Canon 5D Mark IV -- Time to Discuss!

Charlie wrote in post #18147095
what I'm trying to get at is that micro adjust on a higher mp camera may be different? I'm not completely sure. I'm guessing that the micro movements are even finer than the 5D3 by a tune of 50%, so on a 5D3, dot tune of +6 is equivalent to +9 on a 5D4, since the pixel pitch is finer? just a theory, as I dont know how much exactly is a micro movement.

As TeamSpeed's noted: it's done in fractions of the DOF, so I assume it'd be the same regardless of the sensor resolution.

I have now had time to do a much more detailed DotTune and general shooting test, both on a tripod and handheld; and the news isn't good.

There's unquestionably a strong tendency for the camera to front focus when using the left sided AF points. A little with the right sided points, but likely within margins that you wouldn't notice under normal shooting. The left sided points however are essentially unusable with a fast lens.

I've tried a 24-70II, a 70-200II (that was previous calibrated by a Canon service centre) and a 100L macro. All three show the same problem, regardless of whether you're on a tripod or handheld.

By the time I took the two shots shown below I was working under bulb light - hence the noise, but daylight or bulb, handheld or tripod, the results are consistent.

Scrumhalf wrote in post #18147215
Roger Cicala's 5D4 review:

https://www.lensrental​s.com …ews-the-canon-5d-mark-iv/

Good review.

I like how he said the focusing is the strongest update.

Let's face it, you can have the most whiz bang camera in the universe, but if the images are OOF, then they are no good.

sploo wrote in post #18147299
As TeamSpeed's noted: it's done in fractions of the DOF, so I assume it'd be the same regardless of the sensor resolution.

I have now had time to do a much more detailed DotTune and general shooting test, both on a tripod and handheld; and the news isn't good.

There's unquestionably a strong tendency for the camera to front focus when using the left sided AF points. A little with the right sided points, but likely within margins that you wouldn't notice under normal shooting. The left sided points however are essentially unusable with a fast lens.

I've tried a 24-70II, a 70-200II (that was previous calibrated by a Canon service centre) and a 100L macro. All three show the same problem, regardless of whether you're on a tripod or handheld.

By the time I took the two shots shown below I was working under bulb light - hence the noise, but daylight or bulb, handheld or tripod, the results are consistent.

Hosted photo: posted by sploo in
./showthread.php?p=181​47299&i=i146284441
forum: Canon Digital Cameras

Both are crops from full images. The right hand shot had the centre of the chart under the very centre AF point. The left shot had the centre of the chart under the leftmost cross type point on the middle row. I was taking care to ensure the chart was as parallel as I could get to the sensor plane.

It may not be that clear from the shots below (slightly reduced in resolution for posting) but the front "90" on the left image is about the same sharpness as the rear "10" on the right side.

This is pretty much what I get with all shots (with some minor AF variation, of course). AF with Liveview and it's bang on.

I'm at a bit of a loss what to do now. The service centre where I sent the first body claimed they put it on a test rig and checked all the AF points as good, but it showed the same problem.

I guess my only route is to send this body and a couple of the lenses to them to try; obviously more time without the gear, and expense

I have not followed all your posts, but it's odd.

I checked my outer focus points when you first reported them and I think they are fine.

I was a lot less scientific than you though. I adjusted the lenses using centre point, I then chose something with some fine detail like my cats ear, or cd cases in my book case and focused using the outer points, and then zoomed in on camera and inspected. They honestly looked perfect.

Also my lenses are not as good as yours.

sploo wrote in post #18147299
As TeamSpeed's noted: it's done in fractions of the DOF, so I assume it'd be the same regardless of the sensor resolution.

I have now had time to do a much more detailed DotTune and general shooting test, both on a tripod and handheld; and the news isn't good.

There's unquestionably a strong tendency for the camera to front focus when using the left sided AF points. A little with the right sided points, but likely within margins that you wouldn't notice under normal shooting. The left sided points however are essentially unusable with a fast lens.

I've tried a 24-70II, a 70-200II (that was previous calibrated by a Canon service centre) and a 100L macro. All three show the same problem, regardless of whether you're on a tripod or handheld.

By the time I took the two shots shown below I was working under bulb light - hence the noise, but daylight or bulb, handheld or tripod, the results are consistent.

Hosted photo: posted by sploo in
./showthread.php?p=181​47299&i=i146284441
forum: Canon Digital Cameras

Both are crops from full images. The right hand shot had the centre of the chart under the very centre AF point. The left shot had the centre of the chart under the leftmost cross type point on the middle row. I was taking care to ensure the chart was as parallel as I could get to the sensor plane.

It may not be that clear from the shots below (slightly reduced in resolution for posting) but the front "90" on the left image is about the same sharpness as the rear "10" on the right side.

This is pretty much what I get with all shots (with some minor AF variation, of course). AF with Liveview and it's bang on.

I'm at a bit of a loss what to do now. The service centre where I sent the first body claimed they put it on a test rig and checked all the AF points as good, but it showed the same problem.

I guess my only route is to send this body and a couple of the lenses to them to try; obviously more time without the gear, and expense

I wouldn't send any lenses in, I'd just send the body back, with your data and request a replacement body (and cross my fingers this was a good one).

sploo wrote in post #18147299
As TeamSpeed's noted: it's done in fractions of the DOF, so I assume it'd be the same regardless of the sensor resolution.

I have now had time to do a much more detailed DotTune and general shooting test, both on a tripod and handheld; and the news isn't good.

There's unquestionably a strong tendency for the camera to front focus when using the left sided AF points. A little with the right sided points, but likely within margins that you wouldn't notice under normal shooting. The left sided points however are essentially unusable with a fast lens.

I've tried a 24-70II, a 70-200II (that was previous calibrated by a Canon service centre) and a 100L macro. All three show the same problem, regardless of whether you're on a tripod or handheld.

By the time I took the two shots shown below I was working under bulb light - hence the noise, but daylight or bulb, handheld or tripod, the results are consistent.

Hosted photo: posted by sploo in
./showthread.php?p=181​47299&i=i146284441
forum: Canon Digital Cameras

Both are crops from full images. The right hand shot had the centre of the chart under the very centre AF point. The left shot had the centre of the chart under the leftmost cross type point on the middle row. I was taking care to ensure the chart was as parallel as I could get to the sensor plane.

It may not be that clear from the shots below (slightly reduced in resolution for posting) but the front "90" on the left image is about the same sharpness as the rear "10" on the right side.

This is pretty much what I get with all shots (with some minor AF variation, of course). AF with Liveview and it's bang on.

I'm at a bit of a loss what to do now. The service centre where I sent the first body claimed they put it on a test rig and checked all the AF points as good, but it showed the same problem.

I guess my only route is to send this body and a couple of the lenses to them to try; obviously more time without the gear, and expense

Sploo- I am not certain that I follow your post and method correctly so forgive me if I have something wrong but it seems like you are either moving the camera or the target during your testing. If that is the case, without some kind of very strict control there is a lot of potential for inconsistent results. Are you able to reproduce these results consistently even if camera/target movement is involved? I notice you mention that you were 'taking care' to ensure the chart and focal plane were parallel. Before Canon centralized service there was a service/repair centre about 20 minutes from my house and I had the opportunity to go inside on a couple of occasions. I assure you that Canon has tools and methods far more superior and accurate than what we can do in our homes with a few charts. If they gave the 1st unit a good bill of health I'm afraid to say that's likely what you'll get this time too and you still won't be happy with the unit. I would suggest that you save yourself the trouble and go for a return/exchange again.

George Zip wrote in post #18147326
I have not followed all your posts, but it's odd.

I checked my outer focus points when you first reported them and I think they are fine.

I was a lot less scientific than you though. I adjusted the lenses using centre point, I then chose something with some fine detail like my cats ear, or cd cases in my book case and focused using the outer points, and then zoomed in on camera and inspected. They honestly looked perfect.

Also my lenses are not as good as yours.

It'll be more obvious with sharper/faster lenses. At 70mm and, say, f/5.6, it would be harder to spot (though it is a long way off).

I've tried a variety of focal lengths and camera-to-subject distances, and the image results match what DotTune is telling me (i.e. that the left sided AF points would require a significant [about +10] difference in MFA to correct the shot). In fact, I even made that change and fired off a few shots to prove it would work, but obviously then the centre points are way off.

I.e. everything I'm seeing is consistent with the left sided points front focusing, but according to the service centre the first body rated as fine on their rig. To be fair one, of their techs asked me to update them with the results of the new body (which I've done) as he conceded they'd have to feed back to Canon if there was a clear discrepancy between their test system and what the camera was doing.

STIC wrote in post #18147330
I wouldn't send any lenses in, I'd just send the body back, with your data and request a replacement body (and cross my fingers this was a good one).

The problem is, this is already body number 2. The first one went to the service centre, was adjusted and claimed good (but still showed the same problem). That one went back to the vendor, and this one is the replacement.

nqjudo wrote in post #18147360
Sploo- I am not certain that I follow your post and method correctly so forgive me if I have something wrong but it seems like you are either moving the camera or the target during your testing. If that is the case, without some kind of very strict control there is a lot of potential for inconsistent results. Are you able to reproduce these results consistently even if camera/target movement is involved? I notice you mention that you were 'taking care' to ensure the chart and focal plane were parallel. Before Canon centralized service there was a service/repair centre about 20 minutes from my house and I had the opportunity to go inside on a couple of occasions. I assure you that Canon has tools and methods far more superior and accurate than what we can do in our homes with a few charts. If they gave the 1st unit a good bill of health I'm afraid to say that's likely what you'll get this time too and you still won't be happy with the unit. I would suggest that you save yourself the trouble and go for a return/exchange again.

I wouldn't claim to be matching Canon test standards, but I was careful to ensure the sensor plane was as parallel as I could get to the chart plane - so when I'm testing a left sided AF point I move the tripod or chart to get it centred under the relevant AF point, but maintain the parallel nature (i.e. I'm not rotating the camera, which would make the plane of focus different to the plane of the chart).

Finally, I was just doing a lot of handheld shooting - still generally trying to keep the chart parallel, but basically replicating what would be a more normal portrait shoot scenario; the results of that match the tripod results, which match the DotTune results.

Basically, if you (or anyone else) have an f/2.8 70mm lens that you'd MFA'd as good on the centre point then I'd be grateful if you could try a few shots with the leftmost centre AF point. You can see from the image I posted that the left chart itself is clearly not in focus - and that's consistent (it's not a case that I've "cherry picked" one bad image from several).

Colorblinded wrote in post #18146943
Anyone got some challenging landscape or other images with a broad dynamic range that they wouldn't mind sharing RAW files for? I really want to see for myself in LR what can be done with these.

Wow the Dynamic Range is as good as my A7r2 in real world.

Check this out

IMAGE LINK: https://flic.kr/p/MNEb​ES

IMAGE LINK: https://flic.kr/p/MFCc​N5

IMAGE LINK: https://flic.kr/p/MFCb​4y

IMAGE LINK: https://flic.kr/p/MEjP​dL

sploo wrote in post #18147372
I wouldn't claim to be matching Canon test standards, but I was careful to ensure the sensor plane was as parallel as I could get to the chart plane - so when I'm testing a left sided AF point I move the tripod or chart to get it centred under the relevant AF point, but maintain the parallel nature (i.e. I'm not rotating the camera, which would make the plane of focus different to the plane of the chart).

Finally, I was just doing a lot of handheld shooting - still generally trying to keep the chart parallel, but basically replicating what would be a more normal portrait shoot scenario; the results of that match the tripod results, which match the DotTune results.

Basically, if you (or anyone else) have an f/2.8 70mm lens that you'd MFA'd as good on the centre point then I'd be grateful if you could try a few shots with the leftmost centre AF point. You can see from the image I posted that the left chart itself is clearly not in focus - and that's consistent (it's not a case that I've "cherry picked" one bad image from several).

Ok. Interesting. If I get home early enough I'll give it a try. If not, tomorrow.

sploo wrote in post #18147363
The problem is, this is already body number 2. The first one went to the service centre, was adjusted and claimed good (but still showed the same problem). That one went back to the vendor, and this one is the replacement.

I get that...but, the replacement is faulty so, back it goes again with, if it were me, a warning to NOT be sending another defective product...

You don't have to put up with shoddy merchandise, no matter how many examples of it they send you...

STIC wrote in post #18147385
I get that...but, the replacement is faulty so, back it goes again with, if it were me, a warning to NOT be sending another defective product...

You don't have to put up with shoddy merchandise, no matter how many examples of it they send you...

Granted, but what happens if I pull the gamble handle again and get the same issue? I could just spend forever swapping cameras

I've mailed the service centre with the details and I'll see what they suggest.

Anyone know a dealer with no ties to NY that currently has a mark iv in stock? MPEX doesn't have stock yet. Thats where I typically shop for large purchases. Thanks.

sploo wrote in post #18147372
I wouldn't claim to be matching Canon test standards, but I was careful to ensure the sensor plane was as parallel as I could get to the chart plane - so when I'm testing a left sided AF point I move the tripod or chart to get it centred under the relevant AF point, but maintain the parallel nature (i.e. I'm not rotating the camera, which would make the plane of focus different to the plane of the chart).

Finally, I was just doing a lot of handheld shooting - still generally trying to keep the chart parallel, but basically replicating what would be a more normal portrait shoot scenario; the results of that match the tripod results, which match the DotTune results.

Basically, if you (or anyone else) have an f/2.8 70mm lens that you'd MFA'd as good on the centre point then I'd be grateful if you could try a few shots with the leftmost centre AF point. You can see from the image I posted that the left chart itself is clearly not in focus - and that's consistent (it's not a case that I've "cherry picked" one bad image from several).

I just checked it out with a 70-200mm (Tamron) which is MFA'd for the centre point set at @ 70mm - F/2.8 and cannot see any difference in focus/sharpness when using the centre point or any of the peripheral points (both left and right). Are you sure it is not your lens? Did you try it with another similar lens?

sploo wrote in post #18147372
I wouldn't claim to be matching Canon test standards, but I was careful to ensure the sensor plane was as parallel as I could get to the chart plane - so when I'm testing a left sided AF point I move the tripod or chart to get it centred under the relevant AF point, but maintain the parallel nature (i.e. I'm not rotating the camera, which would make the plane of focus different to the plane of the chart).

Finally, I was just doing a lot of handheld shooting - still generally trying to keep the chart parallel, but basically replicating what would be a more normal portrait shoot scenario; the results of that match the tripod results, which match the DotTune results.

Basically, if you (or anyone else) have an f/2.8 70mm lens that you'd MFA'd as good on the centre point then I'd be grateful if you could try a few shots with the leftmost centre AF point. You can see from the image I posted that the left chart itself is clearly not in focus - and that's consistent (it's not a case that I've "cherry picked" one bad image from several).

Sploo - Ok. I just did some quick tests and referred back to the shots of the MFA charts I took just a few days ago. Generally speaking my results are sharp across the entire frame. I can't reproduce your results exactly but I can get something similar by using a shallow DOF and manipulating the focal plane away from parallel with the target. I can see this shift live when I magnify in LiveView mode and make very small shifts using a geared tripod head. That said I can always find the sweet spot to get everything sharp again. I'm not sure how helpful that is to you but I hope it gives you some insight.

sploo wrote in post #18147299
As TeamSpeed's noted: it's done in fractions of the DOF, so I assume it'd be the same regardless of the sensor resolution.

I have now had time to do a much more detailed DotTune and general shooting test, both on a tripod and handheld; and the news isn't good.

There's unquestionably a strong tendency for the camera to front focus when using the left sided AF points. A little with the right sided points, but likely within margins that you wouldn't notice under normal shooting. The left sided points however are essentially unusable with a fast lens.

I've tried a 24-70II, a 70-200II (that was previous calibrated by a Canon service centre) and a 100L macro. All three show the same problem, regardless of whether you're on a tripod or handheld.

By the time I took the two shots shown below I was working under bulb light - hence the noise, but daylight or bulb, handheld or tripod, the results are consistent.

Hosted photo: posted by sploo in
./showthread.php?p=181​47299&i=i146284441
forum: Canon Digital Cameras

Both are crops from full images. The right hand shot had the centre of the chart under the very centre AF point. The left shot had the centre of the chart under the leftmost cross type point on the middle row. I was taking care to ensure the chart was as parallel as I could get to the sensor plane.

It may not be that clear from the shots above (slightly reduced in resolution for posting) but the front "90" on the left image is about the same sharpness as the rear "10" on the right side.

This is pretty much what I get with all shots (with some minor AF variation, of course). AF with Liveview and it's bang on.

I'm at a bit of a loss what to do now. The service centre where I sent the first body claimed they put it on a test rig and checked all the AF points as good, but it showed the same problem.

I guess my only route is to send this body and a couple of the lenses to them to try; obviously more time without the gear, and expense

EDIT: The numbers on the chart roughly correspond to the distance in front/behind the chart plane - so in the example above the left point is front focusing by ~90mm.

I understand what your saying. Due to family and work issues I won't have a chance to see if I can reproduce this until Friday. I'll trying using teamspeeds helpful chart and report back. Good luck on whatever decision you make.