Striations in P45+@800 ISO normal ?

[Dustbak]

Flexcolor also does tend to underexpose slightly compared with ACR, so does Leaf Capture but there we are talking in the range of about half a stop not 2. I always take this into consideration and overexpose slightly while working with Flexcolor (as I did with LC as well).

With Flexcolor I once made the mistake of not setting my exposure back to normal (0), I also once made the mistake of not setting my levels back to default. This resulted in images way off. Could something like this have happened to C1?

I have tried a demo version of C1 about a year ago. I did not like it and have not looked at it again so I cannot say anything about it.

Looking at this I am convinced this image would have looked excellent if it had about 2 more stops.

[stevecoleccs]

No Problem with your Back - just don't underexpose at a high iso - I shoot MOST
of the time with my 30+ at iso800, very clean files, as long as I do not underexpose.

David,

On second thoughts, I guess you're right. I'll dig out a tripod tomorrow and let down the blinds

Edmund
[a href=\"index.php?act=findpost&pid=152918\"][{POST_SNAPBACK}][/a]

[Ray]

I find it quite extraordinary that such professional photographers who spend so much on their equipment should be so ignorant of these basic principles of ETTR.

My mind boggles. Should they all be sacked and sent off to Siberia?

[Jonathan Wienke]

No, but it does demonstrate that just because you photograph models that have a $10000 day rate doesn't automatically mean you have a good grasp of the technical side of things. I find this rather amusing after all of the pomposity posted in the MFDB/ETTR thread. This is an excellent example of why ETTR is a good idea, even for MFDB shooters.

[Dustbak]

I somehow knew you would bring that up again.

[samuel_js]

You still have +1.85 step for this image to be exposed "to the right". Still, the blue channel is clipped in the shadow areas. This is because I'm recovering the exposure and this details has been lost. If you continue to underexpose your captures you'll always find this problem and a lot of noise.

[eronald]

Samuel, I can see the stripes on your screenshot - has anyone else seen this type of fault ?

I think the fruit file ends up with around 230,230, 230  RGB values on the business card, if it is pushed through C1 with the standard film curve (not linear). This is certainly not severe underexposure.

I would be sure, but C1 just crashed on me yet again. It really doesn't like files from my back somehow. I have tried downloading C1 again, but that hasn't helped, it just crashed again. I can post my saved copy of the rendered Tiff if that helps.

I really wonder whether my files aren't corrupted somehow. I have tried using a different memory card but that didn't change much.

Edmund

[Dustbak]

Edmund,

Try exposing for ACR and see how that ends up. In the fruit file the Orange clips first with ACR (the spectacles on the orange that is ofcourse) ! While the business card is still below 255,255,255.

[Jonathan Wienke]

TRC? Are you referring to this?

Yes. You want the TRC set to linear while adjusting the Exposure slider because if you have a curve set to give a smoother rolloff to the highlights and shadows, a large exposure adjustment may make only small adjustments to the white and black points, which makes it much more difficult to tell if you got exposure correct in-camera or not. OTOH, with a linear TRC the movement of the white and black points correlates much more closely with movement of the exposure slider and the amount of exposure in the camera, and it's a lot easier to tell from the RAW converter histogram whether the shot should have had more or less exposure. Once exposure is set optimally, then one can tweak the TRC as needed to get whatever  look is desired from the image.

[samuel_js]

No, but it does demonstrate that just because you photograph models that have a $10000 day rate doesn't automatically mean you have a good grasp of the technical side of things. I find this rather amusing after all of the pomposity posted in the MFDB/ETTR thread. This is an excellent example of why ETTR is a good idea, even for MFDB shooters.
[a href=\"index.php?act=findpost&pid=155535\"][{POST_SNAPBACK}][/a]

Not going again into this topic, but this is just false again. In medium format digital you need to "expose right". Two stops under clipping is an underexposed image. It has nothing to do with ETTR. The histogram of a correct exposure in MFD looks centered.

[Panopeeper]

This image, CF000965, is less underexposed:



compared to the former one, CF000598:



However, it still tops out at level 7500. In fact, only 0.0168% of the green pixels are over 7300.

If the DNG converter's information is true, namely that the numerical range is de facto 0-65535, then it is still 2.95 stops away from the right end.

However, the dark end is not as bad as the other image is: "only" 10.5% of the pixels are zero (i.e. totally lost), but 47% of the pixels are under level 50. This makes the greyish cloth of this jacket or whatever not only noisy but off-color.

Anyway, is the streak you are referring to the one between the arrows here?

Crop of streaky area

[eronald]

Panopeeper, are you sure your software is processing the files in a way similar to the standard Raw converters ?

In the C1 histogram one can clearly distinguish the black cloth, at the left, and the fruit at the center, and the business card more to the right.



Edmund

This image, CF000965, is less underexposed:



[a href=\"index.php?act=findpost&pid=155547\"][{POST_SNAPBACK}][/a]

[Panopeeper]

Panopeeper, are you sure your software is processing the files in a way similar to the standard Raw converters ?

I am sure it does not. It is not a raw processor but a raw analyzer. The histogram I posted reflects the unadultered raw data, that what your camera created, not how it is interpreted. The histogram after setting the white point for the card, white balancing on the card and mapping to sRGB without any other adjustment would look so:



The histogram you see with C1 shows the outcome of the conversion.

[eronald]

Panopeeper

It would seem that using linear scales will push all low-key imagery well into the left 10% of any linear graph. After all, every low-key image has most data 4 stops below the max.

Maybe some base-2 log scales (f-stops) - on the x axis- would be more appropriate to photographic usage, albeit less spectacular ?

Edmund

I am sure it does not. It is not a raw processor but a raw analyzer. The histogram I posted reflects the unadultered raw data, that what your camera created, not how it is interpreted. The histogram after setting the white point for the card, white balancing on the card and mapping to sRGB without any other adjustment would look so:



The histogram you see with C1 shows the outcome of the conversion.
[a href=\"index.php?act=findpost&pid=155559\"][{POST_SNAPBACK}][/a]

[bcroslin]

Ronald,

Go back and open up 1 1/2 stops at ISO 800 and then post that raw file.

I'm really stunned that you don't understand that when you underexpose an image nearly two stops with any digital camera that you're going to get all sorts of funkiness.

Furthermore, the concept of pushing and pulling development might work with film but I've never seen it work the way you're wanting it to with digital.

[Morgan_Moore]

Show us a straight curve on the pic of the woman..

Theses images are way under by the look of it

I am a flash and reflector nut because IMO most scenes need the DR compessing to get a good file ie to lift the shadows that you want to keep

With the card and fruit pic the card needed to be scraping the top of the curve or even 'recovered' by doing HDR

Do some test on recovering highlights - exellent in sinar - dont know about phase

Low key look is created by making a deep S in the Raw Curve not be exposing like that

I assume you expose by chimping the histo not checking the camera meter

and base you noise expectations that the 100ISO base of the P45 is being pumped up with EIGHT times less light than the sweet spot

I have friend who is very happy that his 45 is good at 200

My calculations on getting clean 800 revolve around the D3 which is being trumpeted as 25000000 ISO

S

[godtfred]

So, is this contretemp really all about differences in the way C1 and ACR handle images? A one and 2/3rds stop difference to me is so significant I find it difficult to believe that any major program like C1 could operate like that.

Does the ACR image not look good? There are hundreds of adjustments in ACR you can make if you think the image is not quite right.
[a href=\"index.php?act=findpost&pid=155509\"][{POST_SNAPBACK}][/a]

I started out not believing what my eyes told me in C1, but compared to ACR and Flex, it pushes midtones and highlights upwards between 1 and 2 stops. I frequently push midtones in Flex about 1,05 to 1,15 on the levels adjuster to get a pleasant "modern" result, but don't blow any highlights etc. The same is not true for C1 where i have to "pull back" instead. I have used the H3D-39 since its release date, as well as the H2+CFH-39 before that. I am now in my 3rd week using C1 and my P45+ (not a long time, I know). Yet as far as defaults go, C1 pushes midtones and highlights a lot, and I'm beginning to wonder if this is not the reason why so many users find the C1 conversions pleasing. Its a "fully pushed and pulled" file straight away, contrary to Flex where you start more closely to zero, and work your way up.

-axel

EDIT: See the attatched files for examples of what I mean using Edmunds file as a basis. I can see how someone could be fooled by the histogram in C1 if using it as a basis for determining exposure, as I was initially. The furthermost left part with information in it, is just a stop or two shy of white clipping, but in reality is quite a lot more underexposed.

[attachment=4025:attachment]
[attachment=4026:attachment]
[attachment=4028:attachment]
[attachment=4027:attachment]

[eronald]

Axel,

 Seeing you have a P45+, could you post some underexposed shots at hi iso please ?

Edmund

I started out not believing what my eyes told me in C1, but compared to ACR and Flex, it pushes midtones and highlights upwards between 1 and 2 stops. I frequently push midtones in Flex about 1,05 to 1,15 on the levels adjuster to get a pleasant "modern" result, but don't blow any highlights etc. The same is not true for C1 where i have to "pull back" instead. I have used the H3D-39 since its release date, as well as the H2+CFH-39 before that. I am now in my 3rd week using C1 and my P45+ (not a long time, I know). Yet as far as defaults go, C1 pushes midtones and highlights a lot, and I'm beginning to wonder if this is not the reason why so many users find the C1 conversions pleasing. Its a "fully pushed and pulled" file straight away, contrary to Flex where you start more closely to zero, and work your way up.

-axel
[a href=\"index.php?act=findpost&pid=155638\"][{POST_SNAPBACK}][/a]

[josayeruk]

Even in C1 you are clipping the blacks.

Probably get away with it on 50 ISO but on 800... No chance!  

Reshoot!

Jo S.x

[John Sheehy]

Even in C1 you are clipping the blacks.

Probably get away with it on 50 ISO but on 800... No chance!   

Reshoot!

[a href=\"index.php?act=findpost&pid=155735\"][{POST_SNAPBACK}][/a]

There is no such thing as "clipping the blacks", as far as user expsoure is concerned.  A RAW histogram is not a symmetrical entity; the left side is anchored, and the right side is scalable.

However, bad cameras and converters can clip black, but that is independent of exposure.