Canon EOS 7D Mark II Spec List [CR2]

[rs]

I'd disagree that a lower resolution 7D II would be better. ISO performance can be improved in other ways besides larger photodiode area these days. Only when we hit the maximum achievable Q.E. will we have no other option but to use larger photodiodes. Reach is the key benefit of the 7D line...and reach has to do with pixel density, not sensor size. A 24.1mp APS-C 7D II would bring in the buyers more than anything else, as that is what the 7D line is all about...reach with sports-level performance.

The 7D is also famed for its frame rate and buffer depth. If the MP is kept in check, there's no reason why Canon can't push the boundaries and make that small mirror and shutter move much faster in the mk II. But if they increase it to 24 MP, it's likely to remain at about 8 FPS.

[jrista]

I'd disagree that a lower resolution 7D II would be better. ISO performance can be improved in other ways besides larger photodiode area these days. Only when we hit the maximum achievable Q.E. will we have no other option but to use larger photodiodes. Reach is the key benefit of the 7D line...and reach has to do with pixel density, not sensor size. A 24.1mp APS-C 7D II would bring in the buyers more than anything else, as that is what the 7D line is all about...reach with sports-level performance.

The 7D is also famed for its frame rate and buffer depth. If the MP is kept in check, there's no reason why Canon can't push the boundaries and make that small mirror and shutter move much faster in the mk II. But if they increase it to 24 MP, it's likely to remain at about 8 FPS.

Well, I just ran these calculations the other day. People have gravely miscalculated the data throughput rate of the 1D X at 144-165mb/s per DIGIC 5+ chip. Most of the calculations involve an 18mp image size, and do not account for metadata or other overhead. Additionally, not all throughput calculations people do factor in the 14fps frame rate of the 1D X (which it can do with mirror lockup.) The "real" pixel count of the 1D X is 19.3mp, as Canon masks off a border of pixels for calibration purposes, and those pixels ARE included in the RAW image when it is saved to the CF card. Accounting for a higher "real" RAW image pixel count, 14fps frame rate, and a buffer for any overhead:

Code: [Select]

19,300,000pixels * 14bit / 8bit/byte * 14fps = 473mb/s
Assuming there is some overhead, and the need for a little bit of leeway for metadata for each image, and the need to account for performing compression and the like, I'd say the total throughput of the 1D X is at least 500mb/s. That is 250mb/s per DIGIC 5+ dsp chip. Now, if we run a similar calculation for the 7D II assuming a 24.1mp image size, and a similar 7% additional pixels for the masked border pixels, we have a "real" 7D II pixel count of 25.8mp (25,787,000 pixels). Running the same calculation, only for 10fps instead of 14fps:

Code: [Select]

25,787,000 pixels * 14bit / 8bit/byte * 10fps = 452mb/s
Even with a considerably higher pixel count, the 7D II with a pair of DIGIC 5+ processors should be able to handle 10fps no problem. That would actually be a LOWER data throughput rate than the 1D X at 14fps! We can run the numbers for 11fps, too:

Code: [Select]

25,787,000 pixels * 14bit / 8bit/byte * 11fps = 497mb/s
Even at 11fps, we are still below the 500mb/s total that would be allowed if each DIGIC 5+ chip could process at a throughput rate of 250mb/s (which, to me, seems like a more plausible throughput rate than 165mb/s that you get at 18mp and 12fps). If you factor in the facts that the DIGIC 5+ processors have to not only perform ADC on the incoming sensor pixel data, but also compress and write the output pixels to the memory cards, the raw I/O throughput rate of these chips has to be very high.

I see no reason the 7D II couldn't have both a higher pixel count AND the additional 2fps bonus over the 7D I. The math certainly adds up...

[rs]

Additionally, not all throughput calculations people do factor in the 14fps frame rate of the 1D X (which it can do with mirror lockup.)

At 14 FPS, its not just a lack of mirror movement and AF, but RAW is off the menu too - read page 113 of the 1D X manual if you want to check up on it.

It all depends on whether Canon want to make it fast or really fast. In reality, the small mirror could move faster than the 12 FPS full sized mirror in the 1D X does, but something tells me they won't do that for marketing purposes. 12 FPS and a very healthy sized buffer wouldn't be out of the question if the MP is kept down.

Admittedly, your calculations resized to 12 FPS for the 1D X in RAW mode make a 24MP camera with the same processing power capable of 9 FPS, which is a slight improvement over the current 7D.

[jrista]

Additionally, not all throughput calculations people do factor in the 14fps frame rate of the 1D X (which it can do with mirror lockup.)

At 14 FPS, its not just a lack of mirror movement and AF, but RAW is off the menu too - read page 113 of the 1D X manual if you want to check up on it.

It all depends on whether Canon want to make it fast or really fast. In reality, the small mirror could move faster than the 12 FPS full sized mirror in the 1D X does, but something tells me they won't do that for marketing purposes. 12 FPS and a very healthy sized buffer wouldn't be out of the question if the MP is kept down.

Admittedly, your calculations resized to 12 FPS for the 1D X in RAW mode make a 24MP camera with the same processing power capable of 9 FPS, which is a slight improvement over the current 7D.

The pixels streaming in from the sensor are "raw"...it doesn't become a JPEG until AFTER the DIGIC chip has done its work and the image is rendered, compressed, and written to the CF card. The input rate from sensor to DIGIC would have to be 473mb/s regardless of what format you end up writing to. JPEG would only matter for the data writeout rate to the CF card. Writing to CF card is limited to at most 150mb/s, on top of which I believe the 1D X has only a 768mb frame buffer...which wouldn't actually be able to accommodate 14fps continuous RAW for a sustained 3.2 seconds (which is what you get at 12fps.)

[Bosman]

I know i am kinda late to the party but I keep trying to buy a 7D but keep stopping myself...The current 7D just doesn't out do my 1D Mark III so i guess i can't justify the regular 7D until I have gotten an upgrade. Why why why does this have to come out in the fall. I'd buy it right now like i bought the 5DM3 last March. Oh God I'd be so happy if they released it 2 weeks from now. Heres to unrealistic hope.

Key question - will the current crop of cameras work for you.  By this I mean will they have sufficient IQ, approriately weatherproof, highest enough light sensitivity, .. So you can get the picture you want.   After many people do great work with a phone camera.

Now if lust over the latest and greatest, perfectly okay to do so,the wait (and wait and wait) as something new is always rumored.

Let me put it this way in one summer i will put just shy of 100,000 images on a camera shooting sports. Many of those days are dark mornings thus iso jacked up and colors taking a hit. While my 1dm3 has served me well since 2007 i have had the shutter replaced 3x, the last time was just over 340,000 actuations. I do believe the 7d center point focus beats my 1dm3 because i have guys that shoot for me that use them and their shots in focus ratio is a little better than mine and its not due to operator error. An advanced hi-iso 7D II with new focus tech in it is just what would suit me. I don't see myself dropping 6.5G's so i can shoot sports and weddings better than my 5d3. Also i want the crop factor. It makes my 24LII a 38.4mm focal length and ideally i like to shoot finishline work at 35mm. With my 1dm3 it is just over 31mm. So the sensor crop, speed and focus tech are what i am looking for in a new camera.
I get the do you need it comments. I guess when i make money at it and i can offer a better more consistant product then i am happy too.

[rs]

Additionally, not all throughput calculations people do factor in the 14fps frame rate of the 1D X (which it can do with mirror lockup.)

At 14 FPS, its not just a lack of mirror movement and AF, but RAW is off the menu too - read page 113 of the 1D X manual if you want to check up on it.

It all depends on whether Canon want to make it fast or really fast. In reality, the small mirror could move faster than the 12 FPS full sized mirror in the 1D X does, but something tells me they won't do that for marketing purposes. 12 FPS and a very healthy sized buffer wouldn't be out of the question if the MP is kept down.

Admittedly, your calculations resized to 12 FPS for the 1D X in RAW mode make a 24MP camera with the same processing power capable of 9 FPS, which is a slight improvement over the current 7D.

The pixels streaming in from the sensor are "raw"...it doesn't become a JPEG until AFTER the DIGIC chip has done its work and the image is rendered, compressed, and written to the CF card. The input rate from sensor to DIGIC would have to be 473mb/s regardless of what format you end up writing to. JPEG would only matter for the data writeout rate to the CF card. Writing to CF card is limited to at most 150mb/s, on top of which I believe the 1D X has only a 768mb frame buffer...which wouldn't actually be able to accommodate 14fps continuous RAW for a sustained 3.2 seconds (which is what you get at 12fps.)

I get what you're saying about what comes off the sensor being raw. The digic chip does all the conversion to jpeg after all. But it can't be the CF card which restricts the use of raw, as the data gets stored in the buffer and then it gets dumped onto the CF card in whatever amount of time it takes. After all, a 1D X doesn't slow down with slow cards if the burst can be contained fully within the buffer. Also, the buffer size doesn't impose any limits on speed of the burst, just its depth.

I don't know where the bottleneck with the 1D X is, but it seems like 12 fps is pretty much the upper limit for that FF mirror to flap around as at 14 fps its stays put, disabling AF and the viewfinder. And 12 fps clearly is also the limits of some part of the processing/storage for 18MP raw, as 14 fps is jpeg only. I believe that's why the 1D X has a lower number of MP that other recent FF cameras - the 1D X's electronics couldn't handle more MP while doing the the mirror box justice. No doubt the replacement for the 1D X with its faster electronics will at least get up to what Canon sees as the normal number of MP for FF - low 20's.

As for what is restricting 14 fps raw, quite obviously that sensor gives the read out just fine, the A/D circuitry works, and its processing system can take the input and convert it. Could it simply be the write speed of the buffer isn't up to taking the data at the rate needed for 14 fps in raw format?

Or could it be as simple as Nikon's 14 bit raw dropping to 12 bit raw at higher frame rates on certain cameras? Possibly Canon didn't want to drop down from 14 bit raw, so they just did a quick and dirty solution of limiting it to jpeg only? 14 bit at 12 fps vs 12 bit at 14 fps... If that's the case, 12fps 14 bit raw is the absolute limit of its processing speed, not just the buffer.

[bdunbar79]

Specification list
http://camerarumors.blogspot.in/2013/03/7d-mark-ii-will-have-32mp-sensor.html

Bwahahahahahaha! 32 MP ....thanks for starting my saturday right with a laugh   

I spit up some coffee.  This is great, I haven't been watching any funny movies lately.  Thanks I needed that.

36 or 40 MP would be good news and a raw mode which bins the pixels into 18/20 + 9/10 MP to get rid of the patterns of monochromatic light sources. With a back side illuminated sensor the net photosensor size would be the same as that of a lower MP sensor giving you the freedom to choose between different resolutions, high ISO modes and a mode which avoids demosaicing completely.

Yeah, but it's not going to be in a 7D Mark II. 

[Bosman]

Specification list
http://camerarumors.blogspot.in/2013/03/7d-mark-ii-will-have-32mp-sensor.html

Bwahahahahahaha! 32 MP ....thanks for starting my saturday right with a laugh   

I spit up some coffee.  This is great, I haven't been watching any funny movies lately.  Thanks I needed that.

36 or 40 MP would be good news and a raw mode which bins the pixels into 18/20 + 9/10 MP to get rid of the patterns of monochromatic light sources. With a back side illuminated sensor the net photosensor size would be the same as that of a lower MP sensor giving you the freedom to choose between different resolutions, high ISO modes and a mode which avoids demosaicing completely.

Yeah, but it's not going to be in a 7D Mark II.

^^ Agreed

[neuroanatomist]

It is kind of surprising that you get jpeg but not raw in the highest speed mode. This suggests that is it _some_ (e.g. bus) bandwidth that is the limitation, and not the number-crunching needed to develop into jpeg.

As suggested above, the reflex mirror seems to impose the 12 fps limit as MLU is needed for 14 fps.  The fact that only jpgs are written at 14 fps does suggest a limitation in the write capabilities, since the RAW image has to be acquired and converted to jpg even when shooting RAW only.

[sagittariansrock]

It is kind of surprising that you get jpeg but not raw in the highest speed mode. This suggests that is it _some_ (e.g. bus) bandwidth that is the limitation, and not the number-crunching needed to develop into jpeg.

As suggested above, the reflex mirror seems to impose the 12 fps limit as MLU is needed for 14 fps.  The fact that only jpgs are written at 14 fps does suggest a limitation in the write capabilities, since the RAW image has to be acquired and converted to jpg even when shooting RAW only.

[Krob78]

I'm certainly not cheering the 7D and it's high ISO capabilities, just that for all the complaints about it, decent, usable images can be had at what is considered a high iso for that camera.  My complaint with it is with regard to high iso and portrait work, there it has given me some challenges...

I agree. I liked your photo,  I think there is nothing wrong in that photo that a decent PP could not fix ... I would desaturate the background a bit so the bird would stand out more. But just curious about your comment on high ISO and portrait work ... why would you do portrait work in high ISO, any decent portrait work is usually (if not always) done in good light or studio lighting etc ... I thought most photographers consider it a sacrilege to use high ISO for portrait work.

I will be a candidate for the 7D MK II, most likely within the year after it is available... 

I'll get it when there is a good rebate ... maybe Christmas time

No real issues for studio work so much.  More with regard to location portraiture and in particular, family shoots rather than individual or couples...

The 7d (mine anyway), has a huge issue with noise in the shadows in my opinion and that's the noise I concern myself with in post for family portraiture.  I do a lot of location images and often times my clientele has a place they want images taken at, rather than my recommendations...  it can be trying at times, hence the 5D MK III is now the new pony in my stable since around Christmas time...  It doesn't seem to have those issues and for what it's worth the two cameras seem to get along fine in my bag!   

That being said, the 7D is seeing less and less clicks these days as I work to accentuate my relationship with my 5DIII... I love them both!

[Krob78]

You can also "help" the 7D noise at high ISO by overexposing and pulling it back down in post...try to get as much noise as possible down into the lower DR.  Easier to process out as well.  I never got anything like this at ISO 3200 but considering it's full frame and formatted for the web.  It is quite good.

I'll be rushing out immediately to pick up a new 7D MK II.  What all the hub-bub is about regarding the original 7D is that the complete lack of any high ISO usage, renders images useless and totally unusable for any purpose, as this sample taken at ISO 3200 on my original 7D, clearly shows...  7D MK II please get here soon! 

EOS 7D
EF 100-400mm f/4.5-5.6mm L IS USM
1/1000 sec
F/6.3
ISO 3200
400mm

Good photo (love the colors), but I would say there are some issues with it that are definitely due to the camera and lens combination. There is also no embedded metadata, so there really isn't any way for any of us to verify that it really was shot at ISO 3200.

I've taken a lot of bird photos at ISO 3200 on the 7D (mostly with the 100-400mm lens), and the noise is usually a lot worse than that. Assuming you have not applied any NR, there is definitely a loss of fidelity. There is also a loss of detail. The 100-400mm lens imposes increased softness at f/6.3, and even more at f/5.6...usually, f/7.1 is the sharpest aperture (as f/8 is beyond the DLA of the 7D sensor and just barely a touch softer).  Having used both the 100-400 as well as the 300 & 500 Mark II lenses on my 7D recently, I can attest to the fact that at that magnification/crop, even at ISO 3200, the lens is definitely costing you some sharpness and detail definition there...even though they are noisy, a shot taken the 300 or 500 w/ 1.4x TC at f/6.3 usually has much sharper detail.

I'd also point out that the full-size download shows a fair amount of posterization in the background, which is another one of the issues you encounter when using higher ISO settings with small pixels. Gain is so high for such a low full well capacity that you get a lot more quantization error than with something like the 5D III or 1D X and their significantly higher full well capacities.

Thank you.  In the interest of keeping things above board, I've re-uploaded the file with the metadata embedded as you desired sir! 

I agree that there is certainly some posteriztion in the background and it is certainly far from perfect, I would edit exposure and some other things as well.  That being said, it is not a crop, it's the actual size of that particular image.  I have other bird and wildlife images at high iso's with my 7d as well and some that are really quite usable.  Most with my 70-200mm f/2.8L IS II.  It is much more challenging with the 100-400 as you pointed out. 

I'm certainly not cheering the 7D and it's high ISO capabilities, just that for all the complaints about it, decent, usable images can be had at what is considered a high iso for that camera.  My complaint with it is with regard to high iso and portrait work, there it has given me some challenges...

I love my 7d, I love my 5D MK III even more, not at first though... I will be a candidate for the 7D MK II, most likely within the year after it is available...  My 7D will be passed down or sold sometime in that future... I must say, it's served me well for the over 50,000 images I've taken with it so far!

Thanks East Wind, I didn't really do much to this image, it wasn't a keeper for me, I would apply NR in LR4 and certainly make some adjustments in exp, contrast, saturation, etc. if I thought to keep it for something.  Mostly I just threw it up there, as someone mentioned the crappy 7D and it's lousy high iso capabilities.  My point is just that it's not that terrible... can be trying at times and give one some extra work in post but the 7D is and has been a very formidable and fun camera for a lot of us.  I'm sure I could get years more service out of it...

I still have to say though, the sound of the shutter speed on my 7D sounds incredibly more fast than my 5dIII.  I didn't think it would be that noticeable going from 8fps to 6fps, however it's quite different in my opinion and to my ears...

I shot a soccer game yesterday and could readily see the shots in between that would have been nice to have from my 7d, if I had only brought it... It sounds like my 7d on low speed!  However, I love, love, love the silent high speed shutter mode!  It's excellent and hopefully makes it into the 7dII... 

[neuroanatomist]

since the RAW image has to be acquired and converted to jpg even when shooting RAW only.

When a RAW image is captured, a JPG conversion is done in-camera to create a small preview image, and that JPG preview is saved within the RAW file container.  That JPG preview is what you see on the rear LCD of the camera after the shot, and importantly, it's what's used to generate the histogram data and the highlight alert if you use those features. 

It's worth knowing, because many people think that the in-camera settings are totally irrelevant if recording RAW images - that's mostly true (long exposure NR is an exception), but if you make exposure decisions (e.g. ETTR) based on the histogram, the settings for Picture Style (contrast, etc.), ALO, HTP, white balance, etc., are all applied to the JPG preview image and the histogram data/blinkies, and that can affect your exposure decisions.  Some people actually use a modified Picture Style that makes the JPG image more closely resemble the RAW file in terms of exposure. 

[sagittariansrock]

since the RAW image has to be acquired and converted to jpg even when shooting RAW only.

When a RAW image is captured, a JPG conversion is done in-camera to create a small preview image, and that JPG preview is saved within the RAW file container.  That JPG preview is what you see on the rear LCD of the camera after the shot, and importantly, it's what's used to generate the histogram data and the highlight alert if you use those features. 

It's worth knowing, because many people think that the in-camera settings are totally irrelevant if recording RAW images - that's mostly true (long exposure NR is an exception), but if you make exposure decisions (e.g. ETTR) based on the histogram, the settings for Picture Style (contrast, etc.), ALO, HTP, white balance, etc., are all applied to the JPG preview image and the histogram data/blinkies, and that can affect your exposure decisions.  Some people actually use a modified Picture Style that makes the JPG image more closely resemble the RAW file in terms of exposure.

Makes a lot of sense now that you've explained it. Thanks, that was very helpful.