cjtaylo wrote:
Yes, the technical white paper says 10m indoors and 15 meters outdoors. I'm looking forward to getting rid of my cheap ebay triggers :-)
alundeb wrote:
Sounds like effects of NR combined with sharpening. My 500D can do that sort of things by the default settings in DPP. Relaxing NR and sharpening helps.
This is just a general comment, and should not be taken as defence of the 7D, or denial that Reichmann may have seen something.
if you look at the RAW samples from IR you can see lots of sensor artifacting from imbalanced channels, hopefully only a few pre-pro units have that issue....
it is a little bad that MR seems to hint that maybe he has that too....
then again maybe for the test units they didnt have time to calibrate the sensors....
skibum5 wrote:
if you look at the RAW samples from IR you can see lots of sensor artifacting from imbalanced channels, hopefully only a few pre-pro units have that issue....
it is a little bad that MR seems to hint that maybe he has that too....
then again maybe for the test units they didnt have time to calibrate the sensors....
Are you saying the 7D sensor is not that great after all? Hmmm...
I have yet to come across a single report that says the production copy behaves differently from a pre-production copy.
thw2 wrote:
Are you saying the 7D sensor is not that great after all? Hmmm...
I have yet to come across a single report that says the production copy behaves differently from a pre-production copy.
I guess skibum is talking about the imbalance between the two green channels. That is certainly something that should be fixed by adjustment or firmware or correct parameters for the RAW conversion software.
alundeb wrote:
I guess skibum is talking about the imbalance between the two green channels. That is certainly something that should be fixed by adjustment or firmware or correct parameters for the RAW conversion software.
alundeb wrote:
I don't know why RG uses the words he does.
"A lot like they've come from a 50D" can mean anything from worse to better, and no one knows which direction, and by how much.
Why not check comparisons rather than uncomparable test shots in arbitrary conditions?
One example, the presentation of samples from IR, by thedigitalbean on page 41 in this thread.
headroom wrote:
The Picture in the Link above
What for is this good? Its Pixel Leveld based and has nothing to do the real Image sizes....
Wrong, the sizes are equalized, according to thedigitalbean who posted them:
"I have put together a camera/ISO matrix based off the Imaging Resource RAW files and ACR 5.5/LR 2.5. Click to link view actual pixels. All other cameras were resized (up or down) to the 7D's size to keep the comparison fair."
skibum5 wrote:
yeah they should've at least tossed us the f/4 high-precision for the center since it does make all the f/4 dodgier for AF.
Well duh, they tossed us high precision for the center top, center middle and center bottom sensors working up to f/5.6 by incorporating interleaved horizontal line sensors at those positions into the cross type sensor.
Citing from CPN:
"f/5.6 sensors are especially good at predictive tracking focus, however, the designers wanted to find a solution that also offers high focus accuracy as well. Increasing focus accuracy can be achieved by using a higher resolution focus sensor however, simply increasing the pitch, and therefore resolution, of a single line too much leads to a lower signal to noise ratio in terms of light levels recorded. To combat this, the centre top, centre middle and centre bottom sensors (when the camera is held in a landscape orientation) have a dual zigzag horizontal line sensor.
The dual zigzag arrangement solves this problem by using two focus sensors that are slightly offset from each other. By performing AF calculations based on the two lines sampling slightly different areas, the focusing resolution is improved leading to a more precise focus result while still retaining the responsive AF tracking."
Except for the single horizontal line sensor which works up to f/8 the 7D autofocus module is wiping the floor with the 1D series module...
thw2 wrote:
How exactly does such an artifact look like?
I saw an example once, when viewed at 400%, one could see that each 4'th pixel was a tiny bit darker, or something like that. Don't remember where I found it, but skibum may help us here.
Depending on the demosaic algorithm, it may also show up as a maze pattern.
keithreeder wrote:
I think I can live with that level of "damage" to the IQ...
Yes, if that's all
I'm buying it. Preordered at the release date, not knowing for certain, with a free cancel option. After consuming all the information available so far, I am not afraid that any hidden problems are going to break the deal for me.
Most important to me, are resolution, pixel density, noise and dynamic range. I see a distinct improvement over my 500D, but everyone will have to see for themselves, as said here before :-)
alundeb wrote:
I'm buying it. Preordered at the release date, not knowing for certain, with a free cancel option. After consuming all the information available so far, I am not afraid that any hidden problems are going to break the deal for me.
Most important to me, are resolution, pixel density, noise and dynamic range. I see a distinct improvement over my 500D, but everyone will have to see for themselves, as said here before :-)
Ditto. Having checked the raw-files myself, I'm not concerned about the image comment at Luninous Landscape since it most probably stems from poorly tuned raw-converter software at time point (Adobe themselves says support is beta and NOT optimized). I'm also having a feeling that the built-in noise reduction on jpegs isn't that great, but I shoot raw and can always turn it off if jpegs are needed for action burst-shoots.
charlyw wrote:
Well duh, they tossed us high precision for the center top, center middle and center bottom sensors working up to f/5.6 by incorporating interleaved horizontal line sensors at those positions into the cross type sensor.
Citing from CPN:
"f/5.6 sensors are especially good at predictive tracking focus, however, the designers wanted to find a solution that also offers high focus accuracy as well. Increasing focus accuracy can be achieved by using a higher resolution focus sensor however, simply increasing the pitch, and therefore resolution, of a single line too much leads to a lower signal to noise ratio in terms of light levels recorded. To combat this, the centre top, centre middle and centre bottom sensors (when the camera is held in a landscape orientation) have a dual zigzag horizontal line sensor.
The dual zigzag arrangement solves this problem by using two focus sensors that are slightly offset from each other. By performing AF calculations based on the two lines sampling slightly different areas, the focusing resolution is improved leading to a more precise focus result while still retaining the responsive AF tracking."
Except for the single horizontal line sensor which works up to f/8 the 7D autofocus module is wiping the floor with the 1D series module......Show more →
The only focus point which has the extra sensitivity at f2.8 is the centre one as per a direct quote from that canon webpage. I didn't see it said anywhere about it working at f8 (think telephoto with 2X converter), have you got a link to that, I did look for it wondering whether is was something else that canon was keeping for the 1 series (understandably actually).
Beni wrote:
The only focus point which has the extra sensitivity at f2.8 is the centre one as per a direct quote from that canon webpage. I didn't see it said anywhere about it working at f8 (think telephoto with 2X converter), have you got a link to that, I did look for it wondering whether is was something else that canon was keeping for the 1 series (understandably actually).
I wrote that f/8 focusing is the only omission. And I quoted directly from the Canon Professional Network site about the high precision focusing points located at the top, middle and bottom of the center row - which are high precision focusing line sensors all the way to f/5.6.
charlyw wrote:
Except for the single horizontal line sensor which works up to f/8 the 7D autofocus module is wiping the floor with the 1D series module...
I doubt that will likely be the case with the next 1D iteration; Canon will likely continue to use AF as part of their market segmentation strategy.
IF the 7D AF "approaches" the last working 1D focusing performance I'll be a very happy camper. Time will tell. We should know in the next few weeks, as people start using the 7D (it's already available on some streets in the UK).
charlyw wrote:
I wrote that f/8 focusing is the only omission. And I quoted directly from the Canon Professional Network site about the high precision focusing points located at the top, middle and bottom of the center row - which are high precision focusing line sensors all the way to f/5.6.
Read it again, it's only the center focusing point which has extra sensitivity at f2.8.
"For enhanced precision, the central AF point features extra sensitivity when used with lenses having a maximum aperture of f/2.8 or faster thanks to the diagonal arrangement of the cross-type sensor"
Beni wrote:
Read it again, it's only the center focusing point which has extra sensitivity at f2.8.
You should reread the quote I delivered in my original response. The f/2.8 X aligned cross type sensor is an additional with extra precision for those lenses. Then there are the normally precise 19 cross type AF sensors of which there are three that have horizontal (thus sensitive against vertical contrast) line type detectors which are using an interleaved double AF sensor cell array and thus are delivering higher precision information than the rest.
For enhanced precision, the central AF point features extra sensitivity when used with lenses having a maximum aperture of f/2.8 or faster thanks to the diagonal arrangement of the cross-type sensor - a feature first seen on the EOS 50D. This diagonal placement with the sensors forming an X-arrangement means that the f/2.8 sensors do not overlap with the f/5.6 sensors at the same point as they have a horizontal/vertical arrangement. This avoids any interference between the two sensors and offers both fast tracking performance from the f/5.6 sensor and extra precision from the f/2.8 sensor.
f/5.6 sensors are especially good at predictive tracking focus, however, the designers wanted to find a solution that also offers high focus accuracy as well. Increasing focus accuracy can be achieved by using a higher resolution focus sensor however, simply increasing the pitch, and therefore resolution, of a single line too much leads to a lower signal to noise ratio in terms of light levels recorded. To combat this, the centre top, centre middle and centre bottom sensors (when the camera is held in a landscape orientation) have a dual zigzag horizontal line sensor.
The dual zigzag arrangement solves this problem by using two focus sensors that are slightly offset from each other. By performing AF calculations based on the two lines sampling slightly different areas, the focusing resolution is improved leading to a more precise focus result while still retaining the responsive AF tracking.
Iguess you are both correct, sort of.
The centre point is f2.8 (enhanced precision), but there are three f5.6 points with 'dual zigzag' (more precise focus).
BUT are those numbers reversed? I think the ST-E2 has more range indoors with near 50 ft and 33 ft outdoors. 