p.1 #1 · Limited evidence mystery camera (5DX) is 30MP+
Ok, I'd like to start by saying I am not one to speculate on these new camera threads. I usually sit back and read with amusement all of the wild theories people post.
That being said, the recently leaked "5DX" camera images are of high enough quality that we can make some calculations of total sensor megapixel count through simple mathematical relation of the image on the screen to what we believe is a reasonable screen resolution canon may use for this new model. Why are we able to do this with some limited accuracy
The photographer has the guinea fowl on screen at 100% magnification and the size of the zoomed window relative to sensor size is clearly demonstrated on screen!
Here is what we find:
Screen size from African spy photo shot = 49 pixels on X axis by 49 pixels on Y axis. This gives an approximate screen area of 2,401 pixels.
The zoomed sub window is (and this is approximate to within one pixel) 6 pixels on the X axis and 4 pixels on the Y axis. This gives an approximate screen sub area of 24pixels or 1% of the total sensor sample area.
If we further hypothesize that the LCD screen used in the next generation of camera is the standard 902k dot 640x480 (and not a new 1280x720 version), this ratio provides a final sensor resolution of APPROXIMATELY 6,400 x 4,800 or 30.7MP. I urge everyone to take this with a huge grain of salt but it is an indication that this new camera supports a higher resolution than the 1DX. This calculation is only valid if the bird is at 100% on screen. any zoom amaount lower than that only increases the megapixel count.
It's a full res capture of the spy shot, based on the orginal RAW. I adjusted the exposure, streightend and croped it. Now... redo the math
Maybe it's the perspective look, but the LCD screen looks "wide". It could be the same LCD in 60D (1040k)
Thanks for posting this image since my upload is not enabled. Another way of thinking about this is to calculate the zoomed windows % of coverage of the larger total image area. Really the perspective of the photo doesn't impact what total area the 100% zoom covers. I calculate that this 100% zoomed in area is approximately 1% of the total image area. That is sufficient information, assuming we know the screen to be 640x480 (307,200 pixels) to work out total image area at 640x480x100 = 30.7MP
p.1 #9 · Limited evidence mystery camera (5DX) is 30MP+
John Shultz wrote:
Thanks for posting this image since my upload is not enabled. Another way of thinking about this is to calculate the zoomed windows % of coverage of the larger total image area. Really the perspective of the photo doesn't impact what total area the 100% zoom covers. I calculate that this 100% zoomed in area is approximately 1% of the total image area. That is sufficient information, assuming we know the screen to be 640x480 (307,200 pixels) to work out total image area at 640x480x100 = 30.7MP
Yeah, but if it's 2% then 640x480x50 = 15MP, or 1.2% then 640*480*83=25MP, or....
p.1 #11 · Limited evidence mystery camera (5DX) is 30MP+
John Shultz wrote:
Thanks for posting this image since my upload is not enabled. Another way of thinking about this is to calculate the zoomed windows % of coverage of the larger total image area. Really the perspective of the photo doesn't impact what total area the 100% zoom covers. I calculate that this 100% zoomed in area is approximately 1% of the total image area. That is sufficient information, assuming we know the screen to be 640x480 (307,200 pixels) to work out total image area at 640x480x100 = 30.7MP
I am also "Upload disabled" but I used ImageShack.
I thought about your reasoning and I think you have nothing to sustain the 1% ratio. Think about 50D, 7D and 5D2 - they use the same LCD, but the similar ratios range from 1.46% to 2%. 50D and 7D share the same sensor format and the same LCD, but the ratio is largely different.
p.1 #12 · Limited evidence mystery camera (5DX) is 30MP+
Ciprian: John's 1% ratio refers to his estimate of the area of the white dot compared to the white box around it.
I did my own calculations on this image. I assumed a white dot width of 8 pixels (I cannot make it 6, John; if anything it might be 9!), a white box width of 74 pixels, and an LCD display width (on the camera) of 720 pixels. Thus the figure I arrived at was 66 megapixels. From that I assumed the photographer must have been zoomed-in to 200% to check focus (do Canon SLRs allow that?), which would put the camera at 17 megapixels.
If this image were twice as large we could make more accurate deductions from it. As it is, it's not very useful.
p.1 #13 · Limited evidence mystery camera (5DX) is 30MP+
John Shultz wrote:
Ok, I'd like to start by saying I am not one to speculate on these new camera threads. I usually sit back and read with amusement all of the wild theories people post.
That being said, the recently leaked "5DX" camera images are of high enough quality that we can make some calculations of total sensor megapixel count through simple mathematical relation of the image on the screen to what we believe is a reasonable screen resolution canon may use for this new model. Why are we able to do this with some limited accuracy
The photographer has the guinea fowl on screen at 100% magnification and the size of the zoomed window relative to sensor size is clearly demonstrated on screen!
Here is what we find:
Screen size from African spy photo shot = 49 pixels on X axis by 49 pixels on Y axis. This gives an approximate screen area of 2,401 pixels.
The zoomed sub window is (and this is approximate to within one pixel) 6 pixels on the X axis and 4 pixels on the Y axis. This gives an approximate screen sub area of 24pixels or 1% of the total sensor sample area.
If we further hypothesize that the LCD screen used in the next generation of camera is the standard 902k dot 640x480 (and not a new 1280x720 version), this ratio provides a final sensor resolution of APPROXIMATELY 6,400 x 4,800 or 30.7MP. I urge everyone to take this with a huge grain of salt but it is an indication that this new camera supports a higher resolution than the 1DX. This calculation is only valid if the bird is at 100% on screen. any zoom amaount lower than that only increases the megapixel count.
Thanks. I was thinking of doing something like that, but I figured I'd sit back and let someone else bother with it all and I also wasn't sure about it since the tiny white sizing dot was so small that I had doubts about precision and being able to tell 18MP from 30MP.
Anyway if you are correct, and say it is 32MP, then with eosfun saying that one of these was wrong:
22MP
6fps
19 pt AF with digic 4 assist
well that would give us 32MP, 6fps, 19pt digic 4
the only think that doesn't match is why would eosfun say many might keep their 1Ds3, who would keep a 1Ds3 over that? maybe the AF is just a feeble 7D update still much worse than 1 series or maybe he simply was thinking of those few for who weather sealing and 100% VF are simply critical at all costs? maybe it is simply plain old 7D AF and the digic 4 AF processing assist was a minor point he simply wasn't commenting on (7D AF would be a bit disappoint IMO since I feel the 5D2 center point actually works better as often as not, other than that the specs seem fantastic though)
this body would leave them matching Nikon res and the 6fps would beat D800 even and be just enough to be useful and yet not quite so much as the 7/5fps of the CR/aqua rumor as to hit the 1DX hard.
it fits when he said the canon marketing talk said it will make the 5D stills landscape/wedding crowd happy while bring in 7D goodness which a 32MP, 6fps with 19pt AF would do.
But then again this probably can't the specs because
30MPish, 6fps, between 7D and 1DX AF, moire-free video are the exact specs I've been whining about for them to put in the 5D3 for the last 2-3 years and when have they ever listened to my crazed forum rants?
p.1 #15 · Limited evidence mystery camera (5DX) is 30MP+
John Shultz wrote:
Ok, I'd like to start by saying I am not one to speculate on these new camera threads. I usually sit back and read with amusement all of the wild theories people post.
That being said, the recently leaked "5DX" camera images are of high enough quality that we can make some calculations of total sensor megapixel count through simple mathematical relation of the image on the screen to what we believe is a reasonable screen resolution canon may use for this new model. Why are we able to do this with some limited accuracy
The photographer has the guinea fowl on screen at 100% magnification and the size of the zoomed window relative to sensor size is clearly demonstrated on screen!
Here is what we find:
Screen size from African spy photo shot = 49 pixels on X axis by 49 pixels on Y axis. This gives an approximate screen area of 2,401 pixels.
The zoomed sub window is (and this is approximate to within one pixel) 6 pixels on the X axis and 4 pixels on the Y axis. This gives an approximate screen sub area of 24pixels or 1% of the total sensor sample area.
If we further hypothesize that the LCD screen used in the next generation of camera is the standard 902k dot 640x480 (and not a new 1280x720 version), this ratio provides a final sensor resolution of APPROXIMATELY 6,400 x 4,800 or 30.7MP. I urge everyone to take this with a huge grain of salt but it is an indication that this new camera supports a higher resolution than the 1DX. This calculation is only valid if the bird is at 100% on screen. any zoom amaount lower than that only increases the megapixel count.
Anyway I seriously doubt more than 32MP so we will assume the bird is at 100% zoom.
If the screen is 1280x720, which seems like something they might do, then the MP count is only 8MP though and that seems a bit unlikely to say the least. Or is it....
So I think we can eliminate the bird not being at 100% and we can probably even eliminate a new 1280x720 LCD since even if you count is some degree off it would have to be wayyyy off to make 1280x720 lead to anything remotely reasonable.
Hmm another calculation though is this 49 pixels across and dot is 6 pixels across so (((49/6) * 640)^2)*0.66666666 = uh-oh the same old 18MP FF sensor from the 1DX....
So it is 18MP FF, 6fps, with 19pt AF from 7D (perhaps with digic 4 to make AF better than 7D but still well worse than 1DX since it uses inferior AF sensor quality and larger points)??
That wouldn't be too bad if not for the D800. But 36MP FF at 4fps with a 16MP APS-C at 6fps option and Nikon's best AF effort does sound somewhat more enticing I think.
p.1 #17 · Limited evidence mystery camera (5DX) is 30MP+
skibum5 wrote:
Hmm another calculation though is this 49 pixels across and dot is 6 pixels across so (((49/6) * 640)^2)*0.66666666 = uh-oh the same old 18MP FF sensor from the 1DX....
But the photo is in portrait orientation, so your 0.66 should be 1.5.
That's why I considered the possibility of a 200% view. I have Nikons which do this. Do the existing Canons?
Also, the LCD is bound to be 720 x 480 pixels, like the 1D X.
p.1 #19 · Limited evidence mystery camera (5DX) is 30MP+
(I quickly tried something like this with my 5D2, very quickly and sloppily, and got 30MP for my 5D2 yay!!!)
More (or maybe less) seriously:
OK I just tried it on the '5DX' and get 69-70 across and 8-9 across for something like a 20-22MP total....
Not anything like 30MP....
OTOH if I base it off the other dimension, I get some crazy high MP, truly absurd stuff so I don't think this will work out very well or be able to tell us much.
.
