PDA

View Full Version : New AA filter in 1DIIN - yes or no?


dhphoto
November 22nd, 2006, 05:31 AM
I'd really appreciate an input from someone who knows, definitively, whether there is indeed a 'milder' AA filter in the 'N' as the images I am getting are far more pleasing straight out of camera (RAW of course) and need much less post-processing.

It seems there are two camps - 1DII owners (unsurprisingly) say no and 1DIIN owners (who previously owned the original 1DII) say definitely yes.

Can someone who knows please speak up?

Thanks

David

Bart_van_der_Wolf
November 22nd, 2006, 06:28 AM
I'd really appreciate an input from someone who knows, definitively, whether there is indeed a 'milder' AA filter in the 'N' as the images I am getting are far more pleasing straight out of camera (RAW of course) and need much less post-processing.

I assume you mean post-processing in terms of sharpening?

Can someone who knows please speak up?

I don't know, but I will speak up. If nobody reports in with an absolute answer, I may be of help in finding out. I can suggest two methods of determining the differences, if any.

Method 1:
I have created a target file that is both suited for testing Digicam sensors (with or without AA-filter) and analog film, and you can make your personal resolution test target from it at home with a decent inkjet printer.
For HP/Canon inkjet printers (http://www.xs4all.nl/~bvdwolf/main/downloads/Jtf60cy-100mm_600ppi.gif) (3.8MB).
For Epson inkjet printers (http://www.xs4all.nl/~bvdwolf/main/downloads/Jtf60cy-100mm_720ppi.gif) (5.3MB).

Print it at the indicated ppi without printer enhancements on glossy Photopaper which should produce a 100x100mm target, and shoot it with your (digi)cam from a (non-critical) distance like between 25-50x the focal length. Refocus between shots. For comparison between bodies, one would ideally use the same lens at the same aperture setting.

The resulting "blur"center diameter is a measure of "on-sensor resolution" of the whole optical chain (lens+AA-filter+sensor), and can be expressed as cy/mm after calculating "(60/pi)/diameter". The diameter can be expressed as number of pixels multiplied by the pixel pitch.

This method allows a visual impression, as well as a quantified limiting resolution determination expressed in pixels or cy/mm.

Method 2:
Print a 'Slanted Edge Target', and use a program like Imatest (http://www.imatest.com/docs/sharpness.html#newtest) to determine an MTF type of response graph.

The latter method is more accurate, but requires the use of Imatest which I can assist with. The preparation of a slanted edge target can be as simple as using one of the downloadable versions (http://www.imatest.com/docs/sfr_instructions.html#download).


Bart

Chuck Westfall(Canon USA)
November 22nd, 2006, 06:31 AM
Hi, David:

Our internal documentation clearly confirms that the low pass filter used in the EOS-1D Mark II N is identical to the one used in the EOS-1D Mark II.

Best Regards,

Chuck Westfall
Director/Media & Customer Relationship
Camera Marketing Group/Canon U.S.A., Inc.

dhphoto
November 22nd, 2006, 06:32 AM
Thanks for that Bart, unfortunately I no longer have the original 1DII for comparison

David

dhphoto
November 22nd, 2006, 06:34 AM
Hi, David:

Our internal documentation clearly confirms that the low pass filter used in the EOS-1D Mark II N is identical to the one used in the EOS-1D Mark II.

Best Regards,

Chuck Westfall
Director/Media & Customer Relationship
Camera Marketing Group/Canon U.S.A., Inc.

Really!

That is the info that I wanted, thanks very much Chuck.

So the reason the default files are so different must be down to picture styles or something else (because they surely are very different - even with the 1DIIN in 'faithful')

David

Nill Toulme
November 22nd, 2006, 06:42 AM
David I'm confused. You're say you're shooting RAW, but then you speak of picture styles. Unless you're processing with DPP, picture styles are (or should be) irrelevant. But unless you're processing/viewing your RAW files identically, you can't really make any judgment about whether one is different from the other, as otherwise any differences can obviously be explained by processing...

...or by sample variation between the two particular cameras, as opposed to any general difference between the two models.

Nill
~~
www.toulme.net

dhphoto
November 22nd, 2006, 06:49 AM
David I'm confused. You're say you're shooting RAW, but then you speak of picture styles. Unless you're processing with DPP, picture styles are (or should be) irrelevant. But unless you're processing/viewing your RAW files identically, you can't really make any judgment about whether one is different from the other, as otherwise any differences can obviously be explained by processing...

...or by sample variation between the two particular cameras, as opposed to any general difference between the two models.

Nill
~~
www.toulme.net

I'm viewing both the 1DII and N files 'as is' in C1, DPP & RSP, unsharpened in each case.

I realise the RAW file will, in some converters appear with the attributes given it by the 'faithful' picture style and I am trying to remove these to compare to the pre-picture style 1DII Raw files (for example they don't appear in C1 Pro). AFAIK you can't take a picture on an 'N' without using one or other picture style, even if it's one you've modified!

I no longer have the 1DII so can only go by work I shot previously, but all were shot at the same ISO's, with the same lenses and in the same environments and the N files are just SO MUCH punchier and more contrasty by default.

I'm not alone in thinking this, but from what Chuck said it must be something else that is doing it.

David

John Sheehy
November 22nd, 2006, 03:22 PM
I'm viewing both the 1DII and N files 'as is' in C1, DPP & RSP, unsharpened in each case.

Illusion all around. These RAW converters don't do "as is". Zero sharpening does not mean no sharpening; it means less sharpening, and the amount of sharpening at 0 epends on the camera. A single converter typically uses different recipies to convert different cameras.

Only programs like ImagesPlus, IRIS, MaximDL, and DCRAW will give generic conversions.

dhphoto
November 23rd, 2006, 12:22 AM
Illusion all around. These RAW converters don't do "as is". Zero sharpening does not mean no sharpening; it means less sharpening, and the amount of sharpening at 0 epends on the camera. A single converter typically uses different recipies to convert different cameras.

Only programs like ImagesPlus, IRIS, MaximDL, and DCRAW will give generic conversions.

It's not just the sharpness that is different, it's a quality that is difficult to describe - certainly the contrast from the N files is better too, as I said I'm not alone in thinking this

BTW AFAIK in RSP if you turn the sharpening all the way down, it really is 'off'

David

StuartRae
November 23rd, 2006, 01:09 AM
For RSP the difference in contrast,etc., could certainly be explained by the fact that it uses a different profile (tone curve?) for each camera model.

As you say, moving the sharpness slider to -50 is supposed to turn off sharpening. Additionally it's possible to turn off sharpening for converted files by unchecking the Apply Sharpening box in the Batch Convert tab.

Regards,

Stuart

dhphoto
November 23rd, 2006, 01:29 AM
I'm very much not the only person to have mentioned this.

If I use C1 Pro for example, with default sharpening the two different cameras seem to produce very different files.

There must be some differences as the N RAW files could not be converted in a pre N 1DII converter - needed an upgrade.

David

Jane Auburn
November 27th, 2006, 09:17 AM
I agree that the images I get out of the N are night-and-day better than those I got from the first Mark II I had. Servo focus is superior, far superior, as well.

I'd really appreciate an input from someone who knows, definitively, whether there is indeed a 'milder' AA filter in the 'N' as the images I am getting are far more pleasing straight out of camera (RAW of course) and need much less post-processing.

It seems there are two camps - 1DII owners (unsurprisingly) say no and 1DIIN owners (who previously owned the original 1DII) say definitely yes.

Can someone who knows please speak up?

Thanks

David

John_Nevill
November 27th, 2006, 02:20 PM
I cant really comment on the 1DII, but the 1DN does provide a very clean less noiser image than the 20/30D. Perhaps the lower pixel density helps.

Also, you might want to read this (http://www.eospix.com/?page_id=46), I did extensive testing last year to try and level the playing field for sharpness out of different RCs. I approached it methodically using imatest on the same 20D raw file and btw, RSP even when the sliders are set to zero, it still sharpens. The only way to prevent it is to turn it off in the preferences. The same holds true for NR. In fact I was quite surprised how much sharpening RSP adds by default.

This topic also brings us back to the old adage of pixel mania. Theoretically the 20/30D can resolve ~2050 lw/ph, while the 1DN is much lower, although to my eyes the 1DN RAW seems cleaner and sharper. Since sharpness is function of contrast, this no doubt is why the 1DN seems better.

Asher has also been banging on about better quality sensors for months and I have to agree with him, this megapixel mania is a folly.....sorry for the digression!

Jack Joseph Jr
November 28th, 2006, 09:09 AM
I too think that the 1D2N images look better than those from the 1D2. Of course there is a huge difference in JPEG quality, I replaced two 1D2 bodies for a 1D2N and a 5D simple because I have to shoot a lot JPEGs.

I really like the look of conversions from the current DPP but usually use ACR because of its superior tool set. Cropping/resizing with rotation is essential to my workflow. Standard Picture Style turns blue uniforms purple for some reason and DPP often deletes some of the shooting data (what kind of "professional" application is that!).

Using ACR, which is so dumb it can't even get the WB correct without some help, the 1D2N shots just look better. Of course I have no way of determining the differences that have been programmed into ACR based on the two camera models. Adobe had done a great job of improving image quality in ACR since my 10D days.

MiniRant - It would sure be great if the world's leading SLR company and the world's leading image processing company would work together so that their customers can produce accurate images.

Deborah Kolt
December 2nd, 2006, 10:42 PM
My experience with the two cameras is similar. In a very controlled environment - same gym, same lights, same lens, same settings, I switched overnight last winter from a Mark II to a Mark IIn. The difference was stunning. Much less noise at higher ISO (except possibly 3200, which I rarely have to use), much better contrast - the images are warmer and "pop" much more, and white balance straight out of the camera is far superior. I had intended to keep using the Mark II both as a backup and as a second lens. However, the image out of the N is so much more pleasing (and needs so much less post-processing), that I now almost never use the MII; I'll change lenses whenever possible in order to use the N. When shooting with both, it is immediately apparent which camera was used for a given pic.

From my experience, the N upgrade in image quality has been seriously understated. Thinking perhaps the Mark II needed adjustment, I sent it back to Canon for cleaning and a checkup. Still no comparison. I probably should sell it, since it seems to be logging more hours as a paperweight these days.

Asher Kelman
December 2nd, 2006, 11:56 PM
Deborah,

So you are referring just to the jpg files? You are not suggesting that the RAW files are also different, or are you?

Asher

dhphoto
December 3rd, 2006, 06:23 AM
From my experience, the N upgrade in image quality has been seriously understated.

My feelings exactly.

It isn't that you couldn't probably achieve the same result from either camera with careful processing of the RAW files, just that the 'N' seems to do it automatically.

I'm using *WAY* less USM to crisp up the 'N' files too (none usually :) )

David

Deborah Kolt
December 3rd, 2006, 11:40 PM
Yes, I'm referring to jpegs. Unfortunately, I don't have enough experience with raw files on the old 1DMII to compare. But the jpeg quality is dramatically better.

Jack Joseph Jr
December 4th, 2006, 11:45 AM
Having just finished an ACR processed job part of which is posted at http://jjjphoto.smugmug.com/gallery/2087661 I think that the N RAW files look better too. But let's face it, in reality RAW files have no look to them at all. Whether I use RIT, DPP or ACR each application has its own profile for every camera model so I can't really say whether the camera produces a better picture or whether the converter's camera profile is better.

Add to that the occasionally rumored, but never confirmed, improvement in the N's focus accuracy and it might be that the N's RAW images are sharper but otherwise the same as those from the older 1D2. Maybe it's just perception and human nature. How can you spend four grand on a new camera and not think that the pictures look better than those from the almost identical camera that was replaced?

Ferenc Harmat
December 21st, 2006, 06:48 PM
I'd really appreciate an input from someone who knows, definitively, whether there is indeed a 'milder' AA filter in the 'N' as the images I am getting are far more pleasing straight out of camera (RAW of course) and need much less post-processing.

David


...There is absolutely NO difference in the 1DMKII or 1DMKII-N in terms of sharpening requirements (both from a spectral and strength point-of-view), in RAW files, but that is where the story ends.

There ARE, however, BIG and SUBSTANTIAL differences between the output that these cam's on-board pipeline produce in .JPG format. The differences are tonal/contrast, chroma, spectral and noise-wise.

Spectrally speaking, the 1DMKII .JPGs allow you boost acuity to levels that are basically *unbelievable*. The N .JPGs, on the other hand, can not provide the same astronomical levels of acuity (much lower indeed), but provide a much, much stronger in the mid-frequency range (coarser detail such as textures on clothing, etc.) It also allows to extract an impressive amount of detail from the HIGHLIGHTS (not tonal, but frequency/spectral wise), as well as much, much better overall color, with excellent shadows saturation, and better balance of shadow detail-vs-shadow noise.

I have a few samples available to show this difference, both in synthetic ISO12233 output, as well as in real samples, where differences become even more compelling.

In the shadows is where chroma and noise differences will certainly bring the 1DMKII-N ahead, NOT by virtue of better analog-to-digital processing, but by virtue of much better management of the output, in terms of contrast curves, and Canon's new color processing approach.

In short, there are things that the 1DMKII .JPGs can do that the N would not, and vice-versa. Fortunately, I was able to process the N's output from the get-go, because I had the tools to address these marked spectral differences (that is, again, in .JPG, which I need as much as .CR2)

Happy shooting!

Jane Auburn
December 26th, 2006, 06:24 AM
How can you spend four grand on a new camera and not think that the pictures look better than those from the almost identical camera that was replaced?

Easy. I use my brain and my eyes. Four thousand dollars does not affect my ability to be objective.

Alan T. Price
December 28th, 2006, 07:24 AM
The differences are pretty much due to the interpretation of camera settings as suggested by John Sheehy. You may well have taken two photos with the same setting values, but the way the cameras interpret those values may well be different. The way raw converters interpret them may well be different also, especially if the raw file itself does not explain the precise meaning of every one of the settings to the software.

This effect would be applicable to all relevant settings including colour saturation, sharpness, contrast, WB, etc. It would be far more significant to jpeg files than raw files because in jpegs the cameras have already applied the settings.

DPP makes it easy to apply picture styles to raw files but you may have noticed that changing a picture style setting does not change the other settings such as saturation and sharpness.

That means we are left to experiment with different settings and combinations of settings in a raw converter that is unaware of styles if we are to compare match setting values in one camera to the corresponding setting values in the other camera.