Interview with Canon's Larry Thorpe on C300 - Explains a lot

Status
Not open for further replies.
P

Piolet

Well-known member
Here are some Canon interviews I've found regarding the C300:

Larry Thorpe interview by DPReview:

http://www.dpreview.com/articles/7685106578/canon-eos-c300-launch-interview-with-larry-thorpe

Some notable facts:

  • The sensor can do 60p. The 30p limitation is due to the processor (older Digic DV III) and codec. New cameras with new processors may do 60p.
  • They developed the camera in only two years. The wanted to do it fast so they reused stuff from their XF305, like the codec, processor, and VF. This caused some limitations like the 30p.
  • They have a master plan and the C300 is just one step.
  • Sensor does not do any de-bayering
  • A lower cost model "would make a lot of sense in the marketplace"
  • "We think this is a very very good start but there's no question that 4K is coming, so we have to keep our eye on that."
  • "We've stepped in, and we're never going to stop. We're in for the long term. Wherever the marketplace dictates that we should go, and wherever our technology allows us to go, we'll be there."
**********************

Larry Thorpe interview by Imaging-Resource.com:

http://www.imaging-resource.com/new...-thorpe-on-the-new-cinema-eos-digital-cameras

Some cliff notes:

  • The cine lenses are developed by the Broadcast Division optics group (not the EOS group). The guy that heads up this unit is delighted to be in this new business. He's the same guy that worked on lenses for NHK's 8K UHD project.
  • By 4th quarter next year they'll have two more new cine zooms. These will be lighter and less expensive then the two they already announced. They will still resolve 4K.
  • In 2013, there will two more cine primes. On another site, this is reported to be an ultra-wide and a telephoto prime.
  • He goes into some details about how the sensor works. It puts out 4:4:4 without the need for debayering or interpolation. This cuts down on calculations that need to be done in camera (that's why it uses so little power).
  • The EOS and video group use to live in separate worlds, but that's definitely changing. "This project brought some of them together, and Mr. [Masaya] Maeda, he was banging heads to make sure that as we go forward, you're going to see more and more of that, so there's going to be a lot of cross-fertilization." Masaya Maeda is the Chief Executive of the Image Communications Product group--basically the head cheese of Canon's entire camera/optics group.
  • The C300 sensor was designed by both the EOS and video groups.
  • He gets into some technical stuff about 4K lenses and what they have to do.
  • Interviewer asks a great question about whether Sony is pissed off at them now. Larry gives a diplomatic but honest answer.
**********************

Larry Thorpe audio interview with HD Magazine:

http://www.definitionmagazine.com/j...y-thorpe-on-the-c300s-design-and-origins.html

Some highlights:

0:00 Talks about the sensor and how it works. Lots of technical details
2:22 Talks about the processor
3:11 Goes into why HD-SDI isn't 10 bit
4:00 "We will move from that [8-bit] in future cameras that we have planned."
5:12 Tells story of how he got involved in the camera
6:15 Talks about the camera's origin and development
8:14 Talks about the EOS and video divisions
8:45 Explains why there are two separate, fixed lens mounts
10:40 Explains Canon's strategy and reasoning behind the C300
12:37 Politely talks about competitor's cameras
13:45 "[The C300] is not the next generation. We are very aware of what the next generations are in codecs and things like that. You can be sure down the road we will be playing with things like that."

**********************

Video interview with Larry Thorpe: http://jtown.tv/

Skip to 107:04 for the interview. Most of the info is similar to what you heard before, but there are some interesting PowerPoint slides for the technical guys and a few new snippets:

  • The sensor has four parallel readouts...four wires if you will, one each for red, blue, green1 and green2.
  • An equivalent to the Hollywood Service center is being set up in New Jersey to service the New York area. Will open by the end of the year.
  • By the time the camera ships, there will be a total of 6 service centers in the US that can service the C300.
  • Roller shutter at 24 fps is reduced because the sensor reads out at 60 fps (16 ms) which then gets re-clocked back to 24 fps. If you shoot at 60i, the camera is reading out at 120 fields/sec.
  • Regarding 60p: again the reason is the DSP and codec from the XF cameras are limited to 30p. But "We are totally aware of what's going on [in the marketplace] and we are working accordingly."
  • Explains Canon log with some more detail
  • Engineers virtually eliminated fixed pattern noise in the sensor. FPN is what people dislike about video. Beating it way down is what makes the C300 noise seem reminiscent of film grain (although not exactly the same).
  • The IR filter and ND unit is a sealed chamber.
  • Selling pricing in the US is not released yet because they're still signing contracts with select dealers and rental houses. This will take about another month. Once that's done, we'll know the number.
  • The camera can read LUTs through the SD card. Canon is working with 3rd party vendors for the LUTs.
**********************

FD Times interview with Mr. Masaya Maeda, Canon’s Managing Director and Chief Executive of Image Communication Products Operations.

http://www.fdtimes.com/news/canon/interview-with-canon-managing-director-mr-maeda-2/
 
Last edited:
Great interview and very informative. The C300 looks to be a fantastic camera.

However I have a problem with this statement:

"Wherever the marketplace dictates that we should go, and wherever our technology allows us to go, we'll be there."

Why wait for the market or other companies *cough *cough RED to dictate where you should go? Innovate and break barriers, Canon should (and could) dictate the market.

I have no doubt the c300 will be a great camera and its good to know that Canon is entering this market and sounds like they are in for the longhaul. Sounds like we will see more than a few great cameras from them over the coming years, however you have to respect what RED is doing TODAY with they're feature-set/pricepoint. EPIC on the highend is cheaper than an ALEXA/F35, etc. and now Scarlett is cheaper than this C300, F3, etc. and wins the spec-sheet hands down. Now of course it all comes down to real world image quality and functionality but I doubt the Scarlett will disappoint. In the end we as consumers all win so bring em on.
 
For my video and photo needs, I love Canon. But what annoys me about them is that they're always last to the game. I wish they'd take the bull by the horns and run the show.
 
Mickey G said:
Great interview and very informative. The C300 looks to be a fantastic camera.

However I have a problem with this statement:

"Wherever the marketplace dictates that we should go, and wherever our technology allows us to go, we'll be there."

Why wait for the market or other companies *cough *cough RED to dictate where you should go? Innovate and break barriers, Canon should (and could) dictate the market.
Click to expand...

I think what they're saying is that they will listen to us (the market) and make products that fill our needs and wants. They seem very interested in feedback, hence that Hollywood facility which also serves as an R&D satellite.

Morox said:
For my video and photo needs, I love Canon. But what annoys me about them is that they're always last to the game. I wish they'd take the bull by the horns and run the show.
Click to expand...

That has been true in the past, but it seems like they're really serious this time about motion imaging. If you look at still photography where they're very serious, they're constantly leapfrogging with Nikon to see who has the best technology. I'm optimistic that Canon will eventually be very competitive in motion imaging.
 
The sensor can do 60p. The 30p limitation is due to the processor (older Digic DV III) and codec. New cameras with new processors may do 60p.
They developed the camera in only two years. The wanted to do it fast so they reused stuff from their XF305, like the codec, processor, and VF. This caused some limitations like the 30p.
Click to expand...

"Only two years." So what's normal? Three or four years? People complain about RED taking a long time, but this is yet more evidence that Red isn't slower than others at moving from drawing board to assembly line, just more transparent. This is both excruciating and good, because while Red is announcing stuff that's still a couple years away it gets the other companies moving. "Oh no, Red's coming out with such-and-such. We had better start on something like it" instead of "Oh no, Red has come out with such-and-such. We had better start on something like it."

A lower cost model "would make a lot of sense in the marketplace"
Click to expand...

I shudder to think what they would take away from C300 to bring it down from $20,000.

"We think this is a very very good start but there's no question that 4K is coming, so we have to keep our eye on that."
Click to expand...

4K is coming, but I can't see it. ;) Okay, maybe I can see it, but just barely. :)

Wherever the marketplace dictates that we should go, and wherever our technology allows us to go, we'll be there.
Click to expand...

This is the main problem with most companies. There's no one there with a vision. I know I'm going to sound like a Steve Jobs fanboy. That's okay, because I am one. Designing products by survey and committee will never bring anything beyond mediocrity, except by accident. Steve Jobs started with the engineers. Usually, the way it worked is, the engineers would show him stuff that they could do, and he saw the potential in it and developed it. He notoriously whipped them into doing even more. Engineers like to come up with cool new stuff. The only thing they lack (usually) is salesmanship and perhaps the ability to polish their design. That's where Steve Jobs stepped in. Meanwhile, most companies are doing the exact opposite to their engineers. "Um, if you put a headphone on that, it will compete with this product. So, yeah, I'm gonna ask you to not do that. . . . Um, yeah, if you let flip this switch to let it compress only 10 to 1 instead of a million-to-one, it will be too high quality for this price range. We reserve that (lack of) compression for our $100,000 camera. So thanks. Maybe in 10 years we can use that idea."
 
Piolet said:
I think what they're saying is that they will listen to us (the market) and make products that fill our needs and wants. They seem very interested in feedback, hence that Hollywood facility which also serves as an R&D satellite..
Click to expand...
well, I don't know what they are are reading, or who's feedback they've been listening to, but the cries for RAW have been there on every video forum I've been on since the 5DMk2 was released. Panasonic actually listened and delivered. Sony listened and delivered two cameras at varying price points. Canon delivers the most expensive camera of anyone, late, overpriced, and under-featured. Was this the message they got? If they had delivered a 5D that over-cranked and had RAW for $5000 they would likely have sold every one they made.
 
combatentropy said:
"Only two years." So what's normal? Three or four years? People complain about RED taking a long time, but this is yet more evidence that Red isn't slower than others at moving from drawing board to assembly line, just more transparent.
Click to expand...

It depends on the complexity of the project and how much existing technology is reused. But it can take 3-4 years. Initial concepting, prototyping, human factors and mechanical packaging can take 6 months to a year. You go through several iterations at that time. The electronic design runs in parallel but it's closely tied to the mechanical design because of space claims. After that is a long period of software development, tweaking and testing. This can take 1.5 years or more and if you find a problem (which you will), design changes are needed and possibly another trip down the iteration cycle. Then there is a period of production readiness and supply chain ramp up. This can take up to 6 months.

Red is a bit slower if you consider their products are not really fully done when they ship. Not all features are enabled and there are still bugs being fixed by firmware updates that could last quite a while. Conventional product releases usually have all features enabled and the product is fully ready for volume production. Any firmware updates after release should only be for minor things. The product also usually ships 1 or 2 months after announcement and in large quantities. This means the production of parts was well underway before that and accordingly, the design locked down even earlier.

And no, I don't work for Canon :)
 
Last edited:
Piolet said:
they're constantly leapfrogging with Nikon to see who has the best technology.
Click to expand...

Gonna have to stop you on that one. Leapfrogging to see who has the best technology is not what Nikon and Canon do. They pretty much OWN the market, they keep in close contact with what one another are doing, and they purposely only release products with small, incremental improvements. It's how they keep consumers spending on the "newest and best"; it's what they call business...
 
heh ... I wonder who's the real winner on Nov 3rd ... because someone's gonna sell a **** load of lenses
 
PerroneFord said:
anon delivers the most expensive camera of anyone, late, overpriced, and under-featured. Was this the message they got?
Click to expand...
Has nothing to do with what "message" they got: it's just Canon's nodus operandi. Late, overpriced, and under-featured is exactly what they do and have always done. When Sony, jvc and Panasonic all went hd, canon brought out a standard def dv camera that cost more than any of their competitors DV cameras ever dared. When Panasonic went tapeless HD at $6000 Canon responded with a $9,000 hdv tape camera. This is their pattern. It's what they do. Panasonic and Sony come out with $5-6k large sensor cams, and Canon delivers a $20,000 camera with mpeg-2. It fits perfectly with their pattern and modus operandi. you could have predicted this based on their track record except for how quick they were. I honestly didn't expect them to make this move for another year. I'm sure the C300 is a fine product and it will attract some buyers, just like the XLH1 did. I mean, really, isn't this basically what we should have expected from them? A premium-priced interchangeable lens product using older technology - doesn't that describe their first foray into DV? XL1. Doesn't that describe their first attempt at 24p? XL2. Doesn't that describe their first stab at HD? XLH1. So why should their first large sensor videocam deviate from the pattern?If the pattern tells us anything it tells us that the camera you were all expecting will come, a large sensor xha1 will follow this large sensor xlh1. But settle in and wait a year.
 
So what was all the BS last year with the 4K "hairdryer" concept camera? If Canon was trying to demonstrate it could do 4K, why in such an ugly, unergonomic form factor? If they had no intention of following up with a 4K camera within the year, what was the point?
 
Who says they won't? 4k doesn't have to mean cinema, after all. They might make some manner of 4k product for broadcast in the future.
 
ryan brown said:
Gonna have to stop you on that one. Leapfrogging to see who has the best technology is not what Nikon and Canon do. They pretty much OWN the market, they keep in close contact with what one another are doing, and they purposely only release products with small, incremental improvements. It's how they keep consumers spending on the "newest and best"; it's what they call business...
Click to expand...

I don't know. I disagree. There was some pretty good competition in the early years. Maybe the technology has matured now so they can only do so much - hence the incremental improvements.

I love camera history, so a little trip down memory lane...

  • Before 1999, DSLRs were made by...Kodak! These Kodak DSC cameras body snatched Nikons and Canons, grafted huge digital backs to them and sold for an eye popping $28,000.
  • In 1999, Nikon launches the D1, their first in-house digital camera. It sells for an unheard of $5600 and is a huge hit. Canon has nothing.
  • In 2000, Canon hits back with the $3500 D30, which severely undercuts the D1 in price, but it's no pro model.
  • In 2001, Nikon improves the D1 with a D1H and D1X, which pushes resolution to an amazing 5 MP. The megapixel race is on and doubling in only a year.
  • That same year, Canon launches their first pro DSLR, the EOS-1D, which has a bigger sensor than Nikon's D1H/X and shoots at an amazing 8 fps. Nikon's D1H only manages 5 fps.
  • In early 2002, Nikon releases their first prosumer DSLR, the D100 which tops their own D1X in resolution with a 6 MP sensor and brings prices down to $2000. Canon adds the prosumer D60, but prices it at $2200.
  • In late 2002, Canon drops a bombshell: the EOS-1Ds. World's first ever full frame DSLR at an incredible 11 MP.
  • In early 2003, Canon follows up quickly with the 10D, which lowers prices again to $1500.
  • In 2003 Nikon releases the D2H, which uses their first in-house LBCAST sensor.
  • In late 2003, Canon breaks the $1K price barrier with the EOS-300 Rebel at $900.
  • In 2004, Nikon competes in the low end with the D70 for $999.
  • In late 2004, Nikon reclaims the highest resolution title with the 12.4 MP D2X.
  • But then 5 days later Canon takes back the title with the 16 MP 1Ds MkII
  • In 2005, Canon launches the first "affordable" full frame DSLR, the 5D. This is huge.
  • In 2007 Nikon launches the D300 and D3, which introduces high resolution screens to the back of the camera and amazing low light performance. The high ISO race starts.
  • In 2007 Canon puts Live View in their 40D. Now you can shoot your DSLR like a P&S. But this is really a precursor for what would happen next year...
  • In 2008 a new era begins...720P in the Nikon D90 and 1080P in the 5DMkII.
  • In 2009, Canon releases the 7D, their highest-end APS-C camera yet. It has full manual video control, frame rates 24P/30P/60P, manual audio, HDMI high res out. It out specs the 5DMkII and costs less.
 
Last edited:
Barry_Green said:
Has nothing to do with what "message" they got: it's just Canon's nodus operandi. Late, overpriced, and under-featured is exactly what they do and have always done.
Click to expand...
I will agree with you if you stipulate to the fact we are talking Canon VIDEO division. The stills division has done absolutely ASTONISHING work the past few years. But yes, this is exactly why I had zero faith in their video division. Each of the companies has done about exactly what I expected. Panasonic came in with a "kinda" cine camera at an excellent price point. Sony made a camera that competed in this space and one that a pro could really get to work with. JVC is wandering somewhere in the dark as usual, and Canon hit a homerun.... but in the wrong game.
 
Some additional info:

Diagram of how Canon forms a 1080p image from a 4K sensor(as mentioned in the OP). No debayering . From a 8.3MP sensor (4K2K) it captures 4MP of green, and 2MPs of both blue and red; has complete 1080p image in red, green, and blue. Over a 1,000 lines of resolution.

Cinema_EOS_046.jpg


Picture of built-in ND filter:
canon34.jpg
 
I remember talking to a Sony executive in Washington DC at a Vegas Pro meeting. Sony developed the sensors and innards of the FS100/F3 in 2 years. Totally new design. And most would agree, they hit the mark they were aiming for. So, the excuse of "only" 2 years, well....Ok, sure...what ever!
 
exactly, if you check how much a "packaged scarlet" will cost you (15-16k) and compare it to the canon (with a streetprice of 15-16k) you will find that red has serious competition, think about it:

canon gives you 1080p ON A S35 SENSOR where the scarlet gives you 2k on a 2/3 crop. As someone who is going dvd/bluray + tv + web only, which would be the cheapest (in total,after post) and easiest/fastest option? Dont forget that 1080p was always more than enough for silver screen, and now the canon has REAL 1080p, S35 + 8mp downscaled (properly) PLUS super low light powers :) (oh and the beatiful canon colours ^^, no fanboi :P)

i knida have the feeling that the scarlet is hmmm how should i put it: look, when you are brown (ma momma is white and daddy is black ;) ) whites call you black and blacks call you white - you dont belong to one group (or better: no group fully accepts you) this is for me the scarlet. If you tell me "i need the fidelity of the redraw, I ask you "cant you simply whitebalance properly, light proffesionally and color the image of a F3 the way you want it without any quality downgrades?"


in my eyes, there is only ONE ture "salvation" for the red = the "now" cheap and very fast (many up to 0.95) c-mount glas. Simply shoot 4k? For many, thats totally overkill and coupled with direct and indirect money loss (direct = "more" expensive hardware (production and post) and indircet = all the "time" thats lost with the extra work) that why you see many being totally happy with their panny af102 (i know... i know... different beast 5k and 15k are two diff. worlds, but here, my argument of "not belonging to one group" comes into effect, scarlet should habe been 6-8k or 16-20k)
 
Status
Not open for further replies.
Back
Top