Canon 1DSMK3 test image compared to Phase Backs

[jjj]

There's something here that doesn't seem quite right, Frank. Having myself used a 6mp D60, an 8mp D20 and and a 12.7mp 5D, I have not been struck by any extra sense of reality or 3-dimensionality from the 5D. [a href=\"index.php?act=findpost&pid=163559\"][{POST_SNAPBACK}][/a]

You didn't notice the difference between the cameras.    You are a photographer I assume as you are on here, are you not?  Just looking through the viewfinder, it's blindingly obvious [one would think!] how different and how much better the world looks, when using the larger sensored 5D as compared to the very similar 20D. But then I shoot wide open more than I do at f16  like you do, so we take very different images.
It's the same thing as when using 35mm film over digital imaging [ignoring the recording medium quality], but in movie terms it's decribed as being more cinematic as opposed to being more 3D.

[John Sheehy]

Does no-one read manuals or brouchures as it's been possible to shoot at higher shutter speeds with Canons + Nikons for years now? As long as you have a compatible flash.
[a href=\"index.php?act=findpost&pid=164123\"][{POST_SNAPBACK}][/a]

That's not really the same "flash", though.  High-speed sync does not do a lot of what you expect flash to do, like illuminate the entire frame at the same time, for a very short period of time.  With high speed sync, the flash pulses while the slit between the two curtains slowly slides over the sensor.  If you look closely, the pulses are not even in intensity, and moving subjects can become overlapped strips of different moments in time.  The guide number drops to ridiculously low values in high-speed sync mode, too, especially with high shutter speeds.

It is no substitute for the real thing; It's just a weak auxilliary pulsing lamp during the entire ambient exposure period.

[Ray]

Judging by the grey card (combined with Canon's tendency to have 1.2x - 1.25x the stated ISO), however, I would say that it is properly metered for true ISO 100, so one stop ETTR would have been a true ISO 50, and 1.66 stops of ETTR would result in a true ISO 32.
[{POST_SNAPBACK}][/a]

John,
It appears that Canon is now providing greater accuracy of ISO labelling, at least with the 40D.

I'm not sure about the 1Ds3. However, there's a comparison between the 5D and the 1D3 at Outback Photo which shows in ACR the 5D histogram more to the right with the same exposure in shots of the same scene.

[a href=\"http://www.outbackphoto.com/reviews/equipment/canon_1d_MkIII/1d3_5D_S5.html]http://www.outbackphoto.com/reviews/equipm.../1d3_5D_S5.html[/url]

[attachment=4495:attachment]  [attachment=4496:attachment]

[Ray]

You didn't notice the difference between the cameras.    [a href=\"index.php?act=findpost&pid=164126\"][{POST_SNAPBACK}][/a]

Of course I did. But my comments were in reference to the differences in the images from the cameras and specifically in respect of qualities such as 3-dimensionality.

Please try to read what I actually write, jjj, and not what you think I might have written.

[Frank Doorhof]

Lightmeters should be calibrated for your camera/lens/ISO that you use.

When you use it out of the box it's almost 100% sure you will NOT get a correct exposure.

For my Leaf I have calibrated for (from the top of my head)

ISO25 -8
ISO50 -6
ISO100 -4
ISO200 -6
ISO400 -5

Could be a little wrong (don't have it here).

This settings will give me a perfect exposure on all ISO's.
But that's with the 120MM macro, when I use the 150mm I have to counteract app 2/10th.

[Ray]

Lightmeters should be calibrated for your camera/lens/ISO that you use.

When you use it out of the box it's almost 100% sure you will NOT get a correct exposure.

For my Leaf I have calibrated for (from the top of my head)

ISO25 -8
ISO50 -6
ISO100 -4
ISO200 -6
ISO400 -5

Could be a little wrong (don't have it here).

This settings will give me a perfect exposure on all ISO's.
But that's with the 120MM macro, when I use the 150mm I have to counteract app 2/10th.
[a href=\"index.php?act=findpost&pid=164166\"][{POST_SNAPBACK}][/a]

Frank,
That's an interesting complication; the opacity of lenses affecting the true ISO rating. Are you referring here to the calibration of a stand-alone lightmeter or the camera's built-in light meter?

For me, the built-in light meter of Canon DSLRs is sufficient.

[Frank Doorhof]

the build in lightmeter is almost perfect, but that's because it's behind the glass

When you use a lightmeter for studio work or outside you should calibrate that to the camera and used ISO/lens.

It's very easy to do, on my last DVD I did a small video on it.
In short you have to take a REAL 18% graycard with a gray and white patch.
Measure the light on the card to for example f11.
Set the camera on f11 and take the shot.

Go to your RAW convertor and make sure EVERYTHING is zero'd.
Now select the white/gray patch the graypatch should be deadcenter.

If not, calibrate your meter and start at step one.

Do this for ALL ISO's you use and the lens you use the most, to make sure check other lenses.

It sound like alot of work but if you realise that with flash you often use one ISO setting (for me 50 or 25) and probarbly one or two lenses for studio work it's all done in a minute or 5.

[Dustbak]

Frank,

How useful is that when your camera only allows you to set aperture & shutter speed in halve stops and ISO in full stops?

I use my meter only to be able to determine ratio between light sources within a composition. For all other metering I guestimate and use the histogram which is off but at least I know by how much (my main camera doesn't have any metering).

It is not the work that I would not be willing to do but am I really missing out on something?

[John Sheehy]

John,
It appears that Canon is now providing greater accuracy of ISO labelling, at least with the 40D.

I'm not sure about the 1Ds3. However, there's a comparison between the 5D and the 1D3 at Outback Photo which shows in ACR the 5D histogram more to the right with the same exposure in shots of the same scene.[a href=\"index.php?act=findpost&pid=164147\"][{POST_SNAPBACK}][/a]

That doesn't mean anything.  The brightness of a conversion is totally arbitrary, in an absolute sense.

[jjj]

Of course I did. But my comments were in reference to the differences in the images from the cameras and specifically in respect of qualities such as 3-dimensionality.

Please try to read what I actually write, jjj, and not what you think I might have written.
[a href=\"index.php?act=findpost&pid=164148\"][{POST_SNAPBACK}][/a]

I did read what you wrote and seeing as I was specifically talking about the image produced by each camera, maybe you really shouldn't talk about my ability to read. Besides, you slyly used your own illiteracy  to insult me innacurately, simply to sidestep the fact that you cannot tell the difference between the camera's output, whereas it's damn obvious to me and many others. Just because you cannot tell there's a difference, doesn't mean there isn't one. Just like a colour blind person may not distinguish between red and green.

[jjj]

That's not really the same "flash", though.  High-speed sync does not do a lot of what you expect flash to do, like illuminate the entire frame at the same time, for a very short period of time.  With high speed sync, the flash pulses while the slit between the two curtains slowly slides over the sensor.  If you look closely, the pulses are not even in intensity, and moving subjects can become overlapped strips of different moments in time.  The guide number drops to ridiculously low values in high-speed sync mode, too, especially with high shutter speeds.

It is no substitute for the real thing; It's just a weak auxilliary pulsing lamp during the entire ambient exposure period.
[a href=\"index.php?act=findpost&pid=164140\"][{POST_SNAPBACK}][/a]

It may not be exactly the same type of exposure, but you are still able to shoot at much higher syncs than people keep innacurately say you cannot. Which was the point in question.
And if you are shooting fill, you don't always need a massively high guide no., especially if you are using wide apertures, which is when you'd also need the high shutter speed to balance.
I've used this very useful feature and it's worked with no problems.

[Frank Doorhof]

@Dustbak,

VERY VERY important.

With digital it's very important to have a correct exposure, you have a limited range and you have to fill the bits as good as possible.
Or in other words I never fix my exposure in photoshop I try to get it 100% correct in my setup.

Now technical.

You can set your camera in 1/3 of an fstop.
HOWEVER it register much more accurate, that's normal.

In other words if you have to dial in a -0,8 compensation it would mean that if you measure f11 you will actually have to set your camera at app f9.0 that's a big difference.

Also you can set your lights (if you have a good setup) in 1/10th.

Normally you decide what DOF you want and set your lights for that.
So if I want f11 I will set my camera on f11, and now adjust my lights for f11.
if I set my lights on f8 8/10 I will underexpose, if I set the lights on f11 I will get a perfect exposure.

THAT IS, when the meter is calibrated.

There has been more discussions about the need to set your lights in 1/10th of an fstop whilst the camera can only be set on 1/3.
This in my opinion is a lack of knowledge of how things work, they think the wrong way arround (I mean this with no disrespect by the way ).
even if your camera could only be set in full stops it's still important to set your light in 1/10th stops.

Not to get for example f8 9/10th but to get EXACTLY f11.

Let's say that you set your camera on a full stop.
Let's say you set the meter for full stop measurement.

Set the camera on f11.
Measure the light and the meter says f11.
However your exposure is way off.
That's correct.
It can be f8.6 or even f11.4 both rounded off to f11.

So it's important to set your meter to 1/10th and your lights to 1/10th
Only than you can know exactly that your exposure is 100% right, every time.

sorry for the long explanation.

[Dustbak]

I never have to fix my exposure either in PS (at least not a lot) and don't do the elaborate calibration of my exposure meter.

My mostly used camera can only be set in full stops in shutter and halve in aperture stops (full in ISO again).

I do admit I know what most lenses do and where to adjust but again I usually don't have to adjust more than 1/2 stop in PS and am mostly within the 1/4 stop range. I also set my lights (use Elinchrom as well) in 1/10th which can make a difference indeed.

I take a couple of shots to evaluate on the histogram and adjust my lights accordingly with 1/10th increments.

The ratio between the lights I measure with the meter. Yes, I could be wrong in the range of tenths but does that really matter? IMO, not because vary the distance a little bit and your having different ratios. I settle for ball park and while shooting tethered I can always adjust if I don't like what I see.

Maybe to get it exactly there I could get a meter that can be calibrated (I currently use Minolta V and F which cannot be calibrated if I am not mistaken). I will give it a go one day but it is not very high on my list to be frank

[eronald]

I did a test today of the 1Ds3 vs. my Phase P45+/Mamiya. The MF solution won by a decent margin, although the 80mm Mamiya lens could clearly be improved upon.

Both products make good photos. The finder of the 1Ds3 is much improved on the old 1Ds2 finder, mirror return is *snappy*, the color is good, there is a lot of resolution there but a little of the 1Ds2 "plastic wrapping" effect is still there on skin.

The 1Ds3 is a spectacular product, which might still be improved as regards image quality in daylight. The Phase backs need a better camera than the Mamiya to give their best. Maybe Phase could put one of their sensor inside a Canon, and we'd have the best of both worlds ?

Edmund

[HarperPhotos]

I did a test today of the 1Ds3 vs. my Phase P45+/Mamiya. The MF solution won by a decent margin, although the 80mm Mamiya lens could clearly be improved upon.

Both products make good photos. The finder of the 1Ds3 is much improved on the old 1Ds2 finder, mirror return is *snappy*, the color is good, there is a lot of resolution there but a little of the 1Ds2 "plastic wrapping" effect is still there on skin.

The 1Ds3 is a spectacular product, which might still be improved as regards image quality in daylight. The Phase backs need a better camera than the Mamiya to give their best. Maybe Phase could put one of their sensor inside a slightly enlarged Canon ?

Edmund
[a href=\"index.php?act=findpost&pid=164274\"][{POST_SNAPBACK}][/a]

Hi Edmund,

Mamiya are coming out with a new 80mm lens I've been told in the first quarter of 2008. I am also looking forward to the new Mamiya/Phase camera body plus there new leaf shutter lenses.

Regards

Simon

[Ray]

I did read what you wrote and seeing as I was specifically talking about the image produced by each camera, maybe you really shouldn't talk about my ability to read. Besides, you slyly used your own illiteracy  to insult me innacurately, simply to sidestep the fact that you cannot tell the difference between the camera's output, whereas it's damn obvious to me and many others. Just because you cannot tell there's a difference, doesn't mean there isn't one. Just like a colour blind person may not distinguish between red and green.
[a href=\"index.php?act=findpost&pid=164226\"][{POST_SNAPBACK}][/a]

Nothing sly or inaccurate about my insult. It couldn't be clearer that you either didn't read my post carefully, or you read it and mysteriously thought I had written something else.

Here's what I wrote:

The differences I notice are basically higher resolution and more detailed images from the 5D; a necessity to use f16 with the 5D instead of f10 or f11 that I would use with the D60 or 20D for the greatest DoF without noticeable or significant sacrifice of resolution and detail; a shallower DoF with the 5D at the same f stop which gives the 5D a slight advantage when shallowness of DoF is sought, and a wider FoV with the same lenses.

How anyone can interpret that as meaning I can't tell the differences in output between the D60, 20D and 5D beats me.

You should also have gathered from the lead-up to that post that I am referring to a certain quality that seems to be attributed (by DB owners) to the size of the sensor rather than the resolution of the sensor or total number of pixels.

[EricWHiss]

.
.
.

 Maybe Phase could put one of their sensor inside a Canon, and we'd have the best of both worlds ?

Edmund
[a href=\"index.php?act=findpost&pid=164274\"][{POST_SNAPBACK}][/a]

Actually, I think the biggest downfall of the 1DsIII is not the sensor but rather the lenses.  I sure would love to see some tests with the 1DsIII with Leica or contax glass.  When you consider the pixel density and high ISO performance, their sensor is actually pretty darn good.  Using good lenses, the main differentiator will be DR and color.  When you think about it, to achieve twice as much detail you need not 2x pixels but 4x the pixels so even the p45 is only like 1.4 x better than the 1D3.  Color, DR and the magnification factor currently separate the platforms not pixels.  There's lots of argument about whether the magnification factor is real ( I think so) but I don't think anyone will argue about the differences in color or DR.

[Ray]

That doesn't mean anything.  The brightness of a conversion is totally arbitrary, in an absolute sense.
[a href=\"index.php?act=findpost&pid=164192\"][{POST_SNAPBACK}][/a]

I wasn't looking only at the brightness of the image but how far the histogram was from the right with all other controls in ACR at zero. The 5D image requires a +0.33 EC adjustment to bring the histogram just short of touching the right vertical, producing a value in the brightest part of the image (the centre of the white GM patch) of 252, 250, 240.

The 1D3 image, which also looks darker in the ACR default position, requires a +0.67 EC adjustment to bring the histogram to the right, producing an almost identical value in the centre of the white GM patch of 252, 250, 242.

Since we already know from dpreview tests (and probably other tests that you are aware of) that the 5D ISOs are 1/3rd of a stop understated, the above comparison would tend to suggest that the 1D3 is now accurately stating ISO values, as is the 40D.

This result might not be conclusive. However, if ACR is the preferred RAW converter, it would seem that users of the 1D3 should generally be exposing scenes 1/3rd of a stop more (with the same lens at the same aperture and same ISO setting) than they would with a 5D, for a full ETTR.

[attachment=4515:attachment]  [attachment=4516:attachment]

[Ray]

When you think about it, to achieve twice as much detail you need not 2x pixels but 4x the pixels so even the p45 is only like 1.4 x better than the 1D3.  [a href=\"index.php?act=findpost&pid=164337\"][{POST_SNAPBACK}][/a]

That's true. However, it's better if the increase in pixel count does not result in an increase in pixel density. This is the advantage of the larger sensor. The P45 has double the pixel count of the 1Ds3, but similar pixel density. From the same quality lenses the P45 should therefore deliver a sharper result that really is pretty close to 1.4x the resolution (of the 1Ds3).

Imagine what would happen if Canon were to produce a 39mp 35mm sensor. With current Canon lenses the increase in resolution would fall far short of 1.4x, just as I suspect the increase in resolution of the 1Ds3 compared with the original 1Ds is short of the 1.4x figure.

[Mort54]

That's true. However, it's better if the increase in pixel count does not result in an increase in pixel density. This is the advantage of the larger sensor. The P45 has double the pixel count of the 1Ds3, but similar pixel density. From the same quality lenses the P45 should therefore deliver a sharper result that really is pretty close to 1.4x the resolution (of the 1Ds3).

Imagine what would happen if Canon were to produce a 39mp 35mm sensor. With current Canon lenses the increase in resolution would fall far short of 1.4x, just as I suspect the increase in resolution of the 1Ds3 compared with the original 1Ds is short of the 1.4x figure.
[a href=\"index.php?act=findpost&pid=164344\"][{POST_SNAPBACK}][/a]

One other feature of the P45 which may help it achieve better than the 1.4x factor over the 1DsIII is the lack of an AA filter. I very rarely feel the need to sharpen my P45 shots, or if I do, it's only by a very small amount. One of the things I'd dearly like to see Canon and Nikon do is offer an optional AA filter. Some shooters obviously need an AA filter, but others certainly don't.