As Shot vs Auto

[kami]

I sometimes use AUTO just to see what LR3 does to the photo. Then I adjust the photo the way I like it. You don't need us to tell you if you like the photo or not. Why are all these threads trickling into the "Beyond the Basics" section? This should be in the beginners section.

 

[EchoingWhisper]

I sometimes use AUTO just to see what LR3 does to the photo. Then I adjust the photo the way I like it. You don't need us to tell you if you like the photo or not. Why are all these threads trickling into the "Beyond the Basics" section? This should be in the beginners section.

I thought this was Beyond the Basics? Is the camera algorithm or lightroom algorithm better?

 

[kami]

Well, you always want to start out with a proper white balance in-camera. If you're clueless as to which WB to use on the go I would suggest using AUTO WB and adjust the WB in lightroom if you need to. Lightroom has probably the easiest of WB controls on the planet if you know how to use mouse 1 button and can move a slider.

And not being sure about which WB setting to use, whether Auto or As shot in LR3 is not a "beyond the basics" material IMHO. Maybe asking about gelling of off camera flashes to balance out color temperatures emitted by the available lighting source would fall into this category.

 

[clanthar]

I sometimes use AUTO just to see what LR3 does to the photo. Then I adjust the photo the way I like it. You don't need us to tell you if you like the photo or not. Why are all these threads trickling into the "Beyond the Basics" section? This should be in the beginners section.

I thought this was Beyond the Basics? Is the camera algorithm or lightroom algorithm better?

Alright. There's the question, and it's a legit, good question. In part it comes from your observation of a discrepancy between the two algorithms. Start there. If the two software programs are returning different results then there are three possibilities.

1. The camera is right and LR is wrong.
2. LR is right and the camera is wrong.
3. They're both wrong.

I assure you the answer is #3. Neither algorithm is better and both algorithms can't do the job. Auto white balance doesn't work. Software can't see.

Earlier in the thread you expressed frustration about white balance. The first step toward beating that is to shift your expectations. You're looking for an "auto" solution in the technology and there is none. Frustration in this case is understandable. You've been misled to believe there's possible value in those auto algorithms when in fact there is none. In fact using them typically damages your photo.

Software tools can help, but the job is yours to do and, unlike software, you can see.

Best solution: Shoot RAW files and white balance manually. Using software to set a WB point in a photo can work very well. Good software will do a good job if you can hand it a known reference. I've been walking around my neighborhood lately taking photos. In nearly all my street scenes in the residential area someone has a TV dish on their house -- it's as good as a grey card. When there's no WB reference in the photo, the bottom line is: It's up too you. With practice you'll continue to improve. Software can't see -- you can.

Joe

 

[EchoingWhisper]

I sometimes use AUTO just to see what LR3 does to the photo. Then I adjust the photo the way I like it. You don't need us to tell you if you like the photo or not. Why are all these threads trickling into the "Beyond the Basics" section? This should be in the beginners section.

I thought this was Beyond the Basics? Is the camera algorithm or lightroom algorithm better?

Alright. There's the question, and it's a legit, good question. In part it comes from your observation of a discrepancy between the two algorithms. Start there. If the two software programs are returning different results then there are three possibilities.

1. The camera is right and LR is wrong.
2. LR is right and the camera is wrong.
3. They're both wrong.

I assure you the answer is #3. Neither algorithm is better and both algorithms can't do the job. Auto white balance doesn't work. Software can't see.

Earlier in the thread you expressed frustration about white balance. The first step toward beating that is to shift your expectations. You're looking for an "auto" solution in the technology and there is none. Frustration in this case is understandable. You've been misled to believe there's possible value in those auto algorithms when in fact there is none. In fact using them typically damages your photo.

Software tools can help, but the job is yours to do and, unlike software, you can see.

Best solution: Shoot RAW files and white balance manually. Using software to set a WB point in a photo can work very well. Good software will do a good job if you can hand it a known reference. I've been walking around my neighborhood lately taking photos. In nearly all my street scenes in the residential area someone has a TV dish on their house -- it's as good as a grey card. When there's no WB reference in the photo, the bottom line is: It's up too you. With practice you'll continue to improve. Software can't see -- you can.

Joe

Thanks. I read from somewhere saying white balance is the first step and most important step in post processing, and having a correct one will maximize dynamic range and contrast, so I wanted to make sure what I am doing is correct.

 

[EchoingWhisper]

I sometimes use AUTO just to see what LR3 does to the photo. Then I adjust the photo the way I like it. You don't need us to tell you if you like the photo or not. Why are all these threads trickling into the "Beyond the Basics" section? This should be in the beginners section.

I thought this was Beyond the Basics? Is the camera algorithm or lightroom algorithm better?

Alright. There's the question, and it's a legit, good question. In part it comes from your observation of a discrepancy between the two algorithms. Start there. If the two software programs are returning different results then there are three possibilities.

1. The camera is right and LR is wrong.
2. LR is right and the camera is wrong.
3. They're both wrong.

I assure you the answer is #3. Neither algorithm is better and both algorithms can't do the job. Auto white balance doesn't work. Software can't see.

Joe

Another question, knowing that they are both wrong, which one should be used to achieve more accurate white balance before my own fine-tunings?

 

[unpopular]

Are you people seriously so nit picky that you can't provide an answer to a question about white balance because the horizon is off?

GIVE US SOMETHING MAN!

If Echoing wasn't 100% sure what white balance was, only that in one mode it will look better in some circumstances, but in another circumstance it will look better in a different mode, how can you expect him to know which image to post?

Honestly, I think a night photo is the perfect example to kind of explain exactly what white balance is.

 

[clanthar]

Another question, knowing that they are both wrong, which one should be used to achieve more accurate white balance before my own fine-tunings?

If your camera provides the option to set a custom WB on site and you have the time and means to do that then do that and use that as a starting place. You're talking about LR so I assume you're shooting RAW. A RAW file has no set WB. The converter will allow you to examine the "as shot" WB or a number of presets. The converter will also have a WB point set tool that you can use if the photo has a known WB reference -- do that.

I ignore the WB setting in my camera. I open my RAW files and begin by looking for that WB point reference in the photo. If the photo has a good WB reference I use it and start from there. If the photo has no WB reference then I head straight for the color temp slider and start there.

Joe

 

[unpopular]

Getting the white balance close to accurate in camera is useful if only to render the histogram correctly. I have had on rare occasion my hilight warning giving a false alarm because white point was off.

Though this could be a limitation in my camera, also.

 

[Garbz]

I thought this was Beyond the Basics? Is the camera algorithm or lightroom algorithm better?

Part of being beyond the basics is understanding that photography is an art. You want to know which one is better? Define to me what you mean by better! Then tell me which one of the poor examples more closely matches your ideal "better" result.

Thanks. I read from somewhere saying white balance is the first step and most important step in post processing, and having a correct one will maximize dynamic range and contrast, so I wanted to make sure what I am doing is correct.

White balance has little to do with dynamic range and contrast. A picture does not change tone when you slide the white balance slider (unless you max out one of the channels in the process which itself isn't too much of a problem either). The reason is that if you slide the white balance to the warmer side, green and red increase, and blue proportionally decreases. It doesn't not change the overall balance of a black and white histogram.

Another question, knowing that they are both wrong, which one should be used to achieve more accurate white balance before my own fine-tunings?

This is a trial an error scenario. Different algorithms excel with different images. Just because the camera more accurately renders an overcast day doesn't mean that lightroom may not more accurately render a sunny day etc. Given the wide range of possible white balances the only way to be really sure is to have something grey to work with. But ultimately the goal is to pick what LOOKS best, and not what some mathematical formula tells you to pick. So grab that white balance slider and give it a tweak.

Getting the white balance close to accurate in camera is useful if only to render the histogram correctly. I have had on rare occasion my hilight warning giving a false alarm because white point was off.

Though this could be a limitation in my camera, also.

This is a limitation of the colour gamut, not the camera. Unless you're exceeding the possible visible colour space in which case yes it's a limitation of the camera . Basically there are certain colours that you can try to record which will always result in an out of gamut warning on the histogram regardless of what you try and do. Some colours are just that pure that they will saturate the channel and there's simply nothing you can do about it. Try photographing a red LED for instance without clipping the red channel. If you try and work the image so it's not clipping you'll find you've either darkened it to the point of absurdity or artificially desaturated it making the image no longer a representation of the pure red that an LED is capable of emitting.

 

[EchoingWhisper]

I thought this was Beyond the Basics? Is the camera algorithm or lightroom algorithm better?

Part of being beyond the basics is understanding that photography is an art. You want to know which one is better? Define to me what you mean by better! Then tell me which one of the poor examples more closely matches your ideal "better" result.

Thanks. I read from somewhere saying white balance is the first step and most important step in post processing, and having a correct one will maximize dynamic range and contrast, so I wanted to make sure what I am doing is correct.

White balance has little to do with dynamic range and contrast. A picture does not change tone when you slide the white balance slider (unless you max out one of the channels in the process which itself isn't too much of a problem either). The reason is that if you slide the white balance to the warmer side, green and red increase, and blue proportionally decreases. It doesn't not change the overall balance of a black and white histogram.

Another question, knowing that they are both wrong, which one should be used to achieve more accurate white balance before my own fine-tunings?

This is a trial an error scenario. Different algorithms excel with different images. Just because the camera more accurately renders an overcast day doesn't mean that lightroom may not more accurately render a sunny day etc. Given the wide range of possible white balances the only way to be really sure is to have something grey to work with. But ultimately the goal is to pick what LOOKS best, and not what some mathematical formula tells you to pick. So grab that white balance slider and give it a tweak.

Getting the white balance close to accurate in camera is useful if only to render the histogram correctly. I have had on rare occasion my hilight warning giving a false alarm because white point was off.

Though this could be a limitation in my camera, also.

This is a limitation of the colour gamut, not the camera. Unless you're exceeding the possible visible colour space in which case yes it's a limitation of the camera . Basically there are certain colours that you can try to record which will always result in an out of gamut warning on the histogram regardless of what you try and do. Some colours are just that pure that they will saturate the channel and there's simply nothing you can do about it. Try photographing a red LED for instance without clipping the red channel. If you try and work the image so it's not clipping you'll find you've either darkened it to the point of absurdity or artificially desaturated it making the image no longer a representation of the pure red that an LED is capable of emitting.

Digital Photo Editing Workflow

 

[unpopular]

Getting the white balance close to accurate in camera is useful if only to render the histogram correctly. I have had on rare occasion my hilight warning giving a false alarm because white point was off.

Though this could be a limitation in my camera, also.

This is a limitation of the colour gamut, not the camera. Unless you're exceeding the possible visible colour space in which case yes it's a limitation of the camera . Basically there are certain colours that you can try to record which will always result in an out of gamut warning on the histogram regardless of what you try and do. Some colours are just that pure that they will saturate the channel and there's simply nothing you can do about it. Try photographing a red LED for instance without clipping the red channel. If you try and work the image so it's not clipping you'll find you've either darkened it to the point of absurdity or artificially desaturated it making the image no longer a representation of the pure red that an LED is capable of emitting.

That does make sense, but that does not really change the fact that setting proper white balance affects the histogram readout in the field. Obviously if I set the white too high and then take a photo of green grass, it might appear clipped when in fact it is not. How I approach this image would be different if the grass truly were out of gamut.

I suspect that when exposing to the right this is more prone to affect shooting. I also wonder if a uni-wb workflow might prevent it, though my camera can't "get to" a uni wb afaik.

 

[KmH]

Get one of these to carry around with you:

GENUINE WhiBal G7 Certified Neutral White Balance Card - Pocket Kit (2.1"x3.3")

and/or one of these: X-Rite ColorChecker Passport Software

 

[Garbz]

That does make sense, but that does not really change the fact that setting proper white balance affects the histogram readout in the field. Obviously if I set the white too high and then take a photo of green grass, it might appear clipped when in fact it is not. How I approach this image would be different if the grass truly were out of gamut.

I suspect that when exposing to the right this is more prone to affect shooting. I also wonder if a uni-wb workflow might prevent it, though my camera can't "get to" a uni wb afaik.

Ahh but the grass truly is out of gamut if you photograph it in such a way. The histogram in the field is based on the JPEG preview the camera captures. If you shoot RAW the histogram can't be trusted, not the least because your histogram will look different once you open the file on the computer. If you don't shoot RAW then maybe you should start.

But that is still arguing semantics because what I am trying to say is that clipping one channel in the histogram isn't bad. It's quite different from clipping ALL channels in the histogram.

 

[unpopular]

^^ I am a bit confused by this. Yes, the channel is clipped in the JPEG preview but only because the preview was adjusted from the raw information to meet a determined white balance. If you adjust the white balance away from the clipped channel's color the JPEG preview is no longer clipped in those channels. At least in my camera this is how it works. So, no, the channel is only clipped according to that preview - but if there is actually information there depends on how the image is processed, provided that the color does not exceed the native gamut of the camera's sensor itself.

If the same white balance is applied to both the JPEG preview and the output file, then both histograms will be similar. The histogram can be trusted, but only if white balance is properly set to something that resembles the light available.

I would HIGHLY disagree that clipping in one channel is not "bad". Clipping is always bad, period, and should always be avoided. Yes, you can rebuild one channel from the information from other channels and other such channel chop trick. However, this scenario will always result in lost detail.

Provided that the color itself is within gamut, it is much better to decrease exposure than to allow one channel to blow out. I am also playing with the idea of optically filtering that color and color correcting in post.