November 23, 2014, 08:14:26 PM

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - John Thomas

Pages: 1 2 3 [4] 5 6
46
EOS Bodies / Re: Canon EOS-1D X Delays [CR2]
« on: June 02, 2012, 05:25:15 AM »

I had in mind that they could try NOT to put a "last-minute improvement" (significant or not) but to include a "big" improvement which they have behind the closed doors - something like a BSI sensor - they have the technology...


A BSI sensor, if it ever be true, will place Canon at the same starting line as Nikon/Sony, and should be worth of the long bitter waiting period...

BSI primarily benefits sensors with small pixel pitch, and the aim is to increase QE (quantum efficiency), which only might affect noise levels as more of a byproduct of its primary job. (BSI basically "flips" the fabricated sensor upside down, putting all the R/C activation and readout wiring behind the photodiode, and therefor out of the light path from the pixel/microlens.) The pixel pitch on the 1D X sensor is relatively very large, and it would benefit little from a BSI design. The D800 does not use a BSI design either...it is still FSI.

The difference between Sony cmos sensors and Canon cmos sensors is that Sony integrates FAR more hardware-level noise reduction technology than Canon. Currently, to my knowledge, Canon sensors only employ CDS, Correlated Double-Sampling...however I believe their patents date back nearly a decade. Sony sensors employ a newer and more effective form of CDS, a form of transistor differential compensation to reduce FPN, integrated column-parallel ADC (a smaller, slower ADC for every column of pixels built right into the sensor...slower ADC's produce less noise of their own, and having one ACD per column also helps reduce FPN), and a few other smaller improvements that I currently can't find the patents for. Its these explicit noise reduction features that make a Sony Exmor sensor produce cleaner pictures than a Canon sensor.

Canon could benefit from a BSI sensor in their compact and bridge cameras, but the improvement to QE in a large sensor with a very large pixel pitch like the 1D X would be very small...maybe 1-2%...definitely not enough to put them in the same league as a Sony Exmor. (It should be noted that Canon uses a gapless "microlens" sensor design...but the pixels themselves still have gaps between them...most of the activate and readout wiring exists within the spaces of the gaps, with minimal intrusion into the light path from a pixel. If this were not the case, as might indeed be the case with a very high density full-frame camera (say 60mp or more, the 2-3 micron pixel pitch range), then a BSI design would probably benefit a full-frame high resolution sensor as much as it benefits a tiny point and shoot sensor.)

Your post needs more acronyms  ;D

Nope. It was quite informative as it was.  :P

47
EOS Bodies / Re: Canon EOS-1D X Delays [CR2]
« on: June 01, 2012, 03:45:42 AM »
Given the target market, the price and the fact that their reputation in the professional market rides on this camera, is it any wonder that Canon wants to make sure the thing is as close to perfect as possible before releasing it?

If they think this, they must be really desperate - engineering is always a compromise between dev time and product lifecycle, thus making "near perfection" an idea for geeks, not for a commercial company. It's much more likely that they found flaws (f8 af?) they couldn't get away with given the competition, so rather than "near perfection" its more like "good enough".

1.) By improving 1DX to not leave the impression that 5D3 is abandoned.
Abandoned customers will abandon the brand.

Are you suggesting that Canon would be well-advised not to put any significant last-minute improvement in the 1dx, because it may make the 5d3 look bad? Geez, this is the way of technology - and Canon will hardly stop development because some 5d3 users are a bit touchy when they hear "Nikon" and could not stand internal Canon competition, too. Didn't tons of 5d3 threads establish that the 5d3 is a great body at any price and nothing else matters, esp. not other alternatives :-o ?

I had in mind that they could try NOT to put a "last-minute improvement" (significant or not) but to include a "big" improvement which they have behind the closed doors - something like a BSI sensor - they have the technology, put the link bellow in the Google Translate to see:

http://egami.blog.so-net.ne.jp/2012-01-23

...also have a look here:

http://photorumors.com/2012/01/23/sony-develops-next-generation-back-illuminated-cmos-sensor-with-rgbw-coding-and-hdr-movie-function-canon-files-a-patent-for-a-large-bsi-sensor/

For an expensive product with a long life cycle like 5D3/1DX the early adopters are one of the most valuable assets of a company. It they will feel betrayed then they will betray - a damage which is very hard to repair after. I know this from many years of experience.

Having a camera being twice as expensive and two or even three times better it is understandable. But having a camera ten times better just because they wanted to fool the early adopters and enthusiasts with 5DMk3 - this will backfire. Too few will think that they had a stiff competition from Nikon/Sony and they needed to throw everything against them. The ones who purcased an comparatively over-priced 5D3 will feel cheated and next time who will dare to commit?

OTOH, from my small experience, as I stated elsewhere and it seems that I'm not alone, they now try to overcome some design issues and to improve to the max the firmware (this includes also end-user features like black AF thing but also the internal processing engines like ADC / noise handling etc.). The risk is way to big (imho) to redesign hardware now.

But I think that's critical for them to crowdsource their problems even in the form of "rumors" (this is one the best ways to keep the brand goodwill untouched) because more eyeballs will see better and perhaps a solution will appear from "nowhere" (if it didn't appeared already ;) ).

JustMy0.02c++

JohnTh.

48
EOS Bodies / Re: Canon EOS-1D X Delays [CR2]
« on: May 31, 2012, 03:15:12 AM »
Yes - the most possible scenario is that their "hardware failure" is called Nikon. :)

However, they must take care at some things:

1.) By improving 1DX to not leave the impression that 5D3 is abandoned.
Abandoned customers will abandon the brand.

Also, it will create a pretty nasty chain reaction which will tarnish even more the Canon's reputation which, truth to be said, isn't at its heights nowadays. Nothing worse than the word of mouth from a burned customer.

2.) Improving 1DX in such small amount of time is a sensible matter. But this has two planes: hardware and software (ok, firmware).

Doable but they must be very careful at a big dark trap which any Canon engineer should know: if they do a mistake in firmware and this is found after the product is in the field, this is easily fixable (the cost is low) but if they do a mistake in hardware and this is found in the field then this can be the final tombstone.

Generally speaking, in this situation, significant improvements in firmware has by far the biggest return of investment and they should give much more importance to this now.

However, I think that they are forced to mess up with the hardware and that's why 1DX is delayed.

0.02c++

JohnTh

49
EOS Bodies - For Stills / Re: Odds & Ends [Various]
« on: May 31, 2012, 02:43:54 AM »
Quote
Download the firmware, copy to card, insert card in camera, go into firmware in camera and upgrade, just make sure your battery is 25% or more to last 2-3 minutes. Voila, its done. So easy a caveman can do it.

Yes, we are cavemen 8) - but I think that  MazV-L raised an interesting point: The attitude of (some) Canon employees which in this case is acting like a brake in cavemen's approach to the fire (with all the meanings of this metaphor).

However, I think that Canon should improve the process of learning the phtographers to be more inventive and to have an attitude of continous improvement, knowing that they will receive something as a reward (ie. a better firmware with bugfixes, features etc.).

The product lifecycle for a pro/flagship camera is big and, price drops/rebates aside, only through firmware improvements they will keep the sales up and the customers close.

just my2c,

JohnTh

50
EOS Bodies - For Stills / Re: Odds & Ends [Various]
« on: May 30, 2012, 01:52:55 PM »
Okay, so my 5Diii arrived yesterday, complete with Firmware update 1.1.2 and light leak and DPP software fix. So is there going to be yet another Firmware update so soon :o

God bless them for these! Only to continue in this manner...

Well, perhaps is better to stress that the reality has changed. Nowadays cameras are much more complex than any human (closed) team can handle. The usage cases can be much different from what the design team imagine and the interactions between different camera subsystems are much more complex than several years ago.

Others have also some problems, too. See, for example:

http://www.dpreview.com/news/2012/05/29/Nikon-updates-firmware-D4-D800-D800E-
(see also the comments)

http://nikonrumors.com/forum/topic.php?id=5517
(scroll down to see how many links they post)

I think that's the time to accept the reality and try to improve it. Ok, it would be better to have everything perfect from the beginning - but this just isn't possible practically.

just my2c,

JohnnTh
Having to send the 5Diii to Canon every time there's a firmware update will be a real pain in the neck, but that is what they recommend  :(

?!?!

Why??

You can update the firmware by yourself. The proces is quite straightforward, as you should know, - just to have a changed battery and patience to not touch the camera till the firmware update process is finished.

There are plenty of photographers which download the firware and apply by themselves.

HTH

John Th.

51
EOS Bodies - For Stills / Re: Odds & Ends [Various]
« on: May 30, 2012, 04:05:15 AM »
Okay, so my 5Diii arrived yesterday, complete with Firmware update 1.1.2 and light leak and DPP software fix. So is there going to be yet another Firmware update so soon :o

God bless them for these! Only to continue in this manner...

Well, perhaps is better to stress that the reality has changed. Nowadays cameras are much more complex than any human (closed) team can handle. The usage cases can be much different from what the design team imagine and the interactions between different camera subsystems are much more complex than several years ago.

Others have also some problems, too. See, for example:

http://www.dpreview.com/news/2012/05/29/Nikon-updates-firmware-D4-D800-D800E-
(see also the comments)

http://nikonrumors.com/forum/topic.php?id=5517
(scroll down to see how many links they post)

I think that's the time to accept the reality and try to improve it. Ok, it would be better to have everything perfect from the beginning - but this just isn't possible practically.

just my2c,

JohnnTh

52
Quote
7D black worked comprehensively better for me than the 1D red yesterday in the bright sunlight

This is yet another reason for which a blinking solution (black-red-black) of the AF point is better.

53
There are two paths which I see now:

1. Keep the AF point(s) lit and adjust for the amount of (over)exposure. Complicated, because there are many lookup tables to combine: many AF points, many AF modes (pinpoint, single, expansion etc.) and many exposure programs. However if the amount of light which each red square emits is additive (ie. the amount of light of 2 (adiacent) AF points is double than one AF point), or obeys to a single simple rule, then the things are fixable.

2. Alternate the phases: In fact, what means "AI Servo" - or rather "continous" exposure?

a.) Turn down the AF points (they will be black)
b.) Get the exposure
c.) (After, let's say) 300 ms turn the AF points on (they will became red)
d.) Keep them red for another 300 ms
e.) Go to a.)

The net effect is that the AF points will blink red while the metering is done inbetween.

Focusing should be done in the best case in parallel. Anyway It is not affected by the light in the way in which we discuss it here.

The second solution will have the advantages that A.) we'll see the camera is "working" (focusing, metering etc.) because the AF points are blinking and B.) we'll see also the subject between the AF point's blinks (here perhaps is better to turn off completely the AF point).

Thoughts? Comments?

HTH

54
EOS Bodies / Re: Canon to Announce 2 Cameras in June [CR2]
« on: May 17, 2012, 04:46:19 AM »
But the question is:

It will be better than Sony's???

http://photorumors.com/2012/05/13/sony-a99-specs/

    The Sony A99 camera will have 24.3 MP full frame sensor.
    The A99 will have 102 focus points.
    Two different A99 prototypes are being evaluated - one of them has a flip and touch screen.
    The A99 will have 3 million pixel high-definition EVF.
    To be announced in September 2012.
    Price: 18,000 yuan (around $2,800).
    The A99 will  support 24 video recording formats and a variety of screen style filters.

55
It seems that some people in this thread are looking at the wrong camera. This camera is designed mainly for still photography, NOT video.

Video is just a small "side bonus" on the 5D3, so don't be so surprised if Canon doesn't give you the features you want. They are saving all the good stuff for the 1DC-line. That is where you should be looking.

For video when compared to the D800, it lacks several major / almost crucial features.
Same as with the still photography features.  ;)

Alas, it is a shame that most of the missing features can't be added by firmware. Like more resolution, more DR, and spot metering linked to selected AF point. (Just to name a few).

Spot metering linked to selected AF point can be done, I think. Evaluative metering is linked to selected AF point.

Also, regarding of DR I remember now Nikon D80. In its initial release there were an awful noise at high ISOs. But later they corrected that through firmware.

Ok, I don't think that they can do wonders, but I know that they can tweak the DiGIC registers in order to obtain better Dynamic Range:

Quote
The surprise comes from a little ISO test, which shows that ML can
improve dynamic range and highlight rolloff in movie mode (see
attached files). How much? Let's find the answer together by running a
little experiment.

The above quote is snipped from

https://groups.google.com/forum/?fromgroups#!topic/ml-devel/PK1SX33ewRg

Before putting my wishlist, I think that's quite important to understand that we must give as much feedback as possible (in quantity and quality) in order to make from 5DMk3 the best camera around. And I know that this is possible because the best camera is the one which will allow us to take the best shots not the one which looks best on paper. 

Look at PC / Win vs Mac. Wins the one which helps you better to get your job done.
I can easily prove that "Linux is better" (no flamewars please - thank you). But how many of us have Linux on our computer?

We need to make from 5DMk3 the tool which fits our needs. And for this we must communicate our needs firstly not mainly our solutions. Think outside the box. Sometimes helps.


56
EOS Bodies / Re: 5D3 IQ FB, do I need a D800?
« on: May 08, 2012, 04:23:41 AM »
For your needs D800 is way too low in megapixels.

You might try an H4D-200...

http://www.hasselblad.com/products/h-system/h4d-200ms.aspx

...but I think that neither Hassy will satisfy you  :D

Seriously, tough, I think that at this moment on 5D Mk III we (Canon + community) should concentrate in finding concrete ways to improve it. And I think that the areas are:

- better SNR from sensor (I dunno if this is possible anymore via firmware - but definitely they should try this)
- better customization of camera (imho, they should put more options on more buttons - there are enough options which are available only on one buttons and other buttons are for many just "waste of space")
- ...perhaps they must do something to make the life of camera programmers (Magic Lantern, CHDK etc.) easier? (there are enough ways to do this but first I want to see what's your opinion)

Any other improvement?...

57
For example the EF 24-70 f/2.8 is in Group B while EF 24-70 f//2.8 II is in Group A. Why? And how can I figure it out?

A concrete question: The last Tamron 24-70 f/2.8 VC in which group is in? A certain Sigma f/2.8 lens? Tokina? (in few years) a new Canon lens?

(snip) ....Those will have to be empirically determined.


Yes, this is the gist of the problem. And (AFAIK) we cannot. By pressing the AF chooser (the right-upper corner button) we can conclude for some groups only, based on the fact that for these groups the layout of the Horizontal AF points (which are blinking) is different.

But what can we do for the groups in which only the Double-Cross points are different?

...and now I'm thinking that, being an event / photo journalism / sports AF system, exactly these points can make the difference....

58
Thanks a lot for the info and your intention (seriously) but, in fact, is the same info from the 5DMk3 manual. The problem is that, while there are some hints about aperture, macro and such, there isn't a clear rule to distinguish the groups between them. There doc is formulated with "Most large-aperture lenses...", "Some lenses..." etc., hence one can only guess a new lens in which group will be.

For example the EF 24-70 f/2.8 is in Group B while EF 24-70 f//2.8 II is in Group A. Why? And how can I figure it out?

A concrete question: The last Tamron 24-70 f/2.8 VC in which group is in? A certain Sigma f/2.8 lens? Tokina? (in few years) a new Canon lens?

This is my problem.

Ok, the best solution would be to have the AF engine render exactly the types in the ViewFinder (ie. including the Dual Cross-Type points) and someone publish them.

But thanks a lot anyway.


59
Hi,

In 5DMk3's manual (see pages 78-84), Canon divides the lenses in 9 groups based on how many Dual Cross-Type (if any), Cross-Type and Horizontal AF points a lens can exploit from the total of 61 AF points available.

The charts there are very nice, also the VIewfinder's blinking feature for the Horizontal AF points is ok, but how do I know in which group a lens belongs to, and I mean especially a lens which I do not own yet, a lens which I intend to purchase?

Of course, I mean lenses which aren't mentioned in the tables from the manual, either because are not from Canon either because are newer.

(For the nit-pickers: the aperture isn't sufficient to determine the group. For example there are f/2.8 lenses in groups A, B, C, D, E)

TIA

60
Perhaps someone can enlighten me?

CR's take on the 24-70ii was that Canon had several prototypes, but the one(s) with IS were too heavy or the iq tradeoff was to large to get the weight down. And Canon simply sells stuff for as much money as they can get away with - and they should since I cannot afford it anyhow, better release the "good" stuff at a reasonable price and sell the "stellar" stuff at a premium. For a pro who's in business 1000$ is not a lot of money for better results, and nothing compared to the cost of good tele lenses. When did anyone complain about the price of the 70-200?

Yeah, I'm aware about Canon's strategy. Thanks a lot, anyway. But now the things are changed. And are changed in a significant matter, imho. There IS a viable alternative at a much lower price.

And not a viable alternative (imho) but one which has an important feature which Canon didn't manage to make it - so the price difference doesn't justify. Not the price alone.

See, in the case of of 70-200 (esp. in the case of 70-200 II) there aren't other alternatives which can compete with much success (ok, Sigma perhaps). And hence Canon can settle for a price which market can bear. The same is in the case of, for example, L 8-15 fisheye. It is worth $1400? But since Tammy's & Sigma offerings cannot compete optically and feature wise (or, better to say that the others' offerings are nonexistent), Canon has that price.

But in the case of newly released Tammy's 24-70 VC the things are different. I think that the price / performance ratio is in favor of Tamron by a wide margin, even compared with Canon's 24-70 II if the Canon will keep the same price.

0.02c++ & HTH,

John Th.

Pages: 1 2 3 [4] 5 6