its all about the 5DS, who is using the 5DR

[East Wind Photography]

5DS ... only. Everybody says it is great...

really? I have a friend who has it and she keeps whining about the supposedly terrible noise in higher ISO ranges, going so far as to label it unusable beyond ISO 1600. she said it was much worse than her 5D MKIII.

At a per pixel level, you are correct w.r.t. noise. 5D3 outperforms it.

So in good light / in a studio / on a tripod, the 5DS rigs deliver awesome detail and you get what you paid for.

But in poor light handheld, you may need to burn those pixels down to fight noise. So in that light, a 50 MP 5DS shot downsampled to 22MP compared against the 5D3 is -- depending on who you ask -- about a dead heat.

Spinning it another way, you could say the 5DS is a 5D3, but with a special feature: 2.5X the resolution in lower ISO.

I don't own a 5DS, but that's what various reviewers have shared when studying 5DS noise. I welcome this thread's members who have shot both to corroborate or rebut that notion.

- A

So the 5d3 does have larger pixels and therefore can capture more light compared to the background noise. Downsampling the 5ds will not change that. It will reduce the fine detail captured. In essence you cant create something that isnt there in the first place. The only way to get better signal to noise with smaller pixels is to reduce the noise floor. I dont believe they acheived that with the 50mp sensor.

In fact the pixels size on the 5ds is the same as the 7d2. Everything else being the same, both should have similar signal to noise in raw data.

That really makes no sense at all because that's not the way it works. It's the SENSOR size that counts. Yes the 5Ds and 7D2 have the same pixel size, but the 5Ds has a larger sensor. So the noise characteristic is better. You can't say "everything else being the same" in your statement but then have different sensor sizes, because everything ISN'T the same. The most important factor regarding noise isn't the same.

Well only if you fill the frame with your subject on each sensor. If you select aps-c crop mode on the 5ds and back up you should see the same level of noise as your subject is now spread across the same area as in the 7d2. The only real benefit you get over the 7d2 is more pixels.

So full disclosure here. I own the 5dsr and woukd not have bought it if i didnt think it was better. In fact i traded in my 5d3 to get it...no regrets.

I'll see if i can run some high iso tests with the 7d2 and 5dsr in aps-c crop mode using the same subject and see if there is any noticable difference.

Oh you have both? Cool.

i need the 7d2 for sports. The 5dsr is nice but at 5 fps its difficult to get the timing right.

 

[East Wind Photography]

I just ran through some quick high iso tests and discovered somethig that i didnt know. The high iso noise reduction setting does in fact apply to RAW files, not just jpg. All of my tests were invalid due to inconsistent settings. So back to the drawing board...but for another day.

RAW does not seem to always be RAW.

 

[takesome1]

I just ran through some quick high iso tests and discovered somethig that i didnt know. The high iso noise reduction setting does in fact apply to RAW files, not just jpg. All of my tests were invalid due to inconsistent settings. So back to the drawing board...but for another day.

RAW does not seem to always be RAW.

Can you elaborate on this? You seeing it in DPP? LR?

I have had mine about two months now. Here is something I have noticed but yet to test to see if my suspensions are correct, in LR the 5Ds R you can apply a huge amount of noise reduction. Of course this comes at the loss of resolution, but it appears to me just from observing the pictures I have been pp that if you down sample to the same file size as the 5D III it is never worse than the 5D III image. I have seen a few others mention this but haven't tested it out myself yet to see how much of an ISO advantage if any the 5D III may have over the 5Ds R. My suspension is that it doesn't after PP.

 

[wockawocka]

I noticed with the crop mode overlay in camera these are carried over into camera raw so I'd imaging other settings are too.

 

[East Wind Photography]

I just ran through some quick high iso tests and discovered somethig that i didnt know. The high iso noise reduction setting does in fact apply to RAW files, not just jpg. All of my tests were invalid due to inconsistent settings. So back to the drawing board...but for another day.

RAW does not seem to always be RAW.

Can you elaborate on this? You seeing it in DPP? LR?

I have had mine about two months now. Here is something I have noticed but yet to test to see if my suspensions are correct, in LR the 5Ds R you can apply a huge amount of noise reduction. Of course this comes at the loss of resolution, but it appears to me just from observing the pictures I have been pp that if you down sample to the same file size as the 5D III it is never worse than the 5D III image. I have seen a few others mention this but haven't tested it out myself yet to see how much of an ISO advantage if any the 5D III may have over the 5Ds R. My suspension is that it doesn't after PP.

No actually i was just viewing the images with microsoft image viewer. After seeing a major difference in noise, i retook a couple of shots in raw with the setting to the other extreme and the noise was removed from the raw file. Tested on both cameras and the same. So i need to decide on what setting to use as a baseline. I assume no high iso noise reduction would be best for testing but maybe in practicality it would be better to use some. Would need to compare how well in camera and PP noise filtering work independently....and perhaps even together.

 

[East Wind Photography]

I noticed with the crop mode overlay in camera these are carried over into camera raw so I'd imaging other settings are too.

Yes actually i think the crop mode setting just encodes it ito exif data. I Dont believe it actually excludes areas outside the frame. If your pp app reads it it should crop. If it doesnt then you need to redo the crop. Seems like its cheating a bit.

 

[bdunbar79]

That's really interesting.

 

[East Wind Photography]

I re-ran my shots and the results are interesting and not quite what I expected. The comparisons here show the 5DSR on the left and 7D2 on the right. All were shot using my 85mm F1.2L II lens under a single 60watt CF bulb about 20 ft away at f 1.2. Shots were taken at 3200, 6400, 12800, 12800 -2 stops, and 12800 with in camera NR set to high. All were shot in RAW and snipped using s simple microsoft image viewer with no chance of any external manipulation or hidden settings such as what got me on my first go around with in camera settings. All images were shot from the same distance and cropped to show the same area at about 100%.

Amazingly to me the 7d2 seems to fair a bit better until we use the in camera Hi ISO Noise reduction and this seems to be where things change in favor slightly toward the 5dsR. Both use the Digic 6 so no exactly sure whats going on there...but its for the better.

Edit - Seems that the 5dsr also had AWB W set which accounts for the color balance difference.

 

[takesome1]

I re-ran my shots and the results are interesting and not quite what I expected. The comparisons here show the 5DSR on the left and 7D2 on the right. All were shot using my 85mm F1.2L II lens under a single 60watt CF bulb about 20 ft away at f 1.2. Shots were taken at 3200, 6400, 12800, 12800 -2 stops, and 12800 with in camera NR set to high. All were shot in RAW and snipped using s simple microsoft image viewer with no chance of any external manipulation or hidden settings such as what got me on my first go around with in camera settings. All images were shot from the same distance and cropped to show the same area at about 100%.

Amazingly to me the 7d2 seems to fair a bit better until we use the in camera Hi ISO Noise reduction and this seems to be where things change in favor slightly toward the 5dsR. Both use the Digic 6 so no exactly sure whats going on there...but its for the better.

Edit - Seems that the 5dsr also had AWB W set which accounts for the color balance difference.

Interesting.

I would be interested in seeing the side by side shot in JPEG.
The reason is I think they would both be equally processed and it might give a better representation of the results we could expect in our work flow.
I know from shooting both that there is a fair amount of noise correction that happens with both in JPEG. I think the 5Ds R can handle more correction than the 7D II.

 

[East Wind Photography]

I re-ran my shots and the results are interesting and not quite what I expected. The comparisons here show the 5DSR on the left and 7D2 on the right. All were shot using my 85mm F1.2L II lens under a single 60watt CF bulb about 20 ft away at f 1.2. Shots were taken at 3200, 6400, 12800, 12800 -2 stops, and 12800 with in camera NR set to high. All were shot in RAW and snipped using s simple microsoft image viewer with no chance of any external manipulation or hidden settings such as what got me on my first go around with in camera settings. All images were shot from the same distance and cropped to show the same area at about 100%.

Amazingly to me the 7d2 seems to fair a bit better until we use the in camera Hi ISO Noise reduction and this seems to be where things change in favor slightly toward the 5dsR. Both use the Digic 6 so no exactly sure whats going on there...but its for the better.

Edit - Seems that the 5dsr also had AWB W set which accounts for the color balance difference.

Interesting.

I would be interested in seeing the side by side shot in JPEG.
The reason is I think they would both be equally processed and it might give a better representation of the results we could expect in our work flow.
I know from shooting both that there is a fair amount of noise correction that happens with both in JPEG. I think the 5Ds R can handle more correction than the 7D II.

Im pretty sure thats the case. You can tell from the 12800 comparisons that the noise on the 7d2 is blockier and has quite a few solid artifacts. On the 5dsr it is more random. These are jpgs but AFTER they were enlarged on screen and snipped with the snippet tool. Any jpg effects here would be negligible.

I only create jpgs in my workflow as the last process before publication. I am more interested in wether in camera NR is better or worse than LR or DxO since it can be done to raw images in camera.

 

[privatebydesign]

Your assertion that "no chance of any external manipulation or hidden settings such as what got me on my first go around with in camera settings" is invalid, the WB settings difference proves that.

The only way you can view native RAW files is in something like dcraw, short of that you have to know, definitively, which tags any individual viewer will honor.

 

[East Wind Photography]

Your assertion that "no chance of any external manipulation or hidden settings such as what got me on my first go around with in camera settings" is invalid, the WB settings difference proves that.

The only way you can view native RAW files is in something like dcraw, short of that you have to know, definitively, which tags any individual viewer will honor.

Well i agree on the viewer part. However the WB setting was internal, not external and was my fault for not checking that. Still i dont think it has much affect on the noise reduction....but who knows what the digic is doing.

 

[midluk]

Are you sure your image viewer does not just show the embedded jpeg preview of the raw image?

 

[East Wind Photography]

Are you sure your image viewer does not just show the embedded jpeg preview of the raw image?

I dont believe microsoft image viewer is that sophisticated. Though i could be wrong. ive never had ms image viewer apply noise reduction on any image before. Its not an editor and has no provisions for even adjusting image brightness.

 

[takesome1]

I re-ran my shots and the results are interesting and not quite what I expected. The comparisons here show the 5DSR on the left and 7D2 on the right. All were shot using my 85mm F1.2L II lens under a single 60watt CF bulb about 20 ft away at f 1.2. Shots were taken at 3200, 6400, 12800, 12800 -2 stops, and 12800 with in camera NR set to high. All were shot in RAW and snipped using s simple microsoft image viewer with no chance of any external manipulation or hidden settings such as what got me on my first go around with in camera settings. All images were shot from the same distance and cropped to show the same area at about 100%.

Amazingly to me the 7d2 seems to fair a bit better until we use the in camera Hi ISO Noise reduction and this seems to be where things change in favor slightly toward the 5dsR. Both use the Digic 6 so no exactly sure whats going on there...but its for the better.

Edit - Seems that the 5dsr also had AWB W set which accounts for the color balance difference.

Interesting.

I would be interested in seeing the side by side shot in JPEG.
The reason is I think they would both be equally processed and it might give a better representation of the results we could expect in our work flow.
I know from shooting both that there is a fair amount of noise correction that happens with both in JPEG. I think the 5Ds R can handle more correction than the 7D II.

Im pretty sure thats the case. You can tell from the 12800 comparisons that the noise on the 7d2 is blockier and has quite a few solid artifacts. On the 5dsr it is more random. These are jpgs but AFTER they were enlarged on screen and snipped with the snippet tool. Any jpg effects here would be negligible.

I only create jpgs in my workflow as the last process before publication. I am more interested in wether in camera NR is better or worse than LR or DxO since it can be done to raw images in camera.

What I wonder is how equal the comparisons are in RAW, with firm ware of the two different bodies. The viewer may be reading each differently because of differences Canon has put in the programming. I don't think it would matter how you made the JPG, in camera or other, as long as both are PP to the fullest extent possible. Then compare the two to see what finished product you could get.

But;

To me it probably wouldn't mater to me. I have both bodies and I am not going to use the 7D II unless I need its unique functions for a specific job. Any other time I would be using the 5Ds R.

 

[ahsanford]

With all this noise discussion, why aren't we using TDP's super handy noise comparison charts, which includes turning/off all form of noise reduction, scaling a 5DS down to 5D3 levels, etc.?

http://www.the-digital-picture.com/Reviews/Camera-Noise.aspx?Camera=980&Test=0&ISO=100&CameraComp=963&TestComp=0&ISOComp=0

Look at the pull-down under 'Standard' and tweak what you like. Have at it, yo.

Bryan Carnathan (who runs that site) is also fantastic about responding to technical questions about his methods, exactly how he processed things, etc. Cut him an e-mail if you have questions.

- A

 

[privatebydesign]

Your assertion that "no chance of any external manipulation or hidden settings such as what got me on my first go around with in camera settings" is invalid, the WB settings difference proves that.

The only way you can view native RAW files is in something like dcraw, short of that you have to know, definitively, which tags any individual viewer will honor.

Well i agree on the viewer part. However the WB setting was internal, not external and was my fault for not checking that. Still i dont think it has much affect on the noise reduction....but who knows what the digic is doing.

WB is not "internal" to a RAW file, it is a tag attached in the embedded sidecar file, if a viewer is honouring WB information it is reading that sidecar file, who knows if it is honoring NR tags and or reading the embedded jpeg?

In truth the only relevant comparison is a fully and optimally post processed file vs another optimally processed RAW file both to your personal specific end uses, and coming from different cameras would imply different processing for each to achieve optimal. To be sure 'RAW' is pretty meaningless, it is only the end result we can achieve with that data that is relevant.

 

[East Wind Photography]

Your assertion that "no chance of any external manipulation or hidden settings such as what got me on my first go around with in camera settings" is invalid, the WB settings difference proves that.

The only way you can view native RAW files is in something like dcraw, short of that you have to know, definitively, which tags any individual viewer will honor.

Well i agree on the viewer part. However the WB setting was internal, not external and was my fault for not checking that. Still i dont think it has much affect on the noise reduction....but who knows what the digic is doing.

WB is not "internal" to a RAW file, it is a tag attached in the embedded sidecar file, if a viewer is honouring WB information it is reading that sidecar file, who knows if it is honoring NR tags and or reading the embedded jpeg?

In truth the only relevant comparison is a fully and optimally post processed file vs another optimally processed RAW file both to your personal specific end uses, and coming from different cameras would imply different processing for each to achieve optimal. To be sure 'RAW' is pretty meaningless, it is only the end result we can achieve with that data that is relevant.

Ok so i can buy the concept that what i may be seeing is an embedded jpg image. So in that case the comparisons here are still reasonable, just not pure raw. Only the one pair had NR applied, the rest had it disabled. The original comparison was that the 5dsr would have less noise because it is a full frame sensor and the attempt was to show that the same subject at the same distance on the 7d2 would have similar noise levels as the 5dsr. Both sensors share the same pixel size. From what i have been seeing, the noise is not less on the 5dsr. The pattern is different and perhaps easier to reduce algorythmically.

If you were to change your distance such that the subject filled the frame on the 5dsr, yes the noise would be smaller and less visible. I actually wish instillhad my 5diii to compare as well. It shoud fair a stop or two better due to its larger pixel size.

Anyway, in end as you say, its the finished product that matters the most, no matter how you get there.