon engineers didn’t pursue peripheral sharpening on the 24-105 STM because it hadn't helped the the 24-240.

Would love to see some examples if possible. I have the f71. and 4L versions of the 24-105. Love the size/weight of the f7.1, especially on the RP. I don't yet have SW setup to do RAW conversions so I can only see JPG results. Would love to see how much native distortion the f7.1 has and how it can be cleaned up so to speak at 24mm.

I'm terrible at RAW processing, but I have wondered how the f7.1 version would perform if you did the distortion correction manually and didn't try to get all of it out. In other words partially correct the distortion while leaving some of it in hopes that sharpness would be better. For scenes without straight line this might produce better results.

I didn't have my original 24-105 STM test shots any more, so I took a few new ones. I took a series at 24mm from f/4 through f/16, and f/8 was the clear stand-out winner for best corner performance. I think when I did this last time, I was looking at f/4 or 5.6, which I don't think the in camera corrections can clean up as well. Anyway, when processing them manually, I think it's pretty much a wash compared to in-camera. There may be a bit more detail when corrected manually, but the in-camera corrections are better at chromatic aberration correction, so I think that gives an edge to Canon's JPEG versus what I was doing. Here are all three for reference on the R6. Differences probably become more apparent on the R5 of course.

I did notice on the R6 it has a new setting (or at least I never noticed it on the R) under Digital Lens Optimizer – options for Standard and High. Perhaps these lenses (24-105 STM and 24-240) would benefit from the High setting.

That seems like a pretty big sharpness difference in the edges to me. I will have to try some manual corrections. I'd be a lot happier with the Capture 1 result.

Yeah, I think there is more detail manually, even if it has a little minor chromatic aberration left over. Capture One will be even better once they add a profile for this lens so that the distortion correction is more accurate. But I'm not sure if it's worth the hassle of RAW processing if you're used to using JPEG – maybe if you want to print really big it would be worth it, like 30'x20" and larger.

To be honest, I don't have many complaint about this 24-105 STM lens for everyday casual stuff. I think it's actually quite good. Even in the sample I posted in the OP, it doesn't look bad at screen size on a 28" monitor. I think you could do 13x19" prints on it just fine, even at 24mm f8. For the rest of the range, I find it close enough to the 24-240 to call it equal,.

But it would be nice if we could get sharp corners for more serious shots. The lens is light enough I would like to travel or hike with it. Switching to RAW for detail shots at is not too much of a chore if that's what it takes.

I tried setting DLO on High just now – in-camera JPEG results look over-sharpened across the frame on that setting. I would have thought that DLO would have only amped up the corrective sharpness toward the corners, but it looks like it applies it across the frame. I do have my Picture Style sharpness maxed out, so maybe for the DLO High setting, I need to return Picture Style sharpness to the default/middle.

That's unfortunate. I'm not a fan of oversharpening. It starts to look like a phone photo if you go too far.

I'm a bit of a perfectionist, but printing the OP's 24-105 sample at 13x19 would be a throw away for me. Those corners are just too soft for my landscape work and standards.

Peak freak wrote:

tkbslc wrote:

I tried again as someone mentioned maybe RAW gives better results. But when I viewed the JPEGs, I found the 24-105 seems to be a lot closer today. It could be field curvature? I made sure to focus at the exact spot where these crops were taken. Or maybe the lens just didn't focus as crisply yesterday.

These are DLO JPEGs at 24mm f8. 24-105 on the right. (color difference is from cloud movement between photos)

Top corner, 100% crop from 4K monitor. Light slightly changed while changing lenses.

Tree looks in front of the focal plane to me. House fence railing excellent by comparison.

The focus point is actually a bush that is even with the tree just outside the crop.  But this is 24mm f8 and DOF is basically everything from 5 feet to infinity.

I have to keep things in perspective as these aren't professional lenses, but from my own experience of having an EF 24-105 with soft corners at 24mm, printing either of these wouldn't be good enough - for me. I hate areas that are just a little soft! (As opposed to intentionally OOF).

That's why I have the 24mm prime.  But I was hoping for a little more.

tkbslc wrote:

I tried again as someone mentioned maybe RAW gives better results. But when I viewed the JPEGs, I found the 24-105 seems to be a lot closer today. It could be field curvature? I made sure to focus at the exact spot where these crops were taken. Or maybe the lens just didn't focus as crisply yesterday.

These are DLO JPEGs at 24mm f8. 24-105 on the right. (color difference is from cloud movement between photos)

Top corner, 100% crop from 4K monitor. Light slightly changed while changing lenses.

Sometimes with wide angles, the corners can record rotational movement more than the center and look blurred. And sometimes IBIS has to move the sensor enough that one or two corners suffer because the image circle on these corrected lenses is already being stretched thin.