Knut.
Offline
• • •
Upload & Sell: Off
|
 p.9 #4 · Why Bother Shooting RAW with Cheap Lenses? | |
Steve Spencer wrote:
I think you are misunderstanding a few things here:
1) When the camera converts a file to jpeg, it still uses the RAW file to do so. Remember the RAW file is simply what is read off the sensor, so of course any file has to start with that. The difference between saving in RAW or jpeg is just whether the camera does the processing or the user does the processing with the aid of software.
2) Distortion, vignetting, and lateral CA whether done in camera or automatically as the RAW file is read into the processing software is done by starting with the RAW file and applying automatic corrections. Only if you tweak these automatic corrections (and for these aberrations I almost never do) does it matter whether the camera applies them or the post processing program applies them. Sometimes the camera might have slightly better corrections and sometimes the software might have slightly better corrections, but I have never seen that. These corrections seem to be supplied by the camera maker to the RAW processing programs and work the same way when applied by the camera or the RAW processor. Of course the camera has these corrections from day one and sometimes there is a bit of a delay before a given RAW processor has them as this thread has discussed but once they are available in both they seem to work exactly the same way: the camera or processor applies the same adjustments automatically from the same RAW file read off the sensor.
3) Slower aperture is another matter. A slower max aperture makes it easier to make a lens with fewer aberrations. When a camera maker makes a lens with a narrower max aperture, they can typically make it both cheaper and with fewer aberrations than a faster lens with a wider max aperture. Wider aperture lenses, which typically cost more, therefore often need more corrections than narrower aperture lenses that cost less. These more expensive lenses may benefit more from shooting in RAW because they benefit from non-automatic corrections and more sophisticated processing techniques like local adjustments and layers.
The relation of cost with the need for non-automatic processing of files is probably a pretty weak relation as some very expensive lenses need very little correction with non-automatic processing but some types of non-automatic processing (like trying to correct axial CA or purple fringing) are likely to be needed with fast aperture lenses and these will typically cost more than slower aperture lenses....Show more →
I do not believe that I misunderstand things. It boils down to the question, which raw converter is superior: the in camera raw converter or the raw converter I use in post. In addition, will the raw converter in-camera do certain processing steps for a given lens at alI?
I uderstood the OP in the sense: If you have a cheap lens (third party lenses, legacy lens, adapted lens; whatever), is the extra effort to process it in post worth it? Or are the aberrations of cheap lenses such as nuisance, that it isn’t worth the effort?
It now really comes down to what has made the cheap lenses cheaper.
The practical implications of the statements I make are limited to the existing combination of camera, lens and raw converter (in my case DXO and Sony A1, and many legacy/third party lenses I use). But I imagine, that it can be generalized to other combinations.
1) In camera noise reduction is definitely poorer than noise reduction in post. If the cheaper lenses comparison boils down to, for example:
- GM 50/1.4 II, image shot at f2 and ISO 1600 versus
- Cheap 50mm/ff2.8 lens and ISO 6400
=> I find that putting the raw image through DXO‘s top noise reduction algorithm salvages nearly 2 stops in noise, which I do not get when using the in camera jpeg conversion of the A1.
2) Many third party lenses do not get vignetting correction when I process these images from raw to jpeg in camera. Considering the 8 bit colour depth of jpegs and assuming 2-3 stops that the corners need to be elevated on wide and ultra wide angles, this loses 1-2 bit of colour depth in the corners. Crunched colours, that I may avoid losing when processing a raw file in post.
3) LACA: On many lenses which I use on the Sony A1, the A1 does not correct lateral chromatic aberations. Thus LACA correction is not available at all, if I use in camera body raw to JPEG conversion on many cheap lenses I own.
4) Same story concerning distortion correction: It is just not done in camera for many lenses. If I do it with DXO, result are better if I work on raw files. (I know it is available for the more expensive orignial Sony lenses and some third party lenses like Sigma).
Things apear to boil down to the definition of the cheaper lens: If it is a high-tech lens (for lack of a better wording) that has the relevant parameters (vignetting, distortion etc.) stored in the lens and the lens can pass this data to the camera and the camera accepts this information for processing, then the camera may well provide a reasonable raw to jpeg conversion (but still have poorer noise reduction). If this does not work, as with many cheap legacy lenses or often with lenses of Chinese provenence, then processing raw files in post is the best approach.
I say this with some frustration, since I also do not like the time effort it often takes to do this raw to jpeg conversion in post. I often skip it, make do with the in camera jpegs, but go back to the stored raws when I want to print some of these images.
|
