Canon R7 owners : are you happy with your high iso shots?

[KevinRA]

I have come from an A7C to the R7. I understood the compromises and the benefits, particularly in telephoto lenses size and overall weight.

However...

In my experience, I am struggling with the ISO performance a bit. I feel quite intolerant of most samples at 3200, and even some at 1600. I struggle to find anything at 6400 that I'm pleased with. I hear what everyone is saying with LR, maybe that will make a difference, but I fear it won't make a big enough difference. The A7C was easily a 6400 sensor, and 12800 was recoverable with a bit of work. With the R7 I feel like my threshold is somewhere between 1600 and 3200 is tolerable, 3200-6400 is a struggle.

It is a bit of a double whammy with the 100-400 being f8 at the long end, and the 600/800 being f11. Again, I knew the trade offs going into this, being able to fit the 18-150 and the 100-400 in a tiny Peak Design sling is fantastic, there was no chance of doing that with the Sony 200-600.

I’m pretty much right there with you when it comes to my own results. The output I get at 1600 is a bit hit and miss in terms of noise at 1600. Some look pretty decent to me and some are too noisy even when I’m careful not to underexpose. At 3200 I’m rarely pleased with the results in terms of noise and overall I don’t think I’ve had a shot at 6400 that I was really happy with in terms of noise. This was also pretty much the case for me with my 7d mk2 so I’m not really disappointed with the R7 as it seems to perform as expected and I just do my best to work in the range where I’m happy with the output. High ISO noise performance isn’t in my opinion one of this camera’s strong points but I think that can generally be said about pretty much any crop sensor camera.

Hi - I do suggest having a look with DxO Photolab Deep Prime - it does really help with these Canon sensors at the ISO 800-6400 level. That not the subscription was main reason for me leaving Lightroom (as in most other respects I much preferred it).

IMHO it does not still make ISO 25600+ any good though but YYMV.

 

[Younes B]

I’m High ISO noise performance isn’t in my opinion one of this camera’s strong points but I think that can generally be said about pretty much any crop sensor camera.

I think it is also linked to the MPs of the camera. My quite old K5iis ( 16MP ) has very decent high iso noise. Believe it is a trade off Noise / MPs.

 

[BirdShooter7]

I have come from an A7C to the R7. I understood the compromises and the benefits, particularly in telephoto lenses size and overall weight.

However...

In my experience, I am struggling with the ISO performance a bit. I feel quite intolerant of most samples at 3200, and even some at 1600. I struggle to find anything at 6400 that I'm pleased with. I hear what everyone is saying with LR, maybe that will make a difference, but I fear it won't make a big enough difference. The A7C was easily a 6400 sensor, and 12800 was recoverable with a bit of work. With the R7 I feel like my threshold is somewhere between 1600 and 3200 is tolerable, 3200-6400 is a struggle.

It is a bit of a double whammy with the 100-400 being f8 at the long end, and the 600/800 being f11. Again, I knew the trade offs going into this, being able to fit the 18-150 and the 100-400 in a tiny Peak Design sling is fantastic, there was no chance of doing that with the Sony 200-600.

I’m pretty much right there with you when it comes to my own results. The output I get at 1600 is a bit hit and miss in terms of noise at 1600. Some look pretty decent to me and some are too noisy even when I’m careful not to underexpose. At 3200 I’m rarely pleased with the results in terms of noise and overall I don’t think I’ve had a shot at 6400 that I was really happy with in terms of noise. This was also pretty much the case for me with my 7d mk2 so I’m not really disappointed with the R7 as it seems to perform as expected and I just do my best to work in the range where I’m happy with the output. High ISO noise performance isn’t in my opinion one of this camera’s strong points but I think that can generally be said about pretty much any crop sensor camera.

Hi - I do suggest having a look with DxO Photolab Deep Prime - it does really help with these Canon sensors at the ISO 800-6400 level. That not the subscription was main reason for me leaving Lightroom (as in most other respects I much preferred it).

IMHO it does not still make ISO 25600+ any good though but YYMV.

Thanks, I do have and use the DXO software. It’s good but no miracles in my experience.

 

[charlyw64]

Hi guys,

Not trying to offense anyone here but I have seen quite a few pics coming out of the R7 at ISO 800 or higher than are noisier than one would expect from a current generation sensor.

Are these outliers due to bad use, atmospheric conditions or else?

No it's the prevailing overuse of noise reduction on shots that don't need it coupled with pixel peeping and a lack of knowledge what noise really means. Noise is inevitable due to the nature of light photons and no, noise reduction software isn't a solution, it's a problem all by itself...

 

[Canon_Guy]

Hi guys,

Not trying to offense anyone here but I have seen quite a few pics coming out of the R7 at ISO 800 or higher than are noisier than one would expect from a current generation sensor.

Are these outliers due to bad use, atmospheric conditions or else?

No it's the prevailing overuse of noise reduction on shots that don't need it coupled with pixel peeping and a lack of knowledge what noise really means. Noise is inevitable due to the nature of light photons and no, noise reduction software isn't a solution, it's a problem all by itself...

+1

Thank you for your post. At least someone understands it correctly.

 

[Martin_99]

I’m High ISO noise performance isn’t in my opinion one of this camera’s strong points but I think that can generally be said about pretty much any crop sensor camera.

I think it is also linked to the MPs of the camera. My quite old K5iis ( 16MP ) has very decent high iso noise. Believe it is a trade off Noise / MPs.

Did you try to compare photos at the same size, not the same magnification? I think, that if you view them at 1:1 size, Canon probably show similar or sligthly lower noise level.

I wanted to check DPR sensor comparison tool, but unfortunately, K5IIs is not available.

 

[david31]

Check this out:

Canon R7: ISO Shootout: Testing the R7 vs. R5, 90D and 7D mark II - YouTube

 

[Steve Balcombe]

Steve Balcombe wrote:[..]

this contributes to my assertion that the R7 is not the high-end crop body that so many of us wanted. [..]

I still consider the R7 to be the R90. It's still a very good camera, it just doesn't match the expectations you get from using the 7D series.

That's it exactly.

 

[BirdShooter7]

Steve Balcombe wrote:[..]

this contributes to my assertion that the R7 is not the high-end crop body that so many of us wanted. [..]

I still consider the R7 to be the R90. It's still a very good camera, it just doesn't match the expectations you get from using the 7D series.

That's it exactly.

 

[KevinRA]

Check this out:

Canon R7: ISO Shootout: Testing the R7 vs. R5, 90D and 7D mark II - YouTube

Nice - really shows at 6400 and 12800 how good the R7 is.

 

[Alastair Norcross]

Check this out:

Canon R7: ISO Shootout: Testing the R7 vs. R5, 90D and 7D mark II - YouTube

Nice - really shows at 6400 and 12800 how good the R7 is.

And that video was made before DXO Photolab supported the R7. It's even better now, because of that.

 

[Steve Balcombe]

Steve Balcombe wrote:[..]

this contributes to my assertion that the R7 is not the high-end crop body that so many of us wanted. [..]

I still consider the R7 to be the R90. It's still a very good camera, it just doesn't match the expectations you get from using the 7D series.

That's it exactly.

Yea I wanted higher end as well but that’s not the product that was brought to market. I think I understand why Canon did what they did with the R7. I could be upset about the name but I don’t personally see the point.

Oh, I agree and I'm not "upset about the name" as such, they can call it anything they want. What bothers me is the attitude behind it - "we'll sell more of these if we call it the R7 and imply it's more than it actually is". Such a shame after the R5 and R6 were such worthy successors in the '5' and '6' categories.

The funny thing is, the price is spot on for a 90D successor, so I guess I should be grateful for that

The R7 is what it is and I’m finding that I’m personally doing quite well with it and overall it is a solidly better camera than the old 7d mk2. But yes, I would have been happy to pay more for a higher end RF mount APS-c camera and R7 probably would have been a better name for that camera than the R7 we got.

Yep.

Bottom line, I’m going to enjoy my R7 until something better for me comes along.

Me too.

 

[BirdShooter7]

Check this out:

Canon R7: ISO Shootout: Testing the R7 vs. R5, 90D and 7D mark II - YouTube

 

[RDM5546]

The difference between a 1.6x crop body and a full frame body should be about 1 1/3 stops, all other things being equal, just because the bigger sensor collects 1 1/3 stops more light. I would expect the 1.6X crop sensor to have slightly lower quantum efficiency which would mean a possible difference of 1.5 stops more light collected by the lower density/bigger pixels in the FF sensor.

Still there are other factors such as the sensor amplifier gain adjustments at different ISOs and the effects of parasitic circuit noises which will be a factor. Despite these theoretical considerations the actual results with R7 IMHO seem to be better than this 1.3-1.5 stops and in several cases I checked closer to 1 stop difference between the R7 and R5. In some cases it seemed to be less than 1 stop difference in noise.

Remember where I was coming from - I was explaining to the OP that a difference of 1 1/3 stops is to be expected and not an issue with the R7 specifically. If the actual difference is slightly more or slightly less, it doesn't alter the key point.

The R5 is better for low noise but R7 noise did not disappointment me and in fact seems better than expected. The precise scene dynamic range for the areas of interest and the exposure are playing a role in this comparison.

The R7 noise is within expectations and as such it doesn't disappoint me either, in the sense that I knew what I was buying. It's just frustrating to know that a stacked BSI sensor could be better, and this contributes to my assertion that the R7 is not the high-end crop body that so many of us wanted.

Personally I like to use dynamic range rather than noise when making comparisons. Noise can be reduced very successfully in post, but dynamic range can't be added if it wasn't there, and thanks to Bill Claff we have ready access to some excellent and well-presented data. The Fujifilm X-H2S is about half a stop better than the R7 at ISO speeds above 400. On the other hand the R7's maximum dynamic range (at ISO 100) is about half a stop better than the X-H2S's (which is at ISO 160). So the X-H2S doesn't have a night-and-day lead, but more importantly it gives us a much faster readout without any IQ penalty.

The OM-1's dynamic range is almost identical to the R7 at all but the lowest ISO speeds, so it would appear that the stacked BSI sensor is compensating for the smaller sensor size.

I agree with what you say. Of course a BSI sensor is better but likely lower manufacturing yield and it may take more time to put into production. The $1500 price along with the performance of the R7 at that price is driving huge sales right with the R7 being a very good product. A more rugged, 7D2 like, design could have been made and it would likely cost over 2000 and take more development time. Maybe the $2200 model is in development and still coming. Canon is flat rushing to transition to mirrorless and time is of the essence. I was shocked to get the 12fps mechanical shutter and 30 fps ES with R3 like focus in a $1500 APS-C camera. Not everyone want the same thing but the R7 does seem to connect with lots of buyers out there. I bought one even though I already had the R5.

 

[Alastair Norcross]

The difference between a 1.6x crop body and a full frame body should be about 1 1/3 stops, all other things being equal, just because the bigger sensor collects 1 1/3 stops more light. I would expect the 1.6X crop sensor to have slightly lower quantum efficiency which would mean a possible difference of 1.5 stops more light collected by the lower density/bigger pixels in the FF sensor.

Still there are other factors such as the sensor amplifier gain adjustments at different ISOs and the effects of parasitic circuit noises which will be a factor. Despite these theoretical considerations the actual results with R7 IMHO seem to be better than this 1.3-1.5 stops and in several cases I checked closer to 1 stop difference between the R7 and R5. In some cases it seemed to be less than 1 stop difference in noise.

Remember where I was coming from - I was explaining to the OP that a difference of 1 1/3 stops is to be expected and not an issue with the R7 specifically. If the actual difference is slightly more or slightly less, it doesn't alter the key point.

The R5 is better for low noise but R7 noise did not disappointment me and in fact seems better than expected. The precise scene dynamic range for the areas of interest and the exposure are playing a role in this comparison.

The R7 noise is within expectations and as such it doesn't disappoint me either, in the sense that I knew what I was buying. It's just frustrating to know that a stacked BSI sensor could be better, and this contributes to my assertion that the R7 is not the high-end crop body that so many of us wanted.

Personally I like to use dynamic range rather than noise when making comparisons. Noise can be reduced very successfully in post, but dynamic range can't be added if it wasn't there, and thanks to Bill Claff we have ready access to some excellent and well-presented data. The Fujifilm X-H2S is about half a stop better than the R7 at ISO speeds above 400. On the other hand the R7's maximum dynamic range (at ISO 100) is about half a stop better than the X-H2S's (which is at ISO 160). So the X-H2S doesn't have a night-and-day lead, but more importantly it gives us a much faster readout without any IQ penalty.

The OM-1's dynamic range is almost identical to the R7 at all but the lowest ISO speeds, so it would appear that the stacked BSI sensor is compensating for the smaller sensor size.

I agree with what you say. Of course a BSI sensor is better but likely lower manufacturing yield and it may take more time to put into production. The $1500 price along with the performance of the R7 at that price is driving huge sales right with the R7 being a very good product. A more rugged, 7D2 like, design could have been made and it would likely cost over 2000 and take more development time. Maybe the $2200 model is in development and still coming. Canon is flat rushing to transition to mirrorless and time is of the essence. I was shocked to get the 12fps mechanical shutter

You'll be even more shocked when you discover it's actually 15fps, which is faster than the mechanical shutter burst rate in the R3/5/6.

and 30 fps ES with R3 like focus in a $1500 APS-C camera. Not everyone want the same thing but the R7 does seem to connect with lots of buyers out there. I bought one even though I already had the R5.

 

[Bramble9]

YounesB; That's really a nice set of concert photos and what really grabbed me about them was that they were done w/ the K5II! I used to have a pair of K5IIs years ago and they were considered about the best at high ISO for an APS-C sensor. Anyway, really nice shots. Thanks for the memories.

Bramble9

 

[Younes B]

YounesB; That's really a nice set of concert photos and what really grabbed me about them was that they were done w/ the K5II! I used to have a pair of K5IIs years ago and they were considered about the best at high ISO for an APS-C sensor. Anyway, really nice shots. Thanks for the memories.

Bramble9

Thanks Bramble9.

I still use mine...precisely because of the overall IQ including in high ISO.

But these new offerings from Canon are itching , given the eye AF and the impressive AF performance.

I may run dual system at some point

 

[R2D2]

I have come from an A7C to the R7. I understood the compromises and the benefits, particularly in telephoto lenses size and overall weight.

However...

In my experience, I am struggling with the ISO performance a bit. I feel quite intolerant of most samples at 3200, and even some at 1600. I struggle to find anything at 6400 that I'm pleased with. I hear what everyone is saying with LR, maybe that will make a difference, but I fear it won't make a big enough difference. The A7C was easily a 6400 sensor, and 12800 was recoverable with a bit of work. With the R7 I feel like my threshold is somewhere between 1600 and 3200 is tolerable, 3200-6400 is a struggle.

It is a bit of a double whammy with the 100-400 being f8 at the long end, and the 600/800 being f11. Again, I knew the trade offs going into this, being able to fit the 18-150 and the 100-400 in a tiny Peak Design sling is fantastic, there was no chance of doing that with the Sony 200-600.

I’m pretty much right there with you when it comes to my own results. The output I get at 1600 is a bit hit and miss in terms of noise at 1600. Some look pretty decent to me and some are too noisy even when I’m careful not to underexpose. At 3200 I’m rarely pleased with the results in terms of noise and overall I don’t think I’ve had a shot at 6400 that I was really happy with in terms of noise. This was also pretty much the case for me with my 7d mk2 so I’m not really disappointed with the R7 as it seems to perform as expected and I just do my best to work in the range where I’m happy with the output. High ISO noise performance isn’t in my opinion one of this camera’s strong points but I think that can generally be said about pretty much any crop sensor camera.

Hi - I do suggest having a look with DxO Photolab Deep Prime - it does really help with these Canon sensors at the ISO 800-6400 level.

Agree 100%. Although I don’t think it’ll help some of the shooters here because they only shoots jpegs, and/or use other software.

There’s a huge paradigm shift that’s being driven by these new technologies, both in the equipment and the software. Heck, not too many years ago AF at beyond f/5.6 was barely available! Now new lenses are being produced that aren’t even designed to reach to f/5.6!

Being on the wrong side of the shift will certainly bring up these old complaints. All we can do is make suggestions. ;-)

R2

 

[R2D2]

YounesB; That's really a nice set of concert photos and what really grabbed me about them was that they were done w/ the K5II! I used to have a pair of K5IIs years ago and they were considered about the best at high ISO for an APS-C sensor. Anyway, really nice shots. Thanks for the memories.

Bramble9

Thanks Bramble9.

I still use mine...precisely because of the overall IQ including in high ISO.

But these new offerings from Canon are itching , given the eye AF and the impressive AF performance.

I may run dual system at some point

Definitely! All of this AF goodness with these new bodies actually changes the way you shoot! Highly recommend giving it a try! :-D

R2

 

[Steve Balcombe]

Hi - I do suggest having a look with DxO Photolab Deep Prime - it does really help with these Canon sensors at the ISO 800-6400 level.

Agree 100%. Although I don’t think it’ll help some of the shooters here because they only shoots jpegs, and/or use other software.

There’s a huge paradigm shift that’s being driven by these new technologies, both in the equipment and the software. Heck, not too many years ago AF at beyond f/5.6 was barely available! Now new lenses are being produced that aren’t even designed to reach to f/5.6!

Being on the wrong side of the shift will certainly bring up these old complaints. All we can do is make suggestions. ;-)

This is all true, and one huge improvement in the software is that noise reduction on high ISO images no longer creates a smeary mess. And the days of doing different noise reduction on subject and background are largely gone which is a great time-saver.

But you still have to be alert to problems. One I run into quite often is that Topaz DeNoise AI, so smart in so many ways, sees the compound eyes in many insects as noise and smooths them out. I have to use the history brush in Photoshop to restore the original version.