July 26, 2014, 01:08:59 AM

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 - 3kramd5

Pages: 1 2 3 [4] 5 6 ... 21
46
Lenses / Re: Need a 600mm. Don't want to pay for one
« on: June 05, 2014, 11:45:33 AM »
    600mm for Zoo?? IMHO, 600mm is too long and minimum focusing distance is too long for Zoo unless the animals in your Zoo is kept very far away...

I agree that it's probably too long for a "animals in cages directly adjacent to walkways" style zoo. For the zoo I frequent, 600mm would be very beneficial (hence being on the backorder list for the Tamron, which is the best option in the price range I can justify)

47
But good lenses don't make a system either.


Correct, it is a pool of resources which make a system. On the subject of IQ, flashes and all that can generally be neglected since all that is strictly needed to produce a photograph are 1) a camera and 2) a lens.

Again, ignoring weirdly weighted scores, When comparing the huge volume of data accumulated over measuring 147 lenses, one very surprising result was revealed. The average sharpness scores of the Canon EOS 5D Mark III matched the Nikon D800 and if the results were based solely on the mean average, the Canon actually out-performed the Nikon.

When using specific lenses (such as the new Sigma 35mm f/1.4 DG HSM A) the Nikon can out resolve the Canon. However, taken as a whole, the statistics reveal the EOS 5D MK III is capable of similar sharpness and of achieving a close DxOMark camera/lens score to the Nikon D800. Moreover, that’s despite the latter camera’s 60% extra pixel count.


So, while the sensor Nikon sources for the D800/E/S is undeniably more capable than that Canon builds for the 5D3 (both cameras representing the highest resolution offering from both makes), the lenses for the Nikon are, on average, undeniably poorer performing than those for the Canon.

Within its system, Nikon doesn't yet offer a wide variety of lenses that are a match for its highest resolution body.

Within its system, Canon does offers a wide variety of lenses that are at least good enough for its highest resolution body. Whether they're good enough for a 50% pixel increase has yet to be determined. If and when Canon makes the business decision to enter the higher-res market (and I suspect their lens refresh is a milestone in that direction) we shall see.

Assuming canon releases a 35+ MP body and the lenses resolve that well, it will trounce the D800 in detail. Canon fans will trumpet. Nikon fans will likely say "yah but you need to spend $20,000 for a variety of series II lenses for it." I'll sit back with 23MP and be glad I don't need a metric ton of flash memory and a supercomputer to process it all.

48
The question is why has Canon not released a 36MP'ish FF camera when both Nikon and Sony have done it.

It's a business decision.

So, if according to you, the ability of the processor(s) to handle the data is not the issue, then what is the excuse? Crummy lenses without the resolving power? No. Slow SD and CF cards? No. Inability to actually make such sensors? No. What then?

Their market analysis, likely.


To those chucking Moore's Law around ... yes, possibly the DiG!C architecture can handle the data, but at what cost in terms of power consumption and thermal management? Both these factors can negatively impact on the actual use of the chips ... which leads to the same conclusion.

As has been noted multiple times, canon already churns through 18MPx2 @ 12-14FPS in a two year old body. Clearly processing a 36MP image is NOT beyond their past architecture, so assuming that it's beyond their future architecture is pretty ridiculous.

Canon is very calculating. They won't bring a high dollar camera to the market until a) their infrastructure (lenses, primarily, and I suspect they're there now) is ready, and b) the market can support the investment.

49
What about video?
The 1Dc (same processors as the 1Dx) does 4K video at 24p and 25p
4k = 4096 x 2160 = 8.84mp at 25 frames per second…

Obviously the setup of dual DiG!C 5+ processors with an additional DiG!C 4 clapping hands is sufficient regarding the 1DX and 1DC cameras. That is not the issue. The issue in question is whether this setup is sufficient for what the future brings ...

Why is that the question? I still can't quite get why you are evidently assuming that digic5+ is the best core canon can make. The issue in question is whether what the future brings is sufficient for what the future brings, not whether what they released 2 years ago is.

I never said or implied DIGIC 5+ is the best they can make

I was replying to Sella ;)

50
Who says they aren't?

I think the noise (haha) over a stop or two of additional shadow recovery is fairly inconsequential to canon.

Would it be cool for them to bring in 24-bit ADC? Yah sure, but at the end of the day, most people look at screens with maybe 8 stops of DR, or prints with less. When the recording device is already capturing significantly more range than the end product is capable of displaying, photographers still need to employ methods to compress the DR of a scene, either before it hits the film/sensor (nd filters, lighting techniques, etc), or before it hits the print (exposure stacking, etc).

Having more information to work with when we enter post would be welcome, but the number of instances in which someone wants to show the words on a newspaper under a picnic table in broad daylight as well as the details of the clouds in the sky above it are pretty rare.

That said, I also welcome DR improvements because of what it means for noise.

51
What about video?
The 1Dc (same processors as the 1Dx) does 4K video at 24p and 25p
4k = 4096 x 2160 = 8.84mp at 25 frames per second…

Obviously the setup of dual DiG!C 5+ processors with an additional DiG!C 4 clapping hands is sufficient regarding the 1DX and 1DC cameras. That is not the issue. The issue in question is whether this setup is sufficient for what the future brings ...

Why is that the question? I still can't quite get why you are evidently assuming that digic5+ is the best core canon can make. The issue in question is whether what the future brings is sufficient for what the future brings, not whether what they released 2 years ago is.

52
If they release a digic 6+ with the same kind of performance gains the 5+ saw, it will be very capable.

True, and then I would have been proven wrong. So that is why I am very interested to see what Canon puts inside the upcoming 7D2 camera.

As I recall, they didn't develop a new processor for the 7D, but rather used the same configuration as the 1D4. So maybe a new 7D would use the same config as the current pro sports model (dual digic 5+). However, those cameras came out really close to one another so that could have simplified the decision.

They may want to use a 7D as a test bed for new tech, or they may want to use the economy of scale and leverage of the existing production infrastructure.

In any case, my suspicion is that getting a capable core is the lowest hanging fruit in the higher resolution tree. Developing a line of great lenses is the biggest hurdle, and they've crossed it. The sensor unit is somewhere in the middle, and that's the biggest question in my mind from a tech standpoint, whereas the market is the overarching driver in all of this (if they don't think they can sell sufficient higher res cameras, they won't bother making them).

53
The fact that in a camera developed for release two years ago they didn't use a modern core indicates they don't have a modern core now?

The very fact that there are more capable processors than the (old'ish) DiG!C 5+ can only mean that Canon is behind ... either in terms of actual know-how or due to patent constraints. So it will be very, very interesting to see what processor(s) are placed inside the 7D2.

To me it merely means camera cores developed >2 years ago are behind current cores, but that's an obvious reality.

It doesn't give me any indication as to what Canon will produce for future models (i.e. cameras developed <2 years ago). If they release a digic 6+ with the same kind of performance gains the 5+ saw, it will be very capable.

54
Would be an idea for the Magic Lantern development - if one could identify where everything was saved...

Yah, that's what got me thinking about it. The rule of thumb is to reset your camera to default settings before installing ML, so I was hoping I could offload my settings somewhere first.

55
Canon can do dual processors even in a half sized body (7D), ...

Canon already needs to dual-configure these little rascals to produce a decent frame rate with 18MP sensors.

That was a camera developed for release in 2009. They probably don't need dual processors to do the same in 2014 (the pixel throughput of the 5D3 is nearly 93% of the 7D even though the former only has one processor).

Also, can these DiG!C 5+ or 6 processors even be quad-configured?


Good question. I imagine they could fit four in a full sized body. The PCB might get a little thick, and heat would likely be a major factor, but it's probably doable. I doubt it would be necessary unless they wanted a super high resolution super high frame rate $20,000 market entry. 

... and they can do it with modern cores.

True, but the big question is whether Canon has any "modern cores". The very fact that they have not used it in the 1DX tells me that the answer is a very sorry no, they don't.

Wait, what?
The fact that in a camera developed for release two years ago they didn't use a modern core indicates they don't have a modern core now?

56
bah, humbug.

57
Just once digic5+ is good for at least 180MB/sec (5D3 raws at 6FPS). Given Moore's "law," it would shock me if something like the core logic chip fundamentally precludes higher-resolution cameras.

That's the transfer rate of data, i.e. throughput, of already processed data. What is important is the number of instructions per second that the little guy can do ... and then how many instructions are required to process the data from the sensor's A/D-converter. Not stuff that Canon gives out readily, but it can be deduced from current camera models.

Yes, it's data rate. In any case, one digic 5+ is good for at least 6 23MP captures per second, soup to nuts, all operations. Maybe it doesn't scale linearly, i.e. perhaps it's not safe to assume one can do 3 46MP captures, but I bet it can do 1.5-2. And that's with a single 2+ year old chip. Canon can do dual processors even in a half sized body (7D), and they can do it with modern cores.

58
My conclusion was that Canon cannot really produce a "High-MP" camera (or a decent mirrorless camera) for the very simple reason that the DiG!C processor has hit the wall.

Just once digic5+ is good for at least 180MB/sec (5D3 raws at 6FPS). Given Moore's "law," it would shock me if something like the core logic chip fundamentally precludes higher-resolution cameras.

59
   The funny part is that DPReview said that "Nikon’s 800mm F5.6 FL ED VR and 400mm F2.8 FL ED VR lenses earn the 'FL' in their names from the coating of fluorine applied to the lens elements."

Fluorine, fluorite…what's a calcium ion among friends.   Actually, the 'fluorine coating' is a clever bit of marketing-speak…fluorine is a highly toxic element that's gaseous under standard conditions, so the 'fluorine coating' certainly isn't just fluorine.  It's most likely a polymer of fluorine and carbon…but neither Canon nor Nikon want to say they're using fluorocarbons as lens coatings…   :o

Can't say I blame them. Consumers are easily swayed by words and phrases to which they have a predisposition. Note how almost everyone drops the "RP" from "CFRP" because everyone likes Carbon Fiber but nobody likes (Reinforced) Plastic.

60
Does anyone know if it's possible to back up personal settings for non 1D-series cameras?

Pages: 1 2 3 [4] 5 6 ... 21