How Important is Shooting in Raw?

[Newbie444]

I am so annoyed shooting in Raw. I don't even know what it does, to be honest. I think it makes editing a lot easier? I just find shooting in Raw is more of an annoyance than a help, because I have to convert to a jpg. anyway, because barely anything is compatible with Raw, I find. So should I just not shoot in Raw?

 

[bobn2]

I am so annoyed shooting in Raw. I don't even know what it does, to be honest. I think it makes editing a lot easier?

Raw affords you more latitude. For starters you can change WB, brightness, saturation, contrast, tonal curves, sharpness, and noise reduction.

I'm not trying to be pedantic, but sometimes using the wrong term really does embed a way of thinking that makes it hard to understand. Raw doesn't allow you to change WB, brightness, saturation, tonal curves etc. It enables you to establish what those should be in the first place.

Okay, then shooting Raw allows you to change your mind about those things, or to experiment with them at your leisure.

I think that is better. I wasn't getting at you, I think the way you said it is reasonable until you begin to think what other people might get from it. A long history of discussions here, when you get played back such explanations as the cause of misconceptions has led me to think maybe that one just needs to be aware of other possible interpretations than your own.

The whole point is that a raw file has no defined white balance (apart from WB preconditioning in some cases),

Just to be clear, I have never seen any difference in a Raw file by changing the camera's WB setting.

Then presumably you haven't used a camera with WB preconditioning.

WB preconditioning is a new term to me. This implies that the raw file is not "raw" but is cooked According to some camera presets. Which cameras might do this? This goes against my understanding of what a raw file should be.

It was/is a common Nikon practice and according to a recent post in another thread has been adopted by Canon. It kind of makes sense in the case that ADC range is somewhat limited, not much otherwise. In the beginning, it wasn't 'cooking' the raw, since it was implemented as an analog gain change implemented differentially between the channels (which worked well with CCD read-out). But, I suspect that seeing it as 'cooking' is a bit of a misunderstanding of what raw is anyway. Every camera includes some signal processing of the data from the sensor, whether analog or digital. There is no standard for raw, it's the manufacturer's choice at what arbitrary place in the chain they decide to dump the data.

 

[Tony Beach]

The whole point is that a raw file has no defined white balance (apart from WB preconditioning in some cases),

Just to be clear, I have never seen any difference in a Raw file by changing the camera's WB setting.

Then presumably you haven't used a camera with WB preconditioning.

I wouldn't necessarily make that presumption. I have a Nikon D800 and have used the Sony A850, Nikon D300, Nikon D200, and Nikon D70. I have not seen any difference in files from those cameras as a result of using different WB settings (I usually, but not always, use UniWB). That doesn't mean it might not show up if I push my WB settings in an extreme way or measure the results in a minute way, it just means I haven't seen it.

It is my understanding that WB preconditioning likely does occur on my D800, but that changing the WB does not change that.

 

[bobn2]

...

I think the point AmazingLight was making was that using the default settings as his processing context, a RAW file can result in a "flat" and/or "soft" image.

What would those 'default settings be'. Until processed a raw file has no defined meaning, in terms of how you would see the image.

The "default" settings are whatever your raw processing software use when you don't otherwise change them.

Which means nothing at all, since you don't know what the raw processing software is, what its designers thought the 'defaults' should be or, indeed, which camera or brand you are using. So, there is no universal 'default' which would justify the statement that 'a RAW (sic) file can result in a "flat" or "soft" image', which is, I'm afraid just non-factual.

I believe that a beginner can in fact open a RAW file in PhotoShop, iPhoto, Afintity Photo, etc., without altering any of the offered settings.

I would call this using the "defaults".

'Default' for that particular piece of software, not default for the

Obviously, he will probably not somewhat different images, depending on which software he uses.

So, very clearly the 'default' applies to the software, not the raw file.

Where we differ is whether it makes sense to talk in generalities about the results.

No, that's not where we differ. Where we differ is that you were attributing this 'default' behaviour to be a property of the raw file and trying to use it to pretend that the raw file in some way is in a defined colour space (if loosely defined).

As the above is a workflow that a beginner may first try when shooting RAW, it seems to me that it does make sense to discuss what can happen when you use default settings.

Only if you first talk about which software is being used.

Whether the default settings in some software does in fact produce "flat" or "soft" images is a discussion that I think is reasonable.

That wasn't what you were doing though, and the idea that some undefined software might produce 'flat' images is of no help to anyone.

I think your point was that one shouldn't assume any particular context for processing RAW files, and hence it doesn't make sense to talk about what they "look" like. Depending on the processing, they may result in an image that is flat or vibrant, soft or sharp, etc.

No, that wasn't my point. My point was that until processed, a raw file doesn't have any defined human perceptual meaning.

There is a difference between "any defined meaning" and a "well defined meaning".

I qualified it advisedly, because obviously, since you can use a raw file to create a file with a well defined perceptual meaning, there is some kind of link. But the point is, the numbers in the raw file don't define the 'look' of the image without a set of assumptions the processor has to make, which are not in any way 'default'.

Again, our difference boils down to whether or not the software used by a beginner has default settings.

No, it depends on whether you think investing a characteristic of the software to the file helps anyone understand what is going on. I suspect it just 'helps' them get as confused as you seem to be.

Even without a specific context a raw file will have some associated meaning. Given the embedded metadata, there are areas of the image which will be green, blue, red, etc. Some areas will be darker, and others will be lighter.

What we don't have is an "exact" meaning. Without knowing the particulars of the raw processing, we don't know exactly which shade of green, blue or red we will get. If the particulars are unusual, we might actually get something that isn't green, blue or red.

That amounts to what I said above, clearly since you can derive a perceptually defined link there is some meaning, which is why I qualified the statement as I did. I get the impression now that you're trying to squeeze through that qualification so as not to be wrong. Sure, you can find a route by which what you said isn't complete garbage, but that's actually denying the intention of it, and also ignoring its misinformation potential, which is why I challenged it.

Let's back up a bit. We got here when I said "I think the point AmazingLight was making was that using the default settings as his processing context, a RAW file can result in a "flat" and/or "soft" image."

I stand by that statement. (Other than the typo it should have said "...in his processing...")

We got here in more ways than that. As it happens I agree that might have been what he was meaning to say, which means that he definitely needs the help he was seeking.

 

[Tony Beach]

no reason not to use lossless compressed.

When a Huffman gets corrupted, recovery is close to impossible. With uncompressed formats, on the other hand, the damage is local.

Oddly enough, if the proper backup procedure is in place, that is the only question is the disk size/model and how/when to round-robin the disks, it doesn't matter what is the size of the raw.

But when (and that's mostly how it is, especially in the field - that's why we receive massive recovery requests, and often the only thing we can help with is extraction of embedded JPEGs) the procedure is consumer-level or simply absent, uncompressed files are the safer option.

Isn't the source of corruption nearly always the card? I've had bad cards, which I've replaced, and other than that I can't think of a single incident of a corrupted file (that would be out of tens of thousands).

 

[bobn2]

The whole point is that a raw file has no defined white balance (apart from WB preconditioning in some cases),

Just to be clear, I have never seen any difference in a Raw file by changing the camera's WB setting.

Then presumably you haven't used a camera with WB preconditioning.

I wouldn't necessarily make that presumption. I have a Nikon D800 and have used the Sony A850, Nikon D300, Nikon D200, and Nikon D70. I have not seen any difference in files from those cameras as a result of using different WB settings (I usually, but not always, use UniWB). That doesn't mean it might not show up if I push my WB settings in an extreme way or measure the results in a minute way, it just means I haven't seen it.

It is my understanding that WB preconditioning likely does occur on my D800, but that changing the WB does not change that.

Loading…

www.libraw.org

Actually, I think from looking around that you're right - at least in the Nikons the data is scaled, but the scaling isn't changed according to the WB. Altogether a crazy thing to do, typical of what teams of engineers get into once they've decided something is the way to do things.

 

[Iliah Borg]

Isn't the source of corruption nearly always the card?

The disastrous cases we know are caused by pseudo-reliable backup systems, whole collections are gone. That's why I use properly configured ZFS (FreeNAS is very much OK).

Second place goes to the card reader / interface problems (and people do not check copies they made to HDD before erasing and overwriting the cards).

Bad cards next, and the last place - card interface in the camera.

 

[Tony Beach]

I think your point was that one shouldn't assume any particular context for processing RAW files, and hence it doesn't make sense to talk about what they "look" like. Depending on the processing, they [sic] may result in an image that is flat or vibrant, soft or sharp, etc.

No, that wasn't my point. My point was that until processed, a raw file doesn't have any defined human perceptual meaning.

There is a difference between "any defined meaning" and a "well defined meaning".

I qualified it advisedly, because obviously, since you can use a raw file to create a file with a well defined perceptual meaning, there is some kind of link. But the point is, the numbers in the raw file don't define the 'look' of the image without a set of assumptions the processor has to make, which are not in any way 'default'.

Again, our difference boils down to whether or not the software used by a beginner has default settings.

No, it depends on whether you think investing a characteristic of the software to the file helps anyone understand what is going on.

We get it here all the time at DPR, or we used to anyway, and this site encourages it with their boilerplate approach to converting Raw files from cameras and then offering them up in their studio "comparison." Until recently I was downloading those files to look at them myself and making the mistake of accepting the default sRGB color space, but then I opened a Nikon file using my preferred AdobeRGB color space and what a difference that made as the colors from the sRGB color space are noisier (I would call them muddier) and flatter too.

This is what I got from one of the files:

[IMG width="400px" alt=""Default" sRGB on the top of the patches, AdobeRGB on the bottom of the patches."]https://photos.imageevent.com/tonybeach/mypicturesfolder/sharing/DSC_0532.jpg[/IMG]
"Default" sRGB on the top of the patches, AdobeRGB on the bottom of the patches.

 

[Michael Fryd]

...

I think the point AmazingLight was making was that using the default settings as his processing context, a RAW file can result in a "flat" and/or "soft" image.

What would those 'default settings be'. Until processed a raw file has no defined meaning, in terms of how you would see the image.

The "default" settings are whatever your raw processing software use when you don't otherwise change them.

Which means nothing at all, since you don't know what the raw processing software is, what its designers thought the 'defaults' should be or, indeed, which camera or brand you are using. So, there is no universal 'default' which would justify the statement that 'a RAW (sic) file can result in a "flat" or "soft" image', which is, I'm afraid just non-factual.

I believe that a beginner can in fact open a RAW file in PhotoShop, iPhoto, Afintity Photo, etc., without altering any of the offered settings.

I would call this using the "defaults".

'Default' for that particular piece of software, not default for the

Obviously, he will probably not somewhat different images, depending on which software he uses.

So, very clearly the 'default' applies to the software, not the raw file.

Where we differ is whether it makes sense to talk in generalities about the results.

No, that's not where we differ. Where we differ is that you were attributing this 'default' behaviour to be a property of the raw file and trying to use it to pretend that the raw file in some way is in a defined colour space (if loosely defined).

I think our difference is that you are talking about the file, and I am referring to the system as a whole.

That system does include the RAW processing software.

As the above is a workflow that a beginner may first try when shooting RAW, it seems to me that it does make sense to discuss what can happen when you use default settings.

Only if you first talk about which software is being used.

Whether the default settings in some software does in fact produce "flat" or "soft" images is a discussion that I think is reasonable.

That wasn't what you were doing though, and the idea that some undefined software might produce 'flat' images is of no help to anyone.

Whether or not the idea is helpful is a different issue than whether or not the idea is out there.

My point was that this is the idea that a beginner actually had. The idea probably stems from his experience of getting flat/soft images when he tried shooting RAW.

I think it's helpful to acknowledge that experience, rather then to try to deny its existence.

I think your point was that one shouldn't assume any particular context for processing RAW files, and hence it doesn't make sense to talk about what they "look" like. Depending on the processing, they may result in an image that is flat or vibrant, soft or sharp, etc.

No, that wasn't my point. My point was that until processed, a raw file doesn't have any defined human perceptual meaning.

There is a difference between "any defined meaning" and a "well defined meaning".

I qualified it advisedly, because obviously, since you can use a raw file to create a file with a well defined perceptual meaning, there is some kind of link. But the point is, the numbers in the raw file don't define the 'look' of the image without a set of assumptions the processor has to make, which are not in any way 'default'.

Again, our difference boils down to whether or not the software used by a beginner has default settings.

No, it depends on whether you think investing a characteristic of the software to the file helps anyone understand what is going on. I suspect it just 'helps' them get as confused as you seem to be.

Let's try to stay away from personal commentary.

Even without a specific context a raw file will have some associated meaning. Given the embedded metadata, there are areas of the image which will be green, blue, red, etc. Some areas will be darker, and others will be lighter.

What we don't have is an "exact" meaning. Without knowing the particulars of the raw processing, we don't know exactly which shade of green, blue or red we will get. If the particulars are unusual, we might actually get something that isn't green, blue or red.

That amounts to what I said above, clearly since you can derive a perceptually defined link there is some meaning, which is why I qualified the statement as I did. I get the impression now that you're trying to squeeze through that qualification so as not to be wrong. Sure, you can find a route by which what you said isn't complete garbage, but that's actually denying the intention of it, and also ignoring its misinformation potential, which is why I challenged it.

Let's back up a bit. We got here when I said "I think the point AmazingLight was making was that using the default settings as his processing context, a RAW file can result in a "flat" and/or "soft" image."

I stand by that statement. (Other than the typo it should have said "...in his processing...")

We got here in more ways than that. As it happens I agree that might have been what he was meaning to say, which means that he definitely needs the help he was seeking.

Yes, I think we both agree he would benefit from a better understanding of why he got the results he got. I suspect we even agree on how he got those results.

We seem to disagree on terminology, not the underlying issues. Does it make sense to talk about "colors" in a RAW file? On one hand, it's true that there is no direct and absolute correspondence between the data in the raw file and a particular color. On the other hand, we can certainly make some reasonable predictions of approximately what color we will get, given the data (and meta data) in the RAW file.

 

[briantii]

I am so annoyed shooting in Raw. I don't even know what it does, to be honest. I think it makes editing a lot easier? I just find shooting in Raw is more of an annoyance than a help, because I have to convert to a jpg. anyway, because barely anything is compatible with Raw, I find. So should I just not shoot in Raw?

I COMPLETELY AGREE ...

This is a "BEGINNERS" forum and I feel that the "experienced" here should help promote the "most" images with the emphasis on good "COMPOSITION" instead of the ultimate IQ at this stage of their photography.

Shooting only RAW, (thus requiring later PP), limits the number of images they can process if they have limited time.

Depends entirely on the workflow. RAW adds zero extra processing time in my workflow with Lightroom. I'd import to Lightroom via the same exact process and everything if they were JPEG's. The file sizes are bigger and the files allow more modifications if I choose, but only if I choose to do that. No extra effort though.

 

[Tony Beach]

Isn't the source of corruption nearly always the card?

The disastrous cases we know are caused by pseudo-reliable backup systems, whole collections are gone. That's why I use properly configured ZFS (FreeNAS is very much OK).

Fixable when using Uncompressed files?

Second place goes to the card reader / interface problems (and people do not check copies they made to HDD before erasing and overwriting the cards).

Had that happen, went away when I replaced the bad cards (there was a batch of them).

Bad cards next, and the last place - card interface in the camera.

Same question as above.

I still still think I'm sticking with Lossless Compressed, but this is food for thought. Thanks.

 

[Michael Fryd]

I think your point was that one shouldn't assume any particular context for processing RAW files, and hence it doesn't make sense to talk about what they "look" like. Depending on the processing, they [sic] may result in an image that is flat or vibrant, soft or sharp, etc.

No, that wasn't my point. My point was that until processed, a raw file doesn't have any defined human perceptual meaning.

There is a difference between "any defined meaning" and a "well defined meaning".

I qualified it advisedly, because obviously, since you can use a raw file to create a file with a well defined perceptual meaning, there is some kind of link. But the point is, the numbers in the raw file don't define the 'look' of the image without a set of assumptions the processor has to make, which are not in any way 'default'.

Again, our difference boils down to whether or not the software used by a beginner has default settings.

No, it depends on whether you think investing a characteristic of the software to the file helps anyone understand what is going on.

We get it here all the time at DPR, or we used to anyway, and this site encourages it with their boilerplate approach to converting Raw files from cameras and then offering them up in their studio "comparison." Until recently I was downloading those files to look at them myself and making the mistake of accepting the default sRGB color space, but then I opened a Nikon file using my preferred AdobeRGB color space and what a difference that made as the colors from the sRGB color space are noisier (I would call them muddier) and flatter too.

This is what I got from one of the files:

[IMG width="400px" alt=""Default" sRGB on the top of the patches, AdobeRGB on the bottom of the patches."]https://photos.imageevent.com/tonybeach/mypicturesfolder/sharing/DSC_0532.jpg[/IMG]
"Default" sRGB on the top of the patches, AdobeRGB on the bottom of the patches.

Yes. Depending on how the raw data is processed, you might get a different point in the visible colorspace. On the other hand, with reasonable processing, the greens are still green, the red is still red, etc. There is a good correspondence between the raw data and the approximate area of the visible color space.

Depending on what your default settings are, relying on them might very well produce flat or muddy colors.

 

[Tony Beach]

Depending on what your default settings are, relying on them might very well produce flat or muddy colors.

That makes "default" a rather amorphous target.

 

[Iliah Borg]

It is my understanding that WB preconditioning likely does occur on my D800, but that changing the WB does not change that.

WB pre-conditioning started with D2X, and it is in fact a calibration to avoid using different colour profiles for different cameras. Nikon are applying, right in the camera, to raw data being written, what Adobe call CameraCalibration (1,2) and apply at the raw processing stage - "a calibration matrix that transforms reference camera native space values to individual camera native space values". Thus Nikon are avoiding the disclosure of calibration values.

 

[Tony Beach]

It is my understanding that WB preconditioning likely does occur on my D800, but that changing the WB does not change that.

WB pre-conditioning started with D2X, and it is in fact a calibration to avoid using different colour profiles for different cameras. Nikon are applying, right in the camera, to raw data being written, what Adobe call CameraCalibration (1,2) and apply at the raw processing stage - "a calibration matrix that transforms reference camera native space values to individual camera native space values". Thus Nikon are avoiding the disclosure of calibration values.

Does changing the WB setting on the camera change the WB preconditioning in any way?

 

[Iliah Borg]

Isn't the source of corruption nearly always the card?

The disastrous cases we know are caused by pseudo-reliable backup systems, whole collections are gone. That's why I use properly configured ZFS (FreeNAS is very much OK).

Fixable when using Uncompressed files?

Yes, with manual touch-up.

Second place goes to the card reader / interface problems (and people do not check copies they made to HDD before erasing and overwriting the cards).

Had that happen, went away when I replaced the bad cards (there was a batch of them).

Bad cards next, and the last place - card interface in the camera.

Same question as above.

I still still think I'm sticking with Lossless Compressed, but this is food for thought. Thanks.

I'm not saying one needs to change the ways of setting the camera. I'm saying that one may want to consider a more robust backup policy if using compressed raw.

 

[Tony Beach]

Isn't the source of corruption nearly always the card?

The disastrous cases we know are caused by pseudo-reliable backup systems, whole collections are gone. That's why I use properly configured ZFS (FreeNAS is very much OK).

Fixable when using Uncompressed files?

Yes, with manual touch-up.

Second place goes to the card reader / interface problems (and people do not check copies they made to HDD before erasing and overwriting the cards).

Had that happen, went away when I replaced the bad cards (there was a batch of them).

Bad cards next, and the last place - card interface in the camera.

Same question as above.

I still still think I'm sticking with Lossless Compressed, but this is food for thought. Thanks.

I'm not saying one needs to change the ways of setting the camera. I'm saying that one may want to consider a more robust backup policy if using compressed raw.

Got it. Thanks.

 

[Iliah Borg]

Does changing the WB setting on the camera change the WB preconditioning in any way?

That depends. Adobe recommend to do so, but I never saw Nikon changing those values depending on white balance. However I never payed much attention to that. With other brands, like Canon, the correction factors depend mostly on the sensor temperature - but Canon do not apply those to raw, they have those in metadata. Actually, this whole thing is a huge topic in itself.

 

[Tony Beach]

Does changing the WB setting on the camera change the WB preconditioning in any way?

That depends. Adobe recommend to do so, but I never saw Nikon changing those values depending on white balance.

Then, if I'm reading this right, it might well depend on which Raw converter is being used.

However I never payed much attention to that. With other brands, like Canon, the correction factors depend mostly on the sensor temperature - but Canon do not apply those to raw, they have those in metadata. Actually, this whole thing is a huge topic in itself.

Yes, tangential to the topic at hand, and a proverbial can of worms.

 

[Tony Beach]

I am so annoyed shooting in Raw. [Snip.] So should I just not shoot in Raw?

...just shoot JPEG for now.

Seems this was the answer the OP was looking for.

 

[Mark_A]

Isn't the source of corruption nearly always the card?

The disastrous cases we know are caused by pseudo-reliable backup systems, whole collections are gone. That's why I use properly configured ZFS (FreeNAS is very much OK).

Fixable when using Uncompressed files?

Yes, with manual touch-up.

Second place goes to the card reader / interface problems (and people do not check copies they made to HDD before erasing and overwriting the cards).

Had that happen, went away when I replaced the bad cards (there was a batch of them).

Bad cards next, and the last place - card interface in the camera.

Same question as above.

I still still think I'm sticking with Lossless Compressed, but this is food for thought. Thanks.

I'm not saying one needs to change the ways of setting the camera. I'm saying that one may want to consider a more robust backup policy if using compressed raw

I think I will start experimenting with compressed 12 bit for the time being. Neither my memory cards nor my computer are I think up to 70mb raw files yet.

It would be nice to be able to shoot raw+jpeg and have the best of both worlds and another D800 shooter on here has said they shoot this way and 9.9/10 they end up just using the jpeg and not touching the raw file.

From my experience so far the jpegs out of the D800 are pretty good and with the blinkies, which I hadn't had before, it is easy to see if I have actually blown the highlights.

Mark_A

Thread for Sunrise & Sunset pictures (part 3!)

https://www.dpreview.com/forums/post/60273364