Should i get the 7D or 5dMKII ??

[Don Hamilton Jr.]

Good morning: What are the down sides to this possible acquistion. In the true application; how bad of a loss of "distance" does one expereince with birds, if you lose the 1.6 sensor, and use a full frame. Just trying to make an intelligent long term acquistion. The 7d truly looks to do everything i need, however the reviews, and pictures from the 5d MKII are simply breathtaking!
Any input is welcome, and advice as to the pro and cons, is greatly appreciated!
PS: i currently own a 50d which is great, however i'm not impresssed when using higher iso settings.
Don

[Doug Brown]

Need a bit more information before attempting to answer your question Don. What are your primary photographic subjects? How much reach do you have with your glass. Moving or stationary subjects. How high are you looking to take your ISO. What don't you like about the 7D images you've seen.

[Don Hamilton Jr.]

Hi doug, thks for your reply.
I honestly can't say i'm see any neagative images from the 7d. I'm a beginning "newbie" want- to- be wildlife photograper, however i like to travel, and see the world, so buildings, and landscape etc are also on the radar, however at time, i'm mostly shooting birds, I own the following:
all canon :
50D
200 2.8 IS 7-200
100-400 IS
500 f4
1.4 EXII
24-105 L IS
580 EXII flash
So i think BIF, and statics are of significant interest, BIF as they are challenging to capture the images.
Thks for your time, and expertise here... just trying to think about it from all angles. As far as iso goes, i was disappointed in the noise on my 50d above 800, however this was when i was at 100%, and i'm very new to the processing workflow with photoshop, and lr 2.5.
Don

[Robert Amoruso]

Don,

My recommendation would be the 7D for birds. That will be the best all-around option for you with birds as the primary focus.

I have a 5D Mark II, do use it for birds when I am too tight with my 50D or 1D Mark III and the 600mm. Yes it produces wonderful, detail rich images too. An example when the 5D was my best choice. (1) Gatorland using the 600mm with the 5D I am right at the best focal length and prefer to use the 600mm due to the low DOF making my backgrounds more OOF, and (2) last week at MINWR with high winds making the birds work to small pools near the entrance - with an all foliage BG, the Mark III was difficult in this situation - due to the slower AF response of the 5D, it worked perfect.

[Don Hamilton Jr.]

Hey robert, thks so much for your observations. Very good stuff for me to consider and think about!
My plans are to head to the venice rookery on saturady am, and then little estero as i head back south to boca. However the weather is looking "less than stellar" for saturday. Regardless, i'm on business friday in sarasota, so these are on my way home , i hope to capture some god images, if the weather improves. What are you guys up to for the weekend???
Thks so much for your input... the 5dMKII poped into the process yesterday as i thought about longterm plans. Regardless, to have a second body is such a relief, as now i can quickly pickup the 7d with the 100-400 and grab some handheld shots as well, vs the gimbal'd 500 f4.
Don

[Doug Brown]

I wouldn't expect 5D Mark II noise levels with the 7D, but it'd be my first choice over the 5D Mark II. It has crop factor, frame rate, AF, and a slew of nice features in its favor. Image quality is excellent overall.

If you can afford one, and you're really concerned about ISOs greater than 800, pick up a 1D Mark IV or a 1D Mark III.

[Aravind Krishnaswamy]

One of the interesting IQ questions is whether a lens bare on a 7D is better than using the same lens with a 1.4x TC on a 5D2 (since the effective reach in those cases is more or less the same). My experience suggests that using the lens bare on a 7D is the way to go. The 5D2 does have better high ISO performance somewhere between 2/3-1 stop better, however that advantage gets nullified when you have to bump the ISO to get the same shutter speed. Furthermore the AF speed also suffers as a result of the 1.4x TC.

So if you are shooting subjects where you are focal length limited, the 7D will be the better choice. If however you can manage to get closer to your subject with the 5D2, it will give you the ultimate in image quality.

[Roger Clark]

Don,
Since you already have a 50D, is a 7D that much better to justify a second camera with small pixels? If your not impressed with the 50D at high ISOs you won't be with the 7D either. The difference in pixels on subject between the 50D and 7D is less than 10% (linear dimension). And as we've seen the smaller pixel pitch is a challenge for some lenses.

Yes, the 5D Mark II has wonderful image quality and superb high iso performance. I find a full frame sensor gives more leeway in pointing and tracking moving subjects. And the full frame is great for landscapes.

It's nice to have cameras with different capabilities for different applications.

(I'm currently using 1DII and 5DII.)

Roger

[arash_hazeghi]

Don I had the 50D and I didn't like it due to noise, lack of DR and soft AA filter, didn't make that many nice photos with it and sold it in less than 4 months :eek:. 7D is a definite improvement in every department however it is not going to be a MKII, it is a solid two stops behind MKII in terms of noise performance and doesn't have the dynamic range.
On the other hand, MKII lacks the AF performance and speed needed for flight shots, the MKII AF module is the same as original 5D and basically similar to that of the 20D plus 6 assist points that don't do much. It shoots at less than 4 fps and its shutter lag + black out time is more than 170 msec. So although I own MKII I don't use it for flight shots, it's a wonderful landscape camera.
So if your primary subject is birds I would recommend a 7D, if you have cash left after upgrading to 7D I would pick a used MKII and use it for landscape shots and when photographing stationary birds.

Good luck

[Roger Clark]

Don I had the 50D and I didn't like it due to noise, lack of DR and soft AA filter, didn't make that many nice photos with it and sold it in less than 4 months :eek:. 7D is a definite improvement in every department

Arash,
If you didn't like the noise and lack of dynamic range, the 7D can not be any better. The 7D has smaller pixels and the Canon 1D Mark II, 5D Mark II, 40D and 50D are plotting on the same trend line with pixel size. E.g., see figure 2 at:
http://www.clarkvision.com/articles/...rmance.summary

The initial sensor data I've seen shows no indication the 7D will break the trend and magically be better. Then there are the strange reports of two green filters causing (new word here): mazing.

I do agree about the frames per second and AF system in the 5DII. Bizarre that canon did not update that. It's like the 5Ds are designed and built in a different division and they don't talk to each other.

Roger

[arash_hazeghi]

Arash,
If you didn't like the noise and lack of dynamic range, the 7D can not be any better. The 7D has smaller pixels and the Canon 1D Mark II, 5D Mark II, 40D and 50D are plotting on the same trend line with pixel size. E.g., see figure 2 at:
http://www.clarkvision.com/articles/...rmance.summary

The initial sensor data I've seen shows no indication the 7D will break the trend and magically be better. Then there are the strange reports of two green filters causing (new word here): mazing.

I do agree about the frames per second and AF system in the 5DII. Bizarre that canon did not update that. It's like the 5Ds are designed and built in a different division and they don't talk to each other.



Roger

Roger, I have done my own tests and they indicate cell quantum efficiency has indeed improved from 7D and 50D, this is also confirmed by Canon data presented at ISSCC 2008 proceedings where this industry meets. We have access to engineering samples of these sensors and measure noise directly with spectrum analyzers in controlled lab environment using monochromatic source as it is performed in this industry, I do not find other methods scientifically acceptable and do not comment on them, most of the reports on the internet are plain BS, you cannot measure cell noise at photon gate level without disabling on chip CDS unit first. I guess we have once talked about this and I told you about our standards and procedures...


Any ways, for any photographer 7D will produce better RAW files than 50D and anyone who actually owns this body and knows what they are doing will confirm this.



Best

[Doug Brown]

I'm going to agree with Arash about the 7D compared to the 50D. RAW image files look much nicer at a given ISO coming from the 7D; noise has a much finer grain.

[Robert Amoruso]

Roger,

The problem with the 5D Mark II is its sluggish AF response. Given certain circumstances like the ones I mentioned about, it is great on birds and I do use it for birds but it can't beat the 50D, 7D or 1 series in the AF department.

As an 50D owner I am please with the camera but just know not to push the ISO limit (I try to stay at 400). But that is a big limitation from using my 1D Mark III where ISO800 is very useful and has become my standard until it gets bright out then I lower ISO.

From what I can see, the 7D is out-performing the 50D noise-wise. That conclusion is based on subjective comparison of images and not science. Though I am a tech-geek and appreciate the tech end of things, for my images in processing, it is the subjective results that mean the most to me.

[Flavio Rose]

We have access to engineering samples of these sensors and measure noise directly with spectrum analyzers in controlled lab environment using monochromatic source as it is performed in this industry, I do not find other methods scientifically acceptable and do not comment on them, most of the reports on the internet are plain BS, you cannot measure cell noise at photon gate level without disabling on chip CDS unit first.

The noise model that Roger and Emil Martinec teach on their websites seems to me logical and consistent. It involves simplifying assumptions but so do most models in science, e.g., molecular orbitals, Bloch wavefunctions in crystals :). A model is judged by the validity of its predictions. Are you saying that a model of noise based only on a "full well" value and a read noise value does not predict well the noise that is actually seen in raw images? Or only that "full well" is a misnomer because of CDS?

A major takeaway from Roger and Emil's sites is that there is only slow progress in the "full well" per unit area in their model. Is this incorrect?

Any enlightenment would be much appreciated.

[arash_hazeghi]

The noise model that Roger and Emil Martinec teach on their websites seems to me logical and consistent. It involves simplifying assumptions but so do most models in science, e.g., molecular orbitals, Bloch wavefunctions in crystals :). A model is judged by the validity of its predictions. Are you saying that a model of noise based only on a "full well" value and a read noise value does not predict well the noise that is actually seen in raw images? Or only that "full well" is a misnomer because of CDS?

A major takeaway from Roger and Emil's sites is that there is only slow progress in the "full well" per unit area in their model. Is this incorrect?

Any enlightenment would be much appreciated.

Flavio,

You cannot measure and deduce components of noise (1/f noise, reset noise, dark noise, cross talk etc.) accurately or estimate cell QE by looking at the RAW recordings of an APS sensor, There is just too much happening from the point that the bit line is charged to the point where data is flushed out of the DSP unit. The issue here is not the validity of the noise model but rather the measurement procedures although noise models need be accurate enough as well.

If you are interested to see how we measure noise in the lab and how data is presented in technical articles have a look at this http://isl.stanford.edu/~abbas/group/papers_and_pub/hui_thesis.pdf it is an entire PhD thesis, although some minor details have changed since 2002 since todays chips perform much more functions at circuit level, but fundamentals are the same. Here is another useful reference
http://www-isl.stanford.edu/~abbas/group/papers_and_pub/jssc01_hui.pdf
Do you think that half a dozen tenured faculty members and more than 30 full time PhD candidates and researchers waster a bunch of million dollars a year trying to design, model and measure something that a person can do at home? :D
Here is a photo of one of a dozen setups in the lab we use to measure noise, this is about $ 800,000 worth of equipment just for measuring dark current from 300K down to 4K.
Do you think we do this for fun??


Stanford CIS test lab, photo for illustration only copyright Arash Hazeghi.

With respect to the other question I cannot comment on methods which have not been published in a technical journal or have not been presented at a technical conference, but the so called "noise" or standard deviation you see in a RAW recoding can be very different from the noise at the out put of the photo cell. As far as I am concerned, noise is a number expressed in electrons or micro volts per square root Hertz at the out put node for single wavelength excitation and unless I see a plot of such data I don't take it seriously.

I can give you one example and that is the patterns that you see when you lift the dark areas of the photos from a digital camera is an artifact of CDS because of global and local variation between reference cell banks that are distributed across the chip. Comments like misalignment of CFA and such are absolutely ridiculous...Do these people actually know how a sensor is fabricated and packaged in the fab?

BTW, regarding Bloch model it is a good starting point for bulk semiconductors but with today's nano structures things change dramatically and strong quantum confinement and coherence effects are observed which are being researched and models are being developed, this is actually the title of my PhD thesis!

Anyways, I am not here to discuss these issues but to learn and improve photography skills from the great photographers of this website. As far as any photographer is concerned 7D has lower noise in the final RAW images than the 50D.

best,

[Flavio Rose]

Do you think that half a dozen tenured faculty members and more than 30 full time PhD candidates and researchers waster a bunch of million dollars a year trying to design, model and measure something that a person can do at home.

Thank you very much for your explanation and the references. I was long ago a graduate student in EECS at MIT. For me, crude device and process models with some predictive power are not inconsistent with intensive research into the devices and processes being modeled.

[Roger Clark]

You cannot measure and deduce components of noise ......

Arash,
You have leveled a number of charges here. But the first thing you need to realize is most of us who study sensors do so for the sake of getting better images, not to understand the details of what is every little noise source. Second, you need to realize that scientists have been doing this kind of work for decades and not in your lab with $800,000 in equipment. One can derive quite a lot about sensors with some very simple methods. And in fact the methods in use are those published in the scientific literature. One does not need calibrated light sources in a fancy laboratory to determine noise or quantum efficiency. In fact astronomers have been doing such work for decades. For example there is a large network of standard stars with spectral intensity known to high precision. It is fine to study a sensor and its details and I applaud and encourage labs such as yours to do so, because that is were the advances come from. But scientists need more than sensor performance, as do digital photographers: they need total system performance in the field where actual data are acquired about our environment.

You have presented zero evidence that any of the basic data being derived by different people on DSLR sensors is in error by any significant amount. In fact the papers you note: (e.g. Hui et al 2001) "In a CCD image sensor, temporal noise is primarily due to the photodetector shot noise and the output amplifier thermal and 1/f noise. CMOS image sensors suffer from higher noise than CCDs due to the additional pixel and column amplifier transistor thermal and 1/f noise." The key factor here is SHOT NOISE. Shot noise (noise due to the random arrival of photons) is the dominant noise source in almost all DSLR camera images except the very darkest parts os most images. And people are deducing thermal noise as well.

It is a fact that shot noise dominates the performance of every digital camera. It is BASIC physics that dictates the noise performance of a pixel. For example, a lens delivers X photons per square micron to the focal plane. Chop the focal plane into smaller pixels and each pixel gets fewer photons. Changing QE, fill factor, or filter transmittance only give small changes to the overall trend of less photons per pixel as the pixel size shrinks.

So you can cite your group and its accomplishments all you want. It does not change the fact that there are many very bright people out there using real sensors in real applications and are able to derive basic performance from their measurements. This is done every day at telescopes, on spacecraft, and yes, by the the photographer at home. It doesn't take a rocket scientist to do it.

Oh, and by the way, some are "rocket scientists."

Roger

[Roger Clark]

Roger,

The problem with the 5D Mark II is its sluggish AF response. Given certain circumstances like the ones I mentioned about, it is great on birds and I do use it for birds but it can't beat the 50D, 7D or 1 series in the AF department.

Robert,
I had said I agree that the 5DII AF performance is not the fastest and the frame rate low. I do have a 5DII and I do use it for birds and wildlife action, along with my 1DII. I have used the 5DII on birds in flight with success, and when it works, the images are wonderful.

Roger

[arash_hazeghi]

Arash,

"In a CCD image sensor, temporal noise is primarily due to the photodetector shot noise and the output amplifier thermal and 1/f noise. CMOS image sensors suffer from higher noise than CCDs due to the additional pixel and column amplifier transistor thermal and 1/f noise." The key factor here is SHOT NOISE.

It is a fact that shot noise dominates the performance of every digital camera. It is BASIC physics that dictates the noise performance of a pixel. For example, a lens delivers X photons per square micron to the focal plane. Chop the focal plane into smaller pixels and each pixel gets fewer photons. Changing QE, fill factor, or filter transmittance only give small changes to the overall trend of less photons per pixel as the pixel size shrinks.

So you can cite your group and its accomplishments all you want. It does not change the fact that there are many very bright people out there using real sensors in real applications and are able to derive basic performance from their measurements. This is done every day at telescopes, on spacecraft, and yes, by the the photographer at home. It doesn't take a rocket scientist to do it.

Oh, and by the way, some are "rocket scientists."

Roger

Roger,

"In a CCD image sensor, temporal noise is primarily due to the photodetector shot noise and the output amplifier thermal and 1/f noise. CMOS image sensors suffer from higher noise than CCDs due to the additional pixel and column amplifier transistor thermal and 1/f noise."

Looks like you did not read the sentence carefully, let me interpret for you, it means that CMOS sensor has extra 1/f noise and thermal noise from the active pixel elements in addition to the shot noise, where in the paper is stated that a CMOS sensor is only photon shot noise limited? Also not all shot noise is from photons, there is dark shot noise and electronic shot noise associated with every single transistors. When the term shot noise is used it refers to all shot noise components not just photons.

None of the current commercial sensors on the market is only photon shot noise limited, not at all. The fact that you think sensors are dominantly photon shot noise limited means your analysis is not capturing some vital details, ignoring the fact that CMOS sensors perform on chip CDS and many other tricks and thus overlooking all other sources of noise, all but photon shot noise and deriving trends and conclusions.

My response to you is if you have trust in your methods please publish them in a reputable image sensor related journal or present them at a conference and then if it gets accepted you can run your parallel lab and drive us out of business, I am sure Canon will come to you instead of us and pay you to do research for them. The fact that you have great achievements in the field of astronomy does not mean that you are an expert in image sensor area which is a different field, no one can concentrate on many fields and know all the details.

Also I still remember that we had a heated debate about a 20bit DR sensor which you not only denied its existence but also tried to prove to me with some math that it was impossible to make, of course this was until I showed you a few papers presenting the actual fabricated sensors with 21Bits of DR...so every human can make a mistake.

BTW, now that you love photon shot noise it is good to remember that photons or electron shot noise is a result of commutation relationship between electric and magmatic field operators in occupation number representation in quantum field theory, there is no need to use the analogous random arrival terminology.


Any ways, I am off and reluctant to continue this discussion, you can continue to believe what you like and do your analysis. I have plenty of technical discussions with the relevant crowd and I am here to enjoy photography and learn to improve my skills : )

I apologize to the OP for diversion.

Best and good luck

[arash_hazeghi]

Thank you very much for your explanation and the references. I was long ago a graduate student in EECS at MIT. For me, crude device and process models with some predictive power are not inconsistent with intensive research into the devices and processes being modeled.

Flavio crude models are good for capturing the trends but not the details... MIT is a great school I have many friends there, we have weekly seminars with prof. Antoniadis and Hoyt at EECS and their work is over the top :)

Do you work on image sensor these days?

You know that I am not a big fan of Canon and have given my honest opinion about the 7D AF system but with respect to sensor Canon engineers have worked hard and designed a fine sensor for the 7D and deserve credit, it is absolutely unfair to say 7D has wore noise performance than 50D. You can buy one for yourself and appreciate it (AF aside)

[Roger Clark]

Roger,

"In a CCD image sensor, temporal noise is primarily due to the photodetector shot noise and the output amplifier thermal and 1/f noise. CMOS image sensors suffer from higher noise than CCDs due to the additional pixel and column amplifier transistor thermal and 1/f noise."

Looks like you did not get the sentence right, let me interpret for you, it means that CMOS sensor has extera 1/f noise and thermal noise from the active pixel elements in addition to shot noise, where in the paper is stated that a CMOS sensor is only shot noise limited?

None of the current commercial sensors on the market is photon shot noise limited, the fact that you think sensors are dominantly photon shot noise limited means your analysis is incorrect, and you fail to recognize CMOS sensors perform on chip CDS and many other tricks. I have no further comment nor any interest in discussing this any further, if you have trust in your methods please publish them in a reputable journal or present them at a conference and then if it gets accepted we can talk more.

Also I still remember that we had a heated debate about a 20bit DR sensor which you not only denied its existence but also tried to prove to me with some strange math that it was impossible to make, of course this was until I showed you a few papers presenting the actual fabricated sensors with 21Bits of DR... Based on the history I feel like discussion is pointless.

I am off.


Best and good luck

Arash,
Please get your facts straight. Here is a direct quote from me from that thread:
http://www.birdphotographers.net/for...=dynamic+range
rnclark said:
"I computed 20 bit SNR. If the read noise were only one electron, then the one would need to collect only 2^20 = a hair over one million photons, not a million squared. At an electron density of 1000 electrons per square micron, one only needs on the order of a 33 micron square pixel, and at a few electrons read noise correspondingly larger area. So I agree it is feasible to do a linear 20 bit DR sensor."

Then you might read some of the papers from your lab, or a simple google search of many papers that discuss photon shot noise, e.g.:
http://citeseerx.ist.psu.edu/viewdoc...=rep1&type=pdf
A Method for Estimating Quantum Eciency for CMOS Image
Sensors by Boyd Fowler, Abbas El Gamal, David Yang, and Hui Tian
The QE estimate relies on photon shot noise. At high signal levels photon noise dominates the noise sources. This is very obvious in any plot of variance versus photon cunt and the method outlined in the paper is very similar to the methods used by multiple people to determine gains, saturation levels and noise floor in DSLRs.

If you have evidence that a CMOS sensor with 10,000, 20,000, 50,000+ photons incident on a pixel has noise components greater than the photon shot noise, please give me a reference.

Roger

[arash_hazeghi]

Arash,
Please get your facts straight. Here is a direct quote from me from that thread:
http://www.birdphotographers.net/for...=dynamic+range
rnclark said:
"I computed 20 bit SNR. If the read noise were only one electron, then the one would need to collect only 2^20 = a hair over one million photons, not a million squared. At an electron density of 1000 electrons per square micron, one only needs on the order of a 33 micron square pixel, and at a few electrons read noise correspondingly larger area. So I agree it is feasible to do a linear 20 bit DR sensor."

Then you might read some of the papers from your lab, or a simple google search of many papers that discuss photon shot noise, e.g.:
http://citeseerx.ist.psu.edu/viewdoc...=rep1&type=pdf
A Method for Estimating Quantum Eciency for CMOS Image
Sensors by Boyd Fowler, Abbas El Gamal, David Yang, and Hui Tian
The QE estimate relies on photon shot noise. At high signal levels photon noise dominates the noise sources. This is very obvious in any plot of variance versus photon cunt and the method outlined in the paper is very similar to the methods used by multiple people to determine gains, saturation levels and noise floor in DSLRs.

If you have evidence that a CMOS sensor with 10,000, 20,000, 50,000+ photons incident on a pixel has noise components greater than the photon shot noise, please give me a reference.

Roger

Facts are straight Roger, you were flatly denying the 20Bit sensor until I posted the paper, why did you censor the prior correspondence? After that you corrected your calculation which still has the pixel size off.

The paper does not state that estimation of QE relies on photon shot noise only, rather it states I quote

"We described a new technique for estimating QE, gain, gain FPN, dark charge, and readout noise for CMOS APS.
Our method is similar to Janesick's method in that it relies on the Poisson statistics of shot noise, but it also takesinto consideration the nonlinearity of the APS response, gain variation, ofsets, and readout noise. We provided a least squares method for estimating the gain and QE and demonstrated that the estimators have low error"


Again I need to interpret for you, the above means that Poisson stat was used for modeling shot noise, and as you can read all other sources are included, where does it say that photon shot noise is the only source?



And oh Yeah the methods are similar to yours, especially using RAW files!!!




Did you see Fig 1 on page 2? Doesn't this look like the equipment setup I posted earlier???????????????????????????


Apparently you know better than us what we are doing

Any ways, I am done Roger, seriously. Friday night and time to go out :)

Have a good weekend ;)

[Peregrine Craig Nash]

This has to be one of the most boring discussions on essentially a bird photography site!

[Robert Amoruso]

Robert,
I had said I agree that the 5DII AF performance is not the fastest and the frame rate low. I do have a 5DII and I do use it for birds and wildlife action, along with my 1DII. I have used the 5DII on birds in flight with success, and when it works, the images are wonderful.

Roger

Roger. Looks like we agree. :) Sorry if it looked like is was saying that you did not know this - that was not the case.

For the originally poster, I just wanted to make the point that IMO, if bird photography is your prime mission, the the 5D Mark II is not a good first choice camera for that task. As we have both noted, it can be a good choice in certain circumstances but I would not want it to be the only camera or the first camera I had for bird photography.

[Peregrine Craig Nash]

Robert one of the best bird photographers in Ireland is a guy called Anthony McGeehan and he uses a 5D Mk1 with a 400 f5.6 lens only. Infact one of the best bird photographers I have come across. Ultimately its the guy behind the camera and the ability of their fieldcraft.

[Robert Amoruso]

Robert one of the best bird photographers in Ireland is a guy called Anthony McGeehan and he uses a 5D Mk1 with a 400 f5.6 lens only. Infact one of the best bird photographers I have come across. Ultimately its the guy behind the camera and the ability of their fieldcraft.

I certainly do not discount that the person behind the camera is the deciding factor. If you do not have the skill, then the equipment does not matter. But I do believe that if you do have that skill, then better AF accusation and ability to hold focus is an asset - as well as the significant better frame rate. Thanks for the feedback - I will definitely look up Mr. McGeehan's work.

[Roger Clark]

Facts are straight Roger, you were flatly denying the 20Bit sensor until I posted the paper, why did you censor the prior correspondence? After that you corrected your calculation which still has the pixel size off.

Arash,
It is very simple. I admitted a mistake and corrected it. But I can't find anywhere where I said it was impossible. In fact in the exchange, you made the statement:
"Sorry, I thought you doubted the existence of these sensors from your previous post, my bad."
I simply said for a 20-bit dynamic range you must collect at least 2^20 photons, and I still stand by that,

I asked you a straight forward question about noise sources which you have yet to show any data for or any reference. You only cite multiple noise sources are taken into account. Yes, many people do that; it's not unique.

Again the question is how much noise from each source contributes to the total noise when there is 10,000, 20,000, 50,000 photons collected by a pixel? For example, what is the percentage of each in terms of the total signal and total noise? E.g. assume 1/100 second exposure.

-------------------photons collected
--------------------- 10,000 --- 20,000 --- 50,000
photon noise
read noise
thermal noise
1/f noise
--------------------------------------------------------------
Total noise: -------- 100% --- 100% --- 100%
total noise
referenced to
electrons:
--------------------------------------------------------------
Signal/Noise

Roger

[arash_hazeghi]

Robert one of the best bird photographers in Ireland is a guy called Anthony McGeehan and he uses a 5D Mk1 with a 400 f5.6 lens only. Infact one of the best bird photographers I have come across. Ultimately its the guy behind the camera and the ability of their fieldcraft.

No amount of equipment can make up for lack of skill but you need to have the right tools for the job too, can't make good steak in a soup pot.

I have not seen this gentlman's portfolio but Rob's and Doug's (and other BPN members') photos are over the top as well.

[Roman Kurywczak]

Hi Don,
I'm agreeing with Doug and will give another opinion.....I waited for the Mark lll before my first DSLR purchase.....why?....i wanted the speed of my film camera's.....once the Mark lll came along with the high noise capabilities.....I was hooked! The proof is in the images! Save the pennies.....get the Mark lV or a leftover Mark lll......you can then do all the diciplines of photography equally.....Oh...did I mention the super battery life?!!!

Arash....I will say that I am disappointed as I respect your photography.....you can't make this statement; "...You know that I am not a big fan of Canon and have given my honest opinion..." and be unbiased! The statement ..."not a big fan of Canon" definitely casts a bias on your opinion.....you should remove yourself from this debate or this is about as useful as a Canon/Nikon debate.
Respectfully,
Roman

[Mike Tracy]

Yawn to all the banter and I even understand it (most).

For birds or sports get the 7D, the 5D will excel IMO in all other areas.

[arash_hazeghi]

Hi Don,
I'm agreeing with Doug and will give another opinion.....I waited for the Mark lll before my first DSLR purchase.....why?....i wanted the speed of my film camera's.....once the Mark lll came along with the high noise capabilities.....I was hooked! The proof is in the images! Save the pennies.....get the Mark lV or a leftover Mark lll......you can then do all the diciplines of photography equally.....Oh...did I mention the super battery life?!!!

Arash....I will say that I am disappointed as I respect your photography.....you can't make this statement; "...You know that I am not a big fan of Canon and have given my honest opinion..." and be unbiased! The statement ..."not a big fan of Canon" definitely casts a bias on your opinion.....you should remove yourself from this debate or this is about as useful as a Canon/Nikon debate.
Respectfully,
Roman

Hi Roman,
This was not a debate between Canon and Nikon, for what matters in this thread all I said was good things about 7D and that it had improved relative to 50D, I have both Nikon and Canon cameras and lenses and I use them as tools to take pictures, I am not "in love" with one brand of camera, I think there is difference between using a tool for a job and being a "fan" I do like my Canon gear and that is why I keep shooting with it, but there are some weak points too. No camera is prefect.

When there is a discussion I point out the strengths AND weaknesses :D I don't think pointing out weakneses makes me "biased"

Thanks for mentioning your thoughts.


Best

[Don Hamilton Jr.]

Thks everyone for your thoughts, and wisdom here! I did chose the 7D at this time. However i'm going to sale the 50D down the road, first things first as several guys mentioned, one must be proficient, and develope a sound technique first. Building my skills, and shooting time has no substitute!
I do have my eye on a full frame for the future. It's certainly a good idea, to be able to cover all the disciplines.
Don

[Jay Gould]

Well, I for one has always enjoyed and learned from these technical debates. It takes several rereadings - which I do - to get a basic understanding and these debates certainly further all of our knowledge about the workings of our tools. I am not a technical person - admit to being a wordsmith - and therefore the correctness of either debater's conclusions is not for me to decide. In fact, the ability to decide who is right is not necessary to enjoy and appreciate the debate per se.

The Journey, Not the Arrival, Matters!

[Jay Gould]

Regarding the OP, I have both the 7D and the 5D2. Hands down, the 7D is the camera of choice for fast moving subjects as is the 5D2 for landscapes and macros. I am on the list for a first out of the box - hopefully in the next two weeks - for a 1D4. I was going to sell the 7D; however, seeing the recent photo of Artie with the 70-200 hanging down his back while using a long lens on a tripod has caused me to rethink selling the 7D. What's another body amongst friends?!

[Bryan Hix]

Wow, amazing post responses. Note to self: don't ask anyone what camera model to buy..............That was the original question, right? :eek: