DxO and A900 self-interest thread

[kuau]

Yes, exactly. I personally use the creative styles extensively as I have found files need no to very little tweaking after I select the right style. My favourites are standard, landscape and vivid. Whatever you do, make sure you turn off noise reduction for each file individually. The default setting is auto, which is equivalent to NR off in the camera (which still does some NR even on this setting).

Do I turn NR off on the camera or in the Sony software? When you select a creative stye on the camera does it automatically turn on NR?

 

[edwardkaraa]

Do I turn NR off on the camera or in the Sony software?

Both

When you select a creative stye on the camera does it automatically turn on NR?

No, on the camera it is a separate setting, but in IDC you'll have to do it every time you switch creative styles.

 

[kuau]

What do the DxO Optics Correction Modules actually do? I have the CZ16-35mm and I finally broke down and got the CZ24-70 used on ebay today for 1395.00 plus I got 8% cash back from live.com Both of theses lenses are listed now at Dx0

Steven

 

[Quentin_Bargate]

What do the DxO Optics Correction Modules actually do? I have the CZ16-35mm and I finally broke down and got the CZ24-70 used on ebay today for 1395.00 plus I got 8% cash back from live.com Both of theses lenses are listed now at Dx0

Steven

Hi,

They remove lens defects: CA, vignetting, and correct for distortion.

Quentin

 

[douglasf13]

Edward, the problem with using IDC and the camera's creative styles is that you're not getting an accurate histogram representation of the RAW data, and you're more than likely underexposing more than you realize. If you have zero issues with the camera's noise handling, then this is a non-issue, I guess. However, if you have qualms with the A900's noise, it's essential to set the camera's jpegs settings to achieve the closest histogram to RAW that the camera can muster, and leave it there. This will reduce the noise quite a bit. When using IDC in your workflow, you're taking a shot underexposed, then the camera jpeg engine boosts the histogram with all of it's "special sauces" like contrast, saturation, etc. When you open the RAWs in IDC, it also applies the same "special sauces," leading you to believe that you were exposed properly all along, when, in actuality, your shot was underexposed, and both the camera jpeg engine and IDC boosted the shot, causing the histogram to look correct. This gaining up of the underexposed shot adds lots of noise.

Does all of this babbling make any sense?

 

[jonoslack]

Hi There
such interesting information . . . but I'm wedded to Aperture, I know it isn't perfect, but it does such a good 'general' job for me with the A900 and with the M8, and with a library of thousands of keyworded images . . .

 

[picman]

Hi Bob,

1. I am getting a black dotted rectangle on mine

2. clicking outside (with the crop too still selected) works fine for me

3. Agree on this: I can't sem to get rid of the edges that I have cropped. With Silkypix, it was easy to discard the cropped area

Cheers

Quentin

Thanks Quentin! Are you on Mac or PC?

Cheers, Bob

 

[Quentin_Bargate]

Thanks Quentin! Are you on Mac or PC?

Cheers, Bob

PC
Quentin

 

[picman]

PC
Quentin

Aha, that might explain it, I am on Mac. I'm going to send in a support ticket to DxO.

Thanks again, cheers, Bob.

 

[edwardkaraa]

Edward, the problem with using IDC and the camera's creative styles is that you're not getting an accurate histogram representation of the RAW data, and you're more than likely underexposing more than you realize. If you have zero issues with the camera's noise handling, then this is a non-issue, I guess. However, if you have qualms with the A900's noise, it's essential to set the camera's jpegs settings to achieve the closest histogram to RAW that the camera can muster, and leave it there. This will reduce the noise quite a bit. When using IDC in your workflow, you're taking a shot underexposed, then the camera jpeg engine boosts the histogram with all of it's "special sauces" like contrast, saturation, etc. When you open the RAWs in IDC, it also applies the same "special sauces," leading you to believe that you were exposed properly all along, when, in actuality, your shot was underexposed, and both the camera jpeg engine and IDC boosted the shot, causing the histogram to look correct. This gaining up of the underexposed shot adds lots of noise.

Does all of this babbling make any sense?

Douglas,

I have not really run into this problem, but probably because I neither use the camera meter nor the histogram. I follow the guidelines of DXOmark on A900 iso values which are 1/2 stop less than claimed for the entire iso range except 100 which is 1/3 over, and use an external incident light meter for most of my shots. Doing this, I get right on exposures that I rarely need to tweak in the raw converter.

By the way, I have downloaded trial versions of DXO, Bibble 5, rawtherapee, Silkypix, Acdsee, and still subjectively prefer the output of IDC. Some of these converters may produce sharper pixels, or lower noise, but what matters most to me is the general look of an image, and for me, IDC is still unbeatable. There are of course some additional important features that I would like to see in future upgrades of IDC, such as dust removal, CA removal, arbitrary rotation and more advanced cropping.

 

[douglasf13]

That is kind of my point. If you're getting what looks like perfectly exposed shots with the default settings in ACR, IDC, or C1, etc, then you probably are underexposing. Do me a favor and take a perfectly exposed RAW that you have, open it in C1, and click the linear curve, rather than film like curve. If your histogram isn't snug up against the right side, then you'll have an idea whether or not underexposure is an issue. Exposing for your RAW converter is very helpful, and something I hadn't considered until recently, but it makes a big difference.

 

[picman]

Just got a reply from DxO support confirming all my findings. Apparently the PC and Mac versions work differently and the manual is written for the PC version. They said they would try to get a new one out which covers both platforms.

Cheers, Bob.

 

[edwardkaraa]

That is kind of my point. If you're getting what looks like perfectly exposed shots with the default settings in ACR, IDC, or C1, etc, then you probably are underexposing. Do me a favor and take a perfectly exposed RAW that you have, open it in C1, and click the linear curve, rather than film like curve. If your histogram isn't snug up against the right side, then you'll have an idea whether or not underexposure is an issue. Exposing for your RAW converter is very helpful, and something I hadn't considered until recently, but it makes a big difference.

Ok, now I see what you mean. But isn't this common to all cameras? I have done linear conversions in the past but found it too time consuming and the results were no better than those obtained with the standard conversions. Aren't linear conversions supposed to look very dark? The ones I did long ago with the original 1Ds files looked this way.

 

[douglasf13]

Yeah, that's what I thought, too. I've come to realize that it's because Im underexposing, and the camera histogram was tricking me into believing everything was fine. Now, when I open a RAW in C1, it looks overexposed with blown highlights, but when I switch to linear, everything is fine. This essentially allows more exposure to the right, resulting in less noise. I realize that many don't mind the A900 noise, but for those that do, this improves things a lot.

Another issue with these cameras is where they place middle grey. With the A900, exposing for the highlights at +2.5 is about right. With the 5dii, it's more like +3.5, if I remember correctly.

 

[kuau]

Yeah, that's what I thought, too. I've come to realize that it's because Im underexposing, and the camera histogram was tricking me into believing everything was fine. Now, when I open a RAW in C1, it looks overexposed with blown highlights, but when I switch to linear, everything is fine. This essentially allows more exposure to the right, resulting in less noise. I realize that many don't mind the A900 noise, but for those that do, this improves things a lot.

Another issue with these cameras is where they place middle grey. With the A900, exposing for the highlights at +2.5 is about right. With the 5dii, it's more like +3.5, if I remember correctly.

Douglas, using basic daylight rule, on a sunny day outside, iso 200
so Shutter Speed 200 F16 how much can you over expose the image before you blow out the high lights? Do you base this all in C1 and Linear Response curve?

Just curious. I have ben using DxO lately for lens correction stuff but I guess I can come out of DxO as dng then do the exposure stuff in C1?

Steven

 

[douglasf13]

Douglas, using basic daylight rule, on a sunny day outside, iso 200
so Shutter Speed 200 F16 how much can you over expose the image before you blow out the high lights? Do you base this all in C1 and Linear Response curve?

Just curious. I have ben using DxO lately for lens correction stuff but I guess I can come out of DxO as dng then do the exposure stuff in C1?

Steven

To be honest, I haven't used sunny 16 much since going digital, since it is an estimation, and my cameras have meters built in (or i have a meter with me.) I usually spot meter the highlights, and spot metering the A900 fully detailed highlights at around +2.5 EV is about right on (defocus the highlight before metering***) This works in Raw Therapee and C1 with a linear curve, and the histogram in RAWanalyze is right on. In lightroom, with all sliders set to zero and a linear curve, you still have to set the exposure to around -1.0 EV in the converter, because Adobe is adding some kind of boost (I believe it has to do with how they apply WB.) This is part of the reason that some say LR has such great highlight recovery....because lightroom is actually boosting your shot around a stop to begin with!

I stay away from DxO, because, like LR, DxO adds a baseline NR even when all NR is set to off. Actually, I stay away from most converters that do a lot of special recipe stuff under the hood, because it adds too many variables. As far as DNG, Iliah Borg has gone into a lot of detail as to why it's not great, and, while I can't follow the reasons 100%, it's enough to convince me not to use it. :lecture:

Cheers, douglas


***defocusing the highlights makes them duller by about -.5 EV. In actuality, the green channel on the A900 blows at about +3.2 EV. Defocusing gets you to around +3 in reality, and the last .2 EV leaves room for flare and sample variation. If you want traces of detail in the highlights, rather than full detail, meter at +2 2/3 EV rather than 2.5. If you want to use the midtones to meter, spot meter +.5 EV.

p.s. i am a nerd! :ROTFL:

 

[dhsimmonds]

I have just checked the DxO site and it seems that virtually all of the full frame prime and zoom Zeiss lenses and the Sony 50 F1.4, 70-200, 70-300, 70-400 lenses are now supported by the Pro Elite version. I might just give the trial version a run before buying it.:thumbs:

 

[douglasf13]

I have just checked the DxO site and it seems that virtually all of the full frame prime and zoom Zeiss lenses and the Sony 50 F1.4, 70-200, 70-300, 70-400 lenses are now supported by the Pro Elite version. I might just give the trial version a run before buying it.:thumbs:

fwiw, there is a current forum over on dpreview sony forum showing that, like ACR, DxO always does a little NR whether it is turned off or not. Also, there seems to be a problem with green casts in some cases.

 

[Quentin_Bargate]

fwiw, there is a current forum over on dpreview sony forum showing that, like ACR, DxO always does a little NR whether it is turned off or not...

Honestly, who cares? The end result is what matters and I rate DxO top of the pile in that regard :thumbup:

Quentin

 

[douglasf13]

I'm the opposite. Everytime I see a DXO comparison, the detail of DxO is a bit mushy. However, I could just be seeing bad examples.