*** 5Dc vs 60D resolution test ***

Started Mar 17, 2013 | Discussions thread
schmegg
schmegg MOD
Senior MemberPosts: 5,017
Like?
Re: please don't twist my photos
In reply to qianp2k, Mar 22, 2013

qianp2k wrote:

schmegg wrote:

Here are CR2 RAW Files

Added two with 70-200L/4.0 IS. They are 100% cropped and processed by ACR with zero sharpening and zero NR (all other default) but cut most parts of sky and lawn to save my DPR gallery space.

There is a glitch at the moment to directly insert gallery photos.

ACR 5D + 70-200L/4.0 IS 100% cropped (upsampling 5D2 file to the same size of 60D file 5184x3456 before cropping to 100%)

ACR 60D + 70-200L/4.0 IS 100% cropped

To my eyes, 5D sample not only is sharper but also resolves more details (such as on bricks) at least with this lens. 5D photo also has better default contrast and color.

As I said this time I shoot from too far away under not very good light condition. I will do again in much closer distance under better light condition that will benefits 60D more.

Thank you - most helpful and I do appreciate you making them available.

I am very busy these days in work and preparing the Cancun trip starts this weekend. So finally have a chance to check these.

Now - to the subject at hand ...

Unbelievable to see you twisted my photos.

Peter - please!

Those crops are not processed in any way apart from being imported in LR4 with default settings and the enlarged for comparison.

I have not touched any sliders at all apart from the zoom slider.

I have no idea why you did that. But why you even tried to zoom 3x or 4x?

I did that Peter because it show more clearly the difference between the resolving capabilities of these cameras.

It's not 'dark magic', it's simply zooming in on the raw data to see what has been recorded.

It is clear, even with these images (which were shot wide open with the 60D but stopped down one stop with the 5D) that the 60D resolves more detail.

Not true at all provided you don't twist my photos and processed default out of RAW.

They are the default processed files.

They are not. Anyone can follow my default standard steps to duplicate the result but not yours.

It's not necessary to "follow" your "default standard steps". It's actually more valid to not touch the file at all. Simply zoom in on them and see what scene detail they have resolved. (which is what I did)

Some examples where this is easy to see - be sure to view these at full size to see the full effects ...

1. Look at the detail in the hair, look at the eyes, look at the lips, look at the curved 'frame' around him ...

This is thru your twist and I have no idea how you did that. No idea why you view at 3x or 4x sizes as simply not we view at 100% cropped, not on prints.

I view at that size so I can see the detail rendered. It's very simple.

LOL. Nobody post a 300% or 400% cropped or print from that enlargement. What's the point before I even question if you did correctly?

The point is simple. It's to show the difference in resolving abilities of these cameras.

Nothing more, nothing less.

At 100% the difference is visible, but less easy to see - and it's more easy to be fooled into thinking that artifacts due to higher acutance and lower resolution are actual detail - when they are not. It's a bit like the effect noise has on perceived sharpness really - it's an illusion - not real detail.

We can see difference at 100% clearly. Natural sharpness (as I doesn’t even add any software sharpness) is not artifacts. As a matter of fact, 60D already display lots more artifacts. You're right by your twisting it's an illusion, lol.

There is no such thing as "natural sharpness" - it's yet another term you have invented.

Sharpness is a factor of the lens, the sensor, the AA filter, your (or the cameras) ability to focus, your ability to hold the camera steady, the lighting conditions ... and more.

All those images you posted contain artifacts. The higher the resolution, the less significant those artifacts are. And the 60D exhibits less artifacts at the same viewing size when compared to the 5D.

Your images clearly show that.

I'd say at average viewing sizes the difference is moot - certainly not the order of magnitude in favour of the 5D that you endlessly claim on these forums.

What's your definition of average viewing size? My one is at 2000-pixel wide. I can see the difference at this size. It's pretty big difference at 100% size.

Should I link back to the post where you were insisting on 5000-pixel wide images as the point at which a comparison becomes valid?

Honestly, if you wish to determine which camera is capable of resolving more detail, you simply need to zoom in on the raw images and assess it.

That's what I did with the images I posted.

60D on left, 5D on right.

Here is the real one at 100% cropped and I upsampling 5D to match to 60D size at 5184x3456 thru CS6 Bicubic enlargement, default from ACR7, 0 sharpening and 0 NR.

Here are steps that everyone can duplicate:

  • load CR2 files into Photoshop CS6 or CS7
  • Use default setting, move sharpening bars all the way to left (zero sharpening).
  • Move NR bars all the way to left (zero NR);
  • Now in PS window, upsampling 5D files to 60D size (5184x3456) by using Bicubic enlargement (best for enlargement).
  • Generate JPEG by using +12 max quality

Anyone can download my original RAW and duplicate the default processing that I used. 60D photo left and 5D photo right

Yes, and anyone can also simply zoom in using the zoom slider to see the actual amount of detail recorded - rather than relying on these very small and post processed new samples you have provided here.

No idea how you do it. But I only view or print at 100% cropped size.

Why?

100% is arbitrary. Why not 50%? Or 200%? Why not 1% or even 1000%?

If you wish to see the difference in the abilities of these camera to resolve detail, then you zoom in on the image until you start to see a difference. It's not rocket science Peter!

2. In this example it is clear that the 5D was completely unable to resolve the chain-wire mesh in the fence, whereas the 60D has managed to capture it reasonable well (considering how far away this was!) ...

Again this is your twisted result. No ideal how you did that.

60D on left, 5D on right.

Here is the real one at 100% cropped and I upsampling 5D to match to 60D size

Wow, what a big difference and the one below is the authentic one from two cameras.

3. In this example the DO NOT ENTER sign in the 60D image is readable or close to it whereas the 5D has just rendered a mess. Also, look at the edge on the sign between the red and white areas. Also look at the edges of the slanted timber support in the top left-hand corner - the 60D has done OK but the 5D has rendered a staircase ...

Again this is your twisted one.

60D on left, 5D on right.

Here is the real one at 100% cropped and I upsampling 5D to match to 60D size.

Those three examples are just a few of the many areas where it is obvious. The more you look, the more you see.

Please don't twist, then you will see exactly as above and everyone would see that.

Not twisted at all - just enlarged so the difference is easily visible.

LOL, at 100% cropped above and I even upsampling 5D to 6D file size, I already can see the difference, and I am not alone.

Really? No offence, but I think you are definitely in the minority here.

Sorry you don't like it.

Sure if you twisted it.

How is simply loading the files in Lightroom and then zooming on them twisting things?

It's a lot less manipulation than the list of things you gave us as the way to not twist things!

It's a shame the 60D didn't have the benefit of the one-stop closed aperture that the 5D enjoyed for this test, but, in any case, it made no difference.

Actually F4.0 on 60D = F6.4 on 5D.If I shoot F6.4 on 5D, 5D will lead more

Ah - so now you believe in equivalence? At least we have made some progress then!

Since you raised first, and also reflect in real world photos before subject to respective diffraction.

So you are converted? It's hard to tell from that statement.

But, I guess, if you are claiming that the principles of equivalence are required for this comparison, then I guess you are!

I will mark this down (and use it in our future 'discussions')

The 60D has clearly resolved more detail.

LOL, check above again if you don't twist. 5D not only is sharper but resolves more fine details (such as the bricks), cleaner with much less fringe.

No - it's sharper but resolves less detail - as your files clearly demonstrate - if you are prepared to look.

My untwisted processed photos show 5D in outdoor photos, not only are sharper but actually resolve more fine details. Anyone without an agenda would see the same.

So you go through a process to alter the raw data and then claim that your images are "untwisted" - whilst I simply load them into Lightroom and move the zoom slider - and my process is somehow 'twisting' the situation?

O ... K ...

Since I own both cameras and generated thousand and thousand photos from each of them, I know so clearly 5D takes better portrait and landscape photo while 60D has 'reach' (pixel density) advantage in zoo and safari. They complement each other well.

Explain to me how a 60D could have a reach advantage if it resolves less detail?

When you're unable to be closer or your lens is not long enough where 60D has 'reach' (actually pixel density) advantage.

When compared to the 5D, the 60D has this advantage regardless. It doesn't 'magically happen' at some point related to subject distance.

However when you are able to frame the subject into the same AOV, 5D does outresolve 60D with most lenses. DXOMark tests are right.

No, it clearly does not. As your images show. And DxO does not back you up either.

Why do you keep saying this? Do you not understand any of what is being explained to you? Are you really unable to grasp the reality of your own images? Or would you prefer to transfer your faith to DxO when your own images don't agree with what you, mistakenly, think the DxO data is indicating?

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
WhatNew
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow