Canon thinks the camera market will drop by another 50% over the next two years

[dolina]

The Canon exec is talking about the camera business as a whole that includes point & shoots, consumer ILC, professional ILC and accessories

The most affected by this decline are consumer camera sales.

Consumers tend to keep their dedicated still cameras longer and not upgrade unless their camera cannot be economically repaired. They often have only 1 lens for their ILCs.

What they did upgrade often though are smartphones in which they upgrade every end of contract that typically lasts for 2 years with their carriers/telco or buy on their own every 3-5 years.

People are incentivized to upgrade their smartphone because the camera is always better and the processing is always faster on a newer phone that is smaller than their pocket.

But when it concerns the high end/professional cameras then I expect sales to be maintained or be slightly higher.

So it really isn't exactly new news about declining sales but tacking on the 50% figure decline is what is news.

 

[Ozarker]

The Canon exec is talking about the camera business as a whole that includes point & shoots, consumer ILC, professional ILC and accessories

The most affected by this decline are consumer camera sales.

Consumers tend to keep their dedicated still cameras longer and not upgrade unless their camera cannot be economically repaired. They often have only 1 lens for their ILCs.

What they did upgrade often though are smartphones in which they upgrade every end of contract that typically lasts for 2 years with their carriers/telco or buy on their own every 3-5 years.

People are incentivized to upgrade their smartphone because the camera is always better and the processing is always faster on a newer phone that is smaller than their pocket.

But when it concerns the high end/professional cameras then I expect sales to be maintained or be slightly higher.

So it really isn't exactly new news about declining sales but tacking on the 50% figure decline is what is news.

Where is your data to back up what you say from?

 

[unfocused]

Where is your data to back up what you say from?

Are you questioning what dolina is saying? Because it's pretty much common sense and if you've followed the industry over the past several years and read any of the regular industry reports and interviews that have been covered on this forum, his remarks are consistent with those reports and interviews.

Do you have any data that would cast doubt on his comments?

 

[Photointo]

I'm with canon from 2006, but now I think the future belongs to SONY. It's like canon can't make more innovative matrix than that from 5D Mark IV...

 

[HarryFilm]

Sorry Harry, that's all just a sequence of pseudo-scientific statements; neural networks are very useful and powerful when applied within certain narrow areas but unfortunately not even close to simulate the actual human brain. Moreover they cannot even simulate a behaviour of a simple worm with 300 neurons.

===

Neural Nets are just basically organized groups of software-based weighted results of a group of If-Then-Else/Switch/Case-Of statements done after a matrix operation (i.e. convolution kernel) has been performed on a numeric input which represents a chemical/electrical value. What I am espousing is more at the hardware level where you are actually emulating the electro-chemical gating/pass through of potassium/sodium/phosphorous/lithium/etc since it seems only about 25 trace elements are actually required for general neurochemistry. A simple record-based infrastructure held in a massive multi-dimesional array to represent general organic chemistry reactions as we know them today has in fact allowed us to emulate SPECIFIC neural structures in worms, flies, rats, cats and even humans and is the MOST PROMISING AVENUE to get to 100+ IQ Levels of General Artificial Intelligence.

In rebuttal to your earlier statement of only 300 neurons of a worm being emulated via a neural net, modern simulations of human brain tissue are now into 4.5 Million Neurons and the largest simulation I know of personally is pretty much emulating ALL 100 Billion Neurons of the human brain on a massively parallel, combined CPU/GPU system that used 128-bits wide GaAs SoC chips running at 60 GHz. The amount of local DRAM memory used for the array-of-record superstructure is into the EXABYTES of space and requires truly fast networking (Dense Wave Optical Multiplexing) into the many PETABYTES PER SECOND range at each compute node! So YES we are talking about some supremely EXPENSIVE gear!

See Waterloo prof constructs world’s second largest simulation of a human brain (2017):
https://www.therecord.com/news-stor...-world-s-largest-simulation-of-a-human-brain/

From a computational point of view, we are currently at a functional emulation level that is set at around molecular-level interactions rather than atomic-nucleus levels of emulation which means we can emulate the base organic chemistry of a human synapse. And when we organize them BY function we can actually get self-organization to take place. We haven't yet MAPPED the entire brain BUT we humans HAVE mapped common neuro-functionality to such an extent that we don't actually have to simulate the low-level atomic-level interactions but can emulate upper-level functionality enough to allow us to get to human 100+ IQ levels! Software-wise we have the technology to do so to a high degree of fidelity.

What the CURRENT PROBLEM BEING FACED is the sheer physical size of the hardware! A 200 PetaFLOP supercomputer is about the size of a small warehouse or a common high school basketball court and to make it PRACTICAL we need to shrink that down to the size of a basketball or smaller. That will take a few years! For a basic convolutional neural net that allows analysis of 3D-XYZ edges and surfaces and their comparison to common objects in any given 3D-XYZ orientation and size, we can NOW SHRINK THAT down to using only FOUR Qualcomm SnapDragon 845 SoC chips (used in many high end smartphones) which will allow us to do VERY HIGH END computational photography that can actually get to pro-level skills.

See more info on convolutional neural networks:
https://en.wikipedia.org/wiki/Convolutional_neural_network

And based upon my industry contacts I do know that 16-core and 32-core ARM SoC chips ARE being worked on by BOTH Apple (A15/A16) and Qualcomm (SnapDragon 865/875) to be ready for 2021 to 2022! This means that ONE single CPU chip can be put into a Smartphone, Tablet or Mirrorless camera to allow pro-level still photography and videography to be automatically imaged by the average consumer!
.

 

[Kit.]

I'm with canon from 2006, but now I think the future belongs to SONY. It's like canon can't make more innovative matrix than that from 5D Mark IV...

My Sony laptop is 7 years old and dying, but I am not buying the new Sony one... guess why.

Also, Canon helps killing the P&S market by not releasing G7X III on time. I mean, really, it's a lot of lost sales for the most promising P&S line in the last half a year.

 

[HarryFilm]

If you can't tell the difference between arri and iphone footage then I doubt your video experience is in the market bm, arri and red and higher up canon cine are aimed at. If you've ever edited anything or been involved with lighting complex scenes you'd know a phone can't do what cine cams can. Sure they may look the same for certain situation and lighting when operated by the youtube content creator types but imho they are barely scratching the surface of the capabilities and would be just as well served with a dslr.

Fwiw I am not a serious video guy, although have actual independent cine folks in my family on fathers side (although he is still guy himself but got out at the end of film). None of the cine members in the pro world use dslr's (in my family I mean) never mind phones for a reason. Good luck getting your iphone to output something that has enough editing headroom when they are all compressed 422 at best I imagine. Bet some of my relatives can tell sitting at their grading desk. These cameras are not aimed at the likes of you and I and those that NEED them know why and it isn't meaningful OOC files at same settings look the sameto us posted on facebook compressed to hell. Not to mention phones don't have any choices on lenses, high DR demands of some scenes, flexibility of settings, accessories and compatibility with them and how they fit into lineup with other cameras used on set and so on.

---

You take 4K footage shot on any major Smartphone at 4:2:2 (preferably 4:4:4 if you can use that recording setting on your camera software), use Lanczos-3 or Lanczos-5 resizing (i.e. a 3x3 or 5x5 resampling filter) to get it down to 1920x1080p 4:4:4 uncompressed for editing, increase brightness by 5% to 10%, reduce highlights by 15%, boost shadows by 10% to 15%, increase contrast by 5% to 10%, increase saturation to personal taste AND finally add EITHER a slight Unsharp Mask for edge-enhancement (i.e. sharpening) OR an Antialias Filter...AND BOOM you get near Arri Alexa quality! I do this nearly every day!

I am NOT some average Vlogger but rather a guy with 30 YEARS of video/cinema experience on everything from 1987 era Betacam SP to Arri 2.7k to 4k Canon C700 to 8K Red cameras to Multi-camera 360 degree surround view 4k/8k setups. We do use a LOT of lighting with 5k, 10k and 20k HID cinema lighting setups on our industrial, scientific and commercial photography/videography sets. AND YES we HAVE used Smartphones and GoPros on these complex setups AND MATCHED them to Arri-shot or Canon-shot footage! When you use LOTS of lighting, shooting with smartphones works GREAT!

We have had 30 second and 60 second Car Commercials RUN on major North American, European and Asian Network TV shot ENTIRELY on iPhones! And for SOME REASON, more and more of my requests ARE for using/shooting on low-cost cameras such as smartphones from the major electronics companies! I no longer have a problem with those requests since even our scientific imaging department which I normally work for is getting in on using ARRAYS of smartphones as their imagers!

I have seen our aerospace parent company shoot it's videos in the toughest of Ground and Space conditions on small camera systems so Computational Photography has come a LOOOOONG WAY!

On a technical note, if you shoot 8-or-10-bit 4:4:4 on an iPhone and shrink the 4K footage DOWN to uncompressed 4:4:4 1080p resolution it actually DOES look utterly fantastic! It's just math and pixel averaging which makes it look good! It's in LOW-LIGHT situations where I can tell from basic noise floor measurement which camera is which but since MOST of my imagery uses LOTS of natural and artificial lighting, we can get away with shooting on Smartphones. (this is done BY REQUEST OF THE CUSTOMER as normally I shoot on Canon C700 Global Shutter or Full frame or a Red Monstro and sometimes the odd rented Alexa-65!)

.

 

[Quarkcharmed]

In rebuttal to your earlier statement of only 300 neurons of a worm being emulated via a neural net, modern simulations of human brain tissue are now into 4.5 Million Neurons and the largest simulation I know of personally is pretty much emulating ALL 100 Billion Neurons of teh human brain on a massively parallel, combined CPU/GPU system that used 128-bits wide GaAs SoC chips running at 60 GHz. The amount of local DRAM memory used for the array-of-record superstructure is into the EXABYTES of space and requires truly fast networking (Dense Wave Optical Multiplexing) into the many PETABYTES PER SECOND range at each compute node! So YES we are talking about some supremely EXPENSIVE gear!

The worm *is* being emulated https://en.wikipedia.org/wiki/OpenWorm it just doesn't behave as expected, neither does any of the existing models. Maybe one day the worm will work, but currently we humans are failing to re-create even a simple worm, so recreating a human brain is totally out of question. And it's not about the computational power, the worm's neural network doesn't require much in term of modern computers.

See Waterloo prof constructs world’s largest simulation of a human brain (2017):

https://www.therecord.com/news-stor...-world-s-largest-simulation-of-a-human-brain/

Harry it's like a nuclear explosion, it's a largest possible simulation of the Sun but not the Sun yet - and will never be.
The rest is very interesting topic but well beyond this forum. The point is, we're very far from having a robo-photographer, and advanced AI with scene type recognition will not affect Canon roadmap in the near future.

 

[RayValdez360]

Does it still make sense, for Panasonic, to enter a declining market? Less cake, more slices...
Should Canon try to increase their market share with additional and competitive models instead of slamming the brakes on? Anyway, apart from Canon, nobody knows what their real strategy will be. Unlike most youtubers, they are market -experienced. We can only wait and see...and remain confident!

what the hell are you talking about. Canon is like number 1 they dont need anyone's advice when it comes to market share. So what if Panasonic is entering they still think it is profitable. Also if companies knew everything then none of them would ever had bad sales or go out of business.

 

[tomscott]

Its a catch 22 if they don't innovate and bring new product to the market then how can they expect to sell more? The product line is so stagnant and it has been since the 5DMKIII which was 2012, thats 7 years of not much happening or improvement. The schedule of updates is so long that there isnt really much to keep up with... hard to get excited.

In all honesty I dont frequent CanonRumours that much anymore because there is never any news worth catching up on.

Seems the same across tech generally, look at whats happening over at Apple... iPhones slumping the iMac has gone 602 days since the last update.

People are fed up all over the place waiting for updates but companies are more interested in selling services than hardware. I dont feel like we have hit the ceiling yet but maybe the investment for such improvements is a lot more than it used to be to see similar gains.

 

[Kit.]

Sorry Harry, that's all just a sequence of pseudo-scientific statements; neural networks are very useful and powerful when applied within certain narrow areas but unfortunately not even close to simulate the actual human brain. Moreover they cannot even simulate a behaviour of a simple worm with 300 neurons.

There is a huge difference between emulating human reasoning and emulating human brain. And there is no practical reason to emulate human brain. Human brains suck.

Human reasoning sucks too, but there is one practical reason to emulate it: it will give a robot photographer an easily automated critic. There is no practical need in emulation of human reasoning in the robot photographer itself.

 

[Stuart]

What does "corporate sales than consumer retail sales " mean - is it only in the camera space that canon is doing this.
Might cameras go the way of the lathe and only be for enthusiasts in their sheds? Or was the real decline in enthusiast sales because of waiting for the new mount cameras and then the 6 month price drops.

 

[Pape]

I am more worried about stupid drone auto cameras ,what about actually intelligent cameras . Ten thousend drone randomly flying around, taking random pictures what ever topic they encounter. will eventually produce lot of really good photos. Like those famous monkeys who wrote new version of biblia.

 

[tomscott]

I think its amazing that non of the big camera manufacturers are looking at what Apple/Samsung/Google are doing with computational photography and adding features into cameras that have incredible sensors. Would save so much time when editing and you could choose what is right for your situation or even turn it on or off after the fact during editing etc Night mode on the pixel 3 for example is really impressive.

DSLRs/Mirrorless are essentially the same just a different form factor, its nothing revolutionary. If anything for me they dont change much because when they are made to be ergonomic they are very similar in size and form factor. The nice thing is the benefit to lens design etc but lets be fair the suite of lenses we have is already pretty amazing.

Just feels like the industry has blinkers on and they are doing nothing other than changing the form factor which really is a plaster to sales not the answer. If they are predicting lowering of sales then less R&D will be placed into the sector so there is nothing to get excited about.

Camera phones are incredible for what they are but they are still so poor compared to a full frame DSLR because of physics.

 

[Kit.]

I am more worried about stupid drone auto cameras ,what about actually intelligent cameras . Ten thousend drone randomly flying around, taking random pictures what ever topic they encounter. will eventually produce lot of really good photos. Like those famous monkeys who wrote new version of biblia.

Alone, unguided? Unlikely.

But if you add millions of monkeys clicking "Like" buttons under those pictures, you can train the drones to intentionally produce photos that the monkeys will like.

 

[RayValdez360]

Its a catch 22 if they don't innovate and bring new product to the market then how can they expect to sell more? The product line is so stagnant and it has been since the 5DMKIII which was 2012, thats 7 years of not much happening or improvement. The schedule of updates is so long that there isnt really much to keep up with... hard to get excited.

In all honesty I dont frequent CanonRumours that much anymore because there is never any news worth catching up on.

Seems the same across tech generally, look at whats happening over at Apple... iPhones slumping the iMac has gone 602 days since the last update.

People are fed up all over the place waiting for updates but companies are more interested in selling services than hardware. I dont feel like we have hit the ceiling yet but maybe the investment for such improvements is a lot more than it used to be to see similar gains.

the catch 22 is that cameras can only innovate so much especially on the photo side of things.

 

[YuengLinger]

Many people seem so addicted to their phones they can't put them down to pick up a better camera that takes two hands to operate.

 

[Ozarker]

Are you questioning what dolina is saying? Do you have any data that would cast doubt on his comments?

Why would I? I'm asking. What part is it you think I am disputing? His assumptions or the data? Data is data. Personal assumptions are always open to questions. I'm not disputing anything.

"But when it concerns the high end/professional cameras then I expect sales to be maintained or be slightly higher." That's an opinion and assumption... with no data. I don't have a problem with it, just wondering where that idea comes from.

 

[tomscott]

the catch 22 is that cameras can only innovate so much especially on the photo side of things.

Maybe not from an image quality side as most people are pretty happy with the quality of the images available across the majority of dedicated cameras. Its usability, integration, easy share, internal editing capabilities etc etc

At the end of the day the camera is just that a dedicated stand alone tool. There is a lot they could do which would make owning, making and sharing images more intuitive and make you more likely to grab a dedicated device and reduce the need to move images over to another device to edit and share.

Its the reason the smart phone is so popular - A, because its always with you B, because its use case is much bigger than basic communication.

Im not saying I want a camera with the functionality of a phone etc but what smartphone companies are doing is computational photography because of the limitation of the size of the sensor and optics is pretty incredible. Imagine scaling that up from a phone sensor to a full frame sensor.

Something as simple as auto HDR on the iPhone and compare it to a DSLR and its night and day. Auto HDR that looks natural and is editable after the fact.

That is literally touching the surface.

 

[knight427]

What does "corporate sales than consumer retail sales " mean - is it only in the camera space that canon is doing this.
Might cameras go the way of the lathe and only be for enthusiasts in their sheds? Or was the real decline in enthusiast sales because of waiting for the new mount cameras and then the 6 month price drops.

corporate sales: