Canon 7D: Slow Continuous Shooting Speed?

[prodigy2k7]

page 93 in manual:
In Low-Light Areas or indoors, the continuous shooting speed may become slower even if a fast shutter speed is set.

Anyone know why this is?

 

[tyler_h]

The time it takes to adjust focus between each shot slows the shooting speed.

 

[Overread]

The general argument is processing of noise performed on the shots right after they are taken (as part of the process even for RAW photos). However the whole fact that its FPS slows in darker conditions is something of a quandary that many have wondered about. The additional processing is about the only thing people can work out as the reason since even if you have the camera in full manual mode with manual focusing the FPS still slows in darker shooting conditions (even if your exposures are not at a high ISO and correctly exposed - eg with flash support lighting).

 

[Derrel]

"It's a feature, not a bug!!!" quote attributed to Microsoft, and now to Canon...

I think overread is correct. The images coming off the sensor are fairly noisy...you can see how bad they are when you open full-resolution 7D JPEG files using a number of different image viewing applications....BOATLOADS of noise shows up as the image is first opened, with loads of white speckles and other assorted bits of noise all over the image field, and then the Noise Reduction algorithm that has been applied to the image in processing, is then applied to the image in the image viewer software, and the noise is suddenly sublimated, and hidden. It's kind of weird, opening up Canon 7D files...I've never seen such noisy images before that camera hit the market.

 

[Overread]

*thinks its 7D does pretty good with noise and is certainly rather nice compared to his 400D*

also Derral how are you able to open up the RAWs before Canons built in noise reduction hits them? As I understand it, the effect is instant and part of the RAW itself, not a preset value applied to the JPEG only and then applied in select RAW conversion software as a default starting point for RAW editing.

I agree that the 7D did get some big flack for not being 5DM2 good with noise (which seems to be the major noise complaint about the 7D from many) and that, for some reason, early RAW processing/reading seemed to give more noisey effects in some cases - however that appears to have been cleaned up.

I've also noted that some people will harp on about noise with cameras (not just the 7D, its just the current victim) when they are shooting with a film mindset rather than a digital mindset (ie not exposing to the right when given the opportunity).

 

[Derrel]

Overrevved,

Here are two low-light 7D images, both horizontally framed, which are available for download. First, an ISO 3200 shot at http://www.robgalbraith.com/public_files/Canon_EOS_7D_ISO3200_Reggae.jpg and an ISO 6,400 shot at http://www.robgalbraith.com/public_files/Canon_EOS_7D_ISO6400_Reggae.jpg


Both were shot as .CR2 images, and Noise Reduction was done with Canon's own DPP 3.7 software. When I open these images using a QuickTime-based image viewer, the UNDERLYING, NOISY,SPECKLED image is visible. After about 1 second, the Canon DPP Noise Reduction mask is applied, and the images takes its finalized form. The images as-shot and as-developed in DPP, without NR applied, are quite noisy. I do not own the camera, but have shot it, and assisted as a lighting tech on a commercial shoot where the photographer and I noticed his 7D's noise was rearing its head at ISO 160, under studio lighting...he was used to the 30D and 40D, and was disappointed that at ISO 200, the images looked far too noisy for his taste. Maybe you misunderstood what my original post was meant to convey: that in low light, and at higher ISO settings, the 7D's images right off the sensor are noisy; so noisy that the camera's touted frame rate drops precipitously due to the intensive need for noise reduction on the captured data.

Once again,not all image viewer software will allow you to see past the NR application parameters when viewing JPEGs made from Canon CR2 files using DPP 3.7; my system has software that allows me to see the data before NR, and after the NR instructions are in place. I do not own a 7D, so I'm not worried about discussing its weaknesses openly. I have a Nikon D2x--it too, has a lot of issues with noise. I'm familiar with noise from too many pixels with technology that's not fully capable of dealing with high pixel density.

 

[Drake]

Overrevved,

Here are two low-light 7D images, both horizontally framed, which are available for download. First, an ISO 3200 shot at http://www.robgalbraith.com/public_files/Canon_EOS_7D_ISO3200_Reggae.jpg and an ISO 6,400 shot at http://www.robgalbraith.com/public_files/Canon_EOS_7D_ISO6400_Reggae.jpg


Both were shot as .CR2 images, and Noise Reduction was done with Canon's own DPP 3.7 software. When I open these images using a QuickTime-based image viewer, the UNDERLYING, NOISY,SPECKLED image is visible. After about 1 second, the Canon DPP Noise Reduction mask is applied, and the images takes its finalized form. The images as-shot and as-developed in DPP, without NR applied, are quite noisy. I do not own the camera, but have shot it, and assisted as a lighting tech on a commercial shoot where the photographer and I noticed his 7D's noise was rearing its head at ISO 160, under studio lighting...he was used to the 30D and 40D, and was disappointed that at ISO 200, the images looked far too noisy for his taste. Maybe you misunderstood what my original post was meant to convey: that in low light, and at higher ISO settings, the 7D's images right off the sensor are noisy; so noisy that the camera's touted frame rate drops precipitously due to the intensive need for noise reduction on the captured data.

Once again,not all image viewer software will allow you to see past the NR application parameters when viewing JPEGs made from Canon CR2 files using DPP 3.7; my system has software that allows me to see the data before NR, and after the NR instructions are in place. I do not own a 7D, so I'm not worried about discussing its weaknesses openly. I have a Nikon D2x--it too, has a lot of issues with noise. I'm familiar with noise from too many pixels with technology that's not fully capable of dealing with high pixel density.

I still don't think the image you see for a brief second is actually a photo before Canon's in-camera NR. How I see it is that the camera takes the raw data, applies the noise reduction algorithms (also in case of many entry level cameras sharpens) it and saves the outcome as a CR2 file. And by outcome I mean that the image, pretty much like a JPEG cannot be brought back to the pre-NRed state. Of course in JPEG mode the camera goes further with WB etc, but that's not the case here. Anyway, after that we load the raw photos to a chosen application and it also applies it's own NR on top of it, and that's pretty much the reason why the same photo has different noise characteristics depending on the software used - DPP, ACR or anything else. I think the noisy image you see for a second is actually the middle state - the actual RAW file before the second NR, but after the in-camera one.

 

[Robin Usagani]

Well maybe they are just saying in low light and shoot with AV for example, your shutter speed needs to be 1/30 sec. So that means the fastest burst you can have is just barely over 3 fps.

 

[Overread]

Derral - the noise reduction I was referring was occurring to the RAWs themselves as part of the processing of them inside the camera. I'm not referring to the additional noise reduction applied to the JPEGs made inside the camera (and also embedded into RAWs) which is what your software is showing you when the DPP mask slips in when it reads the cameras scene settings and uses them for the RAw processing default.

The ISO 160 being more noisy is slightly understandable as its not a real ISO value, but a simulated one in the camera (it either under or over exposes (I forget which, but I think its under) and then uses software to bump/lower the lighting values to the simulated ISO.
I also know a fair few who will compare a 100% view of the 18MP shots to 100% views of 15 or even 12 MP shots and state that they are more noisy on the 18MP sensor. Which is true at 100% view, but greatly in part because the view at that scale is much larger than on the smaller MP sensors. Once resized down to the same dimensions (eg preparing for prints or webposting) the noise is far less apparent and, from what I've read, lower than in previous crop sensor bodies (at least models below the 7D - so the 50D, etc..).

Unfortunatly I've no real data to compare the really high ISO values against - 7D is the best I have (have ever had) so its hard for me to be totally objective of its performance at this level. I do think that Canon could have done better with fewer MP and given a much bigger ISO gain - however I also think there a lot of reports about the 7Ds noise levels which are based on sketchy reports or unfair comparisons.


Schwetty - nope its an actual reduction - manual mode and manual focus and you'll still get the slowdown in darker shooting conditions even if the shutter speed is fast enough to allow its fps to be normal. I'm not 100% sure where the slowdown comes from - its either direct noise processing in the RAW itself taking longer and/or processing on the embedded JPEGs.

 

[KmH]

The continuous shooting speed of any DSLR is highly dependent on what camera settings are used.

 

[prodigy2k7]

The continuous shooting speed of any DSLR is highly dependent on what camera settings are used.

Yes, but as Overread said, manual setting was used 1/8000 of a second at F/2.8 with max FPS setting, manual focus, etc etc... I hold down the shutter, move towards the light and would get 8 FPS, move into darkness and would get less...