5D4 Studio ISO comparison shots added at DPR!

[ahsanford]

5D4 ISO comparison tool posted at DPR:

https://www.dpreview.com/reviews/image-comparison?attr18=daylight&attr13_0=canon_eos5dmkiv&attr13_1=canon_eos5dmkiii&attr13_2=nikon_d810&attr13_3=sony_a7rii&attr15_0=jpeg&attr15_1=jpeg&attr15_2=jpeg&attr15_3=jpeg&attr16_0=100&attr16_1=100&attr16_2=100&attr16_3=100&attr171_3=off&normalization=full&widget=1&x=0&y=0

- A

 

[transpo1]

This is awesome! Thanks for posting. Looks like it actually compares pretty favorably (with noise reduction efforts) to the A7RII at 102400

 

[candc]

The 5div images look soft compared to the other cameras, even the the 5diii

 

[Larsskv]

The 5div images look soft compared to the other cameras, even the the 5diii

They do. No doubt it is the lens, or the focus. I noticed, they still use that Canon 85 f/1.8, but put the Nikon 85 f1.4G on the D810... A sharper lens at 85, across the frame, is the 70-200 f/2.8L IS II.

 

[neuroanatomist]

The 5div images look soft compared to the other cameras, even the the 5diii

A word of caution - literally in the sense that yellow means caution. The info (i) button for the 5DIV shot is yellow to indicate that the images are processed with a beta version of ACR. Quite likely, Adobe's first step is to just get the software to open the images, and later they'll refine the software with sensor-specific code to optimize the handling of 5DIV images and it won't be a beta version any more.

 

[ahsanford]

The 5div images look soft compared to the other cameras, even the the 5diii

A word of caution - literally in the sense that yellow means caution. The info (i) button for the 5DIV shot is yellow to indicate that the images are processed with a beta version of ACR. Quite likely, Adobe's first step is to just get the software to open the images, and later they'll refine the software with sensor-specific code to optimize the handling of 5DIV images and it won't be a beta version any more.

From my very untrained eyes -- and just looking at RAW files -- I'd say the 5D4 very crudely looks about a half-stop better (perhaps a bit less) than the 5D3 around the 6400-12800 neighborhood. I top out my 5D4 ideally at 6400 for handheld low light use, but will walk it up to 8K if needed, so that's where I was looking with this comparison.

But this testing (as informal as it is and as crudely as my eyes are looking at it), I am not seeing a high ISO game-changer so much as a small iterative improvement. The 5D3 holds up very well in comparison given its age.

- A

 

[candc]

The 5div images look soft compared to the other cameras, even the the 5diii

A word of caution - literally in the sense that yellow means caution. The info (i) button for the 5DIV shot is yellow to indicate that the images are processed with a beta version of ACR. Quite likely, Adobe's first step is to just get the software to open the images, and later they'll refine the software with sensor-specific code to optimize the handling of 5DIV images and it won't be a beta version any more.

I suppose they want to be first to post samples but they shouldn't use a beta version acr for the studio comparison. Doing so gives a bad first impression that the camera produces soft images.

 

[StudentOfLight]

DPReview also posted their 6-stop push images:
https://www.dpreview.com/reviews/image-comparison/fullscreen?attr144_0=canon_eos5dmkiv&attr144_1=canon_eos5dsr&attr144_2=nikon_d810&attr144_3=sony_a7rii&attr146_0=100_6&attr146_1=100_6&attr146_2=64_7&attr146_3=100_6&attr177_3=off&normalization=full&widget=333&x=0.0010714658086581354&y=0.4399426933149564

All I can say is I am extremely happy with what I see in all the studio samples they've posted. Looks to have the perfect balance of low ISO and mid-to-high ISO performance, manageable file sizes and with a little bit of sharpening and clarity you have excellent output. I've attached a screenshot of LR comparison of the D810 (unedited) and 5D-IV (sharpened). Is the the AA filter ruining the 5D-IV image?

I've also attached a moire comparison with the A7R-II. The 5D-IV image is softer, but the moire is well suppressed. Is the lack of moire ruining the 5D-IV image? AFAIK, there is no quick fix for moire, I think you need to go in by hand and paint it out in lightroom or take an image into photoshop and use layers and blends. Bottom line (AFAIK) less moire is less work. (Who knows maybe dp-RAW microadjust could also help with moire ???)

Anyway, from an IQ perspective I'm extremely pleased by what Canon has put forward.

 

[StudentOfLight]

The 5div images look soft compared to the other cameras, even the the 5diii

A word of caution - literally in the sense that yellow means caution. The info (i) button for the 5DIV shot is yellow to indicate that the images are processed with a beta version of ACR. Quite likely, Adobe's first step is to just get the software to open the images, and later they'll refine the software with sensor-specific code to optimize the handling of 5DIV images and it won't be a beta version any more.

From my very untrained eyes -- and just looking at RAW files -- I'd say the 5D4 very crudely looks about a half-stop better (perhaps a bit less) than the 5D3 around the 6400-12800 neighborhood. I top out my 5D4 ideally at 6400 for handheld low light use, but will walk it up to 8K if needed, so that's where I was looking with this comparison.

But this testing (as informal as it is and as crudely as my eyes are looking at it), I am not seeing a high ISO game-changer so much as a small iterative improvement. The 5D3 holds up very well in comparison given its age.

- A

My eyes are in agreement with your eyes. Based on what DPReview has shared, I believe the 5D-IV will chart very similarly to the A7R-II in terms of DR... so a bit better than the 1Dx-II at ISO 100 but not as high as the D810 at ISO 64. I'm guessing probably in the range of 13.5-13.8 at ISO 100. It looks to have a very good mid-to-high ISO range, very similar to the 1D-X II. From ISO 12,800 upwards the 1Dx-II has a 1/2 stop advantage.

I'd speculatively chart the DR (in the 8MP DXO downsample method) for the 5D-IV as follows:

 

[Cochese]

A much better example of how well this camera can do in low light, check out some real world downloadable RAWS (DNG) from the 5DM IV.
I'm definitely seeing some improvement.

Here's a shot from a dimly lit studio at ISO 102k
http://froknowsphoto.com/canon-eos-5d-mark-iv-raw-files/

 

[StudentOfLight]

A much better example of how well this camera can do in low light, check out some real world downloadable RAWS (DNG) from the 5DM IV.
I'm definitely seeing some improvement.

Here's a shot from a dimly lit studio at ISO 102k
http://froknowsphoto.com/canon-eos-5d-mark-iv-raw-files/

Thanks for sharing.

I think the camera can perform a little better than the ISO 102k shot demonstrates. (See the histogram in attach) To me it looks like the shot is about 1 stop under-exposed. To make the maximum use of the camera's dynamic range we'd need an "Expose-To-The-Right" image, perhaps taken at 1/160 f/4.5 or 1/250 f/3.5 given the available light. Hopefully we'll see in his full review.

 

[ritholtz]

A much better example of how well this camera can do in low light, check out some real world downloadable RAWS (DNG) from the 5DM IV.
I'm definitely seeing some improvement.

Here's a shot from a dimly lit studio at ISO 102k
http://froknowsphoto.com/canon-eos-5d-mark-iv-raw-files/

Thanks for sharing.

I think the camera can perform a little better than the ISO 102k shot demonstrates. (See the histogram in attach) To me it looks like the shot is about 1 stop under-exposed. To make the maximum use of the camera's dynamic range we'd need an "Expose-To-The-Right" image, perhaps taken at 1/160 f/4.5 or 1/250 f/3.5 given the available light. Hopefully we'll see in his full review.

Not sure if I am getting it right. According to dpr sensor is iso invariant above iso 400. Under exposing 1 stop means, this pic is equivalent of shooting at iso 204k.

 

[sebasan]

My eyes are in agreement with your eyes. Based on what DPReview has shared, I believe the 5D-IV will chart very similarly to the A7R-II in terms of DR... so a bit better than the 1Dx-II at ISO 100 but not as high as the D810 at ISO 64. I'm guessing probably in the range of 13.5-13.8 at ISO 100. It looks to have a very good mid-to-high ISO range, very similar to the 1D-X II. From ISO 12,800 upwards the 1Dx-II has a 1/2 stop advantage.

I'd speculatively chart the DR (in the 8MP DXO downsample method) for the 5D-IV as follows:

For being a "all-rounder", this chart makes me feel that the 5D4 is the best in that class.

 

[StudentOfLight]

The ISO 12,800 shot was unfortunately also underexposed. (See attached) As can be seen in the pane on the left, my only edits were raising exposure by 2.7stops and pulling down the highlights by 100.

In fairness, the light conditions were atrocious and I don't think there was an option for him in terms of exposure settings. The chosen shutter speed was at 1/100s (entirely reasonable for the subject) and f/4 was needed for depth of field.

Yes the 5D-IV looks quite ISO invariant so if the camera was bumped up to ISO 51,200 then you'd get a very similar end result to doing it in post, the critical difference being that the EXIF data now shows a mediocre ISO 12,800, when it could have shown an impressive ISO 51,200.

 

[LSXPhotog]

A much better example of how well this camera can do in low light, check out some real world downloadable RAWS (DNG) from the 5DM IV.
I'm definitely seeing some improvement.

Here's a shot from a dimly lit studio at ISO 102k
http://froknowsphoto.com/canon-eos-5d-mark-iv-raw-files/

Thanks for sharing.

I think the camera can perform a little better than the ISO 102k shot demonstrates. (See the histogram in attach) To me it looks like the shot is about 1 stop under-exposed. To make the maximum use of the camera's dynamic range we'd need an "Expose-To-The-Right" image, perhaps taken at 1/160 f/4.5 or 1/250 f/3.5 given the available light. Hopefully we'll see in his full review.

Fro isn't exactly what I'd call a talented photographer. The fact he underexposed a low light shot with LOADS of room to work with on the shutter speed should say it all. Expose this about 2/3-stop to a full stop over and bring it back down and you'll have a downright usable shot for certain applications.

 

[Cochese]

A much better example of how well this camera can do in low light, check out some real world downloadable RAWS (DNG) from the 5DM IV.
I'm definitely seeing some improvement.

Here's a shot from a dimly lit studio at ISO 102k
http://froknowsphoto.com/canon-eos-5d-mark-iv-raw-files/

Thanks for sharing.

I think the camera can perform a little better than the ISO 102k shot demonstrates. (See the histogram in attach) To me it looks like the shot is about 1 stop under-exposed. To make the maximum use of the camera's dynamic range we'd need an "Expose-To-The-Right" image, perhaps taken at 1/160 f/4.5 or 1/250 f/3.5 given the available light. Hopefully we'll see in his full review.

Fro isn't exactly what I'd call a talented photographer. The fact he underexposed a low light shot with LOADS of room to work with on the shutter speed should say it all. Expose this about 2/3-stop to a full stop over and bring it back down and you'll have a downright usable shot for certain applications.

You're literally confusing the entire point of this "real world test."
Nobody is going to be purposefully shooting in such lighting at that ISO professionally. These are literal tests where the ISO is going. Your critique of a few snapshots who's literal purpose was to show off what the noise looks like at the extreme end are off. >.>

 

[LSXPhotog]

A much better example of how well this camera can do in low light, check out some real world downloadable RAWS (DNG) from the 5DM IV.
I'm definitely seeing some improvement.

Here's a shot from a dimly lit studio at ISO 102k
http://froknowsphoto.com/canon-eos-5d-mark-iv-raw-files/

Thanks for sharing.

I think the camera can perform a little better than the ISO 102k shot demonstrates. (See the histogram in attach) To me it looks like the shot is about 1 stop under-exposed. To make the maximum use of the camera's dynamic range we'd need an "Expose-To-The-Right" image, perhaps taken at 1/160 f/4.5 or 1/250 f/3.5 given the available light. Hopefully we'll see in his full review.

Fro isn't exactly what I'd call a talented photographer. The fact he underexposed a low light shot with LOADS of room to work with on the shutter speed should say it all. Expose this about 2/3-stop to a full stop over and bring it back down and you'll have a downright usable shot for certain applications.

You're literally confusing the entire point of this "real world test."
Nobody is going to be purposefully shooting in such lighting at that ISO professionally. These are literal tests where the ISO is going. Your critique of a few snapshots who's literal purpose was to show off what the noise looks like at the extreme end are off. >.>

In a "Real World Test" a photographer would not use the settings he used. Testing a camera by underexposing an image is poor practice. I've seen enough from that guy to see that he doesn't use common sense settings for many shooting situations and, as an educator, he shouldn't do that. I believe he posted an image the other day at 8000 ISO and a radical shutter speed like 1/4000 and his defense to his setting was "I use the highest shutter speed to eliminate the risk of camera shake and motion blur" which is asinine.

As for that lighting being so poor, that is a real world situation in itself. At a wedding, you rarely find ideal light and you need to know workaround with that. He should have shown what it looked like under and over exposed and shown the difference in the file quality in low light.

Just my personal opinion from viewing many of his videos purely from an entertainment perspective.

 

[K-amps]

The 5div images look soft compared to the other cameras, even the the 5diii

Probably has to do with initial RAW converters. The 80D was deemed soft when it was announced, and I hesitated buying it, now with mature converters/ profile, I have no such concerns. I love it

The file can take a few swipes of sharpening nicely.

 

[3kramd5]

The 5div images look soft compared to the other cameras, even the the 5diii

Probably has to do with initial RAW converters. The 80D was deemed soft when it was announced, and I hesitated buying it, now with mature converters/ profile, I have no such concerns. I love it

The file can take a few swipes of sharpening nicely.

b'am!

nice photo

 

[crashpc]

Very nice photos, but it doesn´t tell us much about the absolute amount of detail from the camera this way...