There will not be an EOS 5D Mark V [CR2]
- Thread starter Canon Rumors
- Start date
Certainly for those who prefer OVF, this is sad. But "stupid" on Canon's part? When times are tough and very competitive for a business, tough decisions are not made lightly. Yes, some customers are going to be disappointed, and some employees are going to be either let go or transferred. Factories must be reconfigured. But consider how much turmoil the photography industry has gone through in the past 20 years and how well Canon has done relative to the still-standing competition.
Reducing menu size, offering what sells best and can be produced and served (distributed) with coherent marketing seems a pretty standard, prudent move in the kind of business environment we live in today. Succesful in the long run? Can't say yet. But "stupid"?
PS I used to feel like you about EVF. I got the R so I could shoot the incredible rf 50mm f/1.2L. Sold my 5DIV with no regrets. (I definitely regret selling my ef 100-400mm II, but I needed the money!)
Upvote
0
Seconded on this. Canon have made it clear that the R system is their future. The 5D IV as people keep saying is an incredible camera. It's also one that is STILL AVAILABLE. It remains a compelling choice for photographers who want the feature set that it offers.
It seems there's a false equivalence in fans minds that the R5 should somehow also be offered as a DSLR, and therefore the 5D V or whatever. At this point, a 5D series with a bigger sensor, or better video would have to compromise something more - data speeds, image capture speed, frame size, etc etc.
Those feeling the sting of it ending - this is also beneficial to Canon, and a calculated business move. By releasing compelling R products, they are encouraging you to switch. And in so doing, upgrading your system and feeding the profits. Thats how businesses work. You as a consumer don't need to do that. You still have access to a 5D IV, and it still remains an amazing camera. If you want faster, bigger, and better - then it's time to make the switch.
Upvote
0
As above, in terms of multithreading, if it's used, then it's used to speed up the delivery to the closest frame update, not to stretch processing across several frames.
Image processing, especially 2D, is very well parallelisable.
Upvote
0
I have a much simpler example to show how EVF fps isn't necessarily the biggest contributor to the lag: I film a 40 second 4k 120fps clip, stop recording and hand you the camera. You then watch that clip through the EVF.
120 fps input, 120 fps output on the EVF and for some reason there's like a minute of lag between the light hitting the sensor and you seing it displayed on the EVF.
Upvote
0
I agree with most of your post, but I'd just whisper here: the ergonomics of a big body, and top-notch durability.
Upvote
0
This example is totally irrelevant to the problem in question.
Hint: of you wanted to introduce a one minute delay inside the camera, you'd have to store 1 minute of 120fps video in a huge internal memory buffer.
Or write and read it at the same time to/from a memory card.
Upvote
0
I've been working with DSLR for years, The past year I changed to eos R, I did about 30 weddings in all this period, and the difference was abismal. I can say at 100% and I will repeat it
mirrorless cameras works a lot better than DSLR to do autofocus in low light conditions. Much faster and accurate 
Peace
Upvote
0
And that's exactly the point we're trying to get accross: there are buffers between the sensor and the EVF! No amount of magic multithreading is going to fix that.
Upvote
0
This is completely false. The difference between 60fps and 120fps is night and day, ask any PC gamer.
Upvote
0
Correct me if I'm wrong, but to me, the positions are as follows:
@Quarkcharmed acknowledges the theoretical possiblility of a pipeline/buffer for a series of multiple images during processing for the viewfinder etc. and a lag of more than 1/(EVF frame rate), but highly doubts that it is actually implemented. One of the arguments for that position is: why bother implementing a 120 Hz viewfinder, which is much faster than the human eye can follow, if not for the main reason that the total lag from capture to display is already LESS than 1/(EVF frame rate), so that the 120 Hz actually make sense.
@koenkooi on the other hand believes that processing an image from capture to display will take MUCH longer than 1/(120 Hz) = 8.333 ms, so it is inevitable that there is extensive buffering of multiple images involved.
And I really have no idea which one of these positions is closer to the truth. While a multithreaded pipeline buffering images from stage to stage seems the most logical thing to implement, because virtually every sort of signal (and therefore also image) processing generally involves multiple stages for amplification, filtering, corrections, etc. and it seems quite natural to implement every such stage as a single thread receiving the image from the previous stage, processing it, and storing it in a buffer that is handed to the next stage. And buffers in the range of a few images with the size of the EVF do not seem overly memory extensive to me. On the other hand, a processing time of over 8 ms for a single image just for (relatively) low-res display does also seem ample to me, but I am not in real-time image processing. Since no compression algorithms are involved, but (likely) rather some very basic processing stages, I do not feel that this requires necessarily more than 8 ms per image.
@Quarkcharmed acknowledges the theoretical possiblility of a pipeline/buffer for a series of multiple images during processing for the viewfinder etc. and a lag of more than 1/(EVF frame rate), but highly doubts that it is actually implemented. One of the arguments for that position is: why bother implementing a 120 Hz viewfinder, which is much faster than the human eye can follow, if not for the main reason that the total lag from capture to display is already LESS than 1/(EVF frame rate), so that the 120 Hz actually make sense.
@koenkooi on the other hand believes that processing an image from capture to display will take MUCH longer than 1/(120 Hz) = 8.333 ms, so it is inevitable that there is extensive buffering of multiple images involved.
And I really have no idea which one of these positions is closer to the truth. While a multithreaded pipeline buffering images from stage to stage seems the most logical thing to implement, because virtually every sort of signal (and therefore also image) processing generally involves multiple stages for amplification, filtering, corrections, etc. and it seems quite natural to implement every such stage as a single thread receiving the image from the previous stage, processing it, and storing it in a buffer that is handed to the next stage. And buffers in the range of a few images with the size of the EVF do not seem overly memory extensive to me. On the other hand, a processing time of over 8 ms for a single image just for (relatively) low-res display does also seem ample to me, but I am not in real-time image processing. Since no compression algorithms are involved, but (likely) rather some very basic processing stages, I do not feel that this requires necessarily more than 8 ms per image.
Upvote
0
It was inevitable that Canon would not introduce a 5DV while the R5 is fresh. Canon obviously want to milk the mirrorless RF mount as much as they can, I just hope they are going to throw Sony a few quid for reinvigorating a mature ILC market with the "mirrorless" concept.
The title of this thread is at odds with Canon going on record as stating that they are continuing to support and develop the EF mount going forwards. Given what Canon have already published I think this rumour should have read " There will not be a 5DV for a while" And that's not at all surprising.
The 5DIV is more than capable, pretty future proof, and probably not going to be under brand rivalry pressure to update as no doubt Nikon will be acting in the same way.
Canon will know that once people by the R series they will get the urge to buy dedicated RF lenses eventually, and so they can resell lenses that are now sat in a very mature and stagnating market, all over again. Smart business move ! Over the next two years the market will decide when or whether we get a 5DV IMHO.
The title of this thread is at odds with Canon going on record as stating that they are continuing to support and develop the EF mount going forwards. Given what Canon have already published I think this rumour should have read " There will not be a 5DV for a while"
And that's not at all surprising.The 5DIV is more than capable, pretty future proof, and probably not going to be under brand rivalry pressure to update as no doubt Nikon will be acting in the same way.
Canon will know that once people by the R series they will get the urge to buy dedicated RF lenses eventually, and so they can resell lenses that are now sat in a very mature and stagnating market, all over again. Smart business move ! Over the next two years the market will decide when or whether we get a 5DV IMHO.
Upvote
0
Your system must be able to process captured frames in less than 1/120s, or it won't be able to feed your EVF at 120fps. Introducing additional buffers to store more frames (in your example - thousands) is absolutely meaningless.
Upvote
0
Because your assessment of the two systems isn't shared by everyone? EF isn't obsolete, your EF lenses will still work just as well if and when you get an RF camera. A non-exhaustive list: EF still offers loads of lenses that don't exist in RF (yet or perhaps ever). Some people value things like OVF over other considerations. Some people won't benefit from the new AF system (even if it were objectively better), because they shoot MF, or their use cases are already adequately covered. Some people don't like change.
Upvote
0
Most likely some processing is happening during exposure and readout. After it finishes, the EVF switches the buffer (likely it uses alternating buffers).
Specialised CPU like DIGIC with additional instructions for image processing converts sensor raw data into EVF-ready buffer much faster than you think, that is, 8ms time seems to be quite ample and plausible.
Upvote
0
For one, people are accustomed to the look of 24p in movies and 30p in TV, and think 60p "looks weird".
Upvote
0
