The RAW/JPEG Discussion

[ajohnw]

Hope you don't mind Colin. Maybe 15min or so including flattening the background to black and as it was done by selection softening the lady a little. Too much really. Shadow was burnt using brightness, saturation and tone map. Latter didn't really do anything despite 5mins of it. Slight unsharp mask to whole image, hence soften. Cropped to how I might crop a shot like that.

Only thing is that the hair doesn't really show up until it's enlarged but realistic shadow ???



-

[pnodrog]

Hi Colin, well spotted and a bit clumsy of me - I do know the difference. However the improvements tend to go hand in hand and both assist the recovery. No photographic advantage in having ultra high resolution A/D on a minimal dynamic range and visa versa.

[ajohnw]

It also possible to change the modelling effects on jpg's. Can take a while to decide on the best one. A moderate change. Clarification and skin glow. A bit too much in places so really needs more work.



-

[pnodrog]

Hi Colin,

Thanks for that. Quite honestly I find these discussions entertaining, humerous and informative. We should never give up on the fact that whilst we might defend our chosen format, there's always a different way which, one day, might benefit us. Jpegs still rule, of course, as we know raw users still can't get it right 'in-camera' ...

Hi Steve

strangely enough this is the basis for a very good argument that beginners should use RAW

[ajohnw]

Hi Colin,

Thanks for that. Quite honestly I find these discussions entertaining, humerous and informative. We should never give up on the fact that whilst we might defend our chosen format, there's always a different way which, one day, might benefit us. Jpegs still rule, of course, as we know raw users still can't get it right 'in-camera' ...

They cause me endless amusement. Camera forums are interesting places but people tend to lock into a way of doing things and there is often no point in showing that they may not be correct. As I stated I shoot raw plus jpg and find that the jpg more often than not does it's job but some cameras are likely to need more subsequent correction than others. I keep the raw and the original jpg so basically if I ever do need either I have it.

One thing that strikes me about this thread is that it should be discussing camera jpg's not straight raw conversion with arbitrary black points even if they happen to be correct. The issue really is that both approaches often need subsequent adjustment. Some like to do it one way other like another. Some find one way easier than the other. Most of my processing has been from jpg so I have a fair idea what to expect from them once I know the camera. Doesn't stop me from using raw but I can not see any point in switching to it completely. I don't think any of the argument presented here will change anybodies thoughts on the subject either ( I've tried before on one aspect ). What people tend to forget is that the camera has a raw converter built into it. They don't just chop resolution off willy nilly they do what people do when they do it themselves with but with some loss of flexibility. How much loss depends on the shot and how clever the manufacturers are.

Anyway each to his own and the topic will probably come up again. It isn't exactly uncommon.

-

[Colin Southern]

Doesn't stop me from using raw but I can not see any point in switching to it completely.

It's probably a glass 1/2 full -v- 1/2 empty kind of thing. I'd take the approach "I can not see any point in NOT switching to it (ie RAW) completely". I just can't see a single down-side to a completely RAW workflow, but several advantages.

I'm just wondering how much of the traditional arguments against it are based on "historical rollover" ie "It takes a lot of time and horsepower to convert the RAWs to JPEG" when folks don't realise that this step has become trivial these days.

[Glenn NK]

It's probably a glass 1/2 full -v- 1/2 empty kind of thing. I'd take the approach "I can not see any point in NOT switching to it (ie RAW) completely". I just can't see a single down-side to a completely RAW workflow, but several advantages.

I'm just wondering how much of the traditional arguments against it are based on "historical rollover" ie "It takes a lot of time and horsepower to convert the RAWs to JPEG" when folks don't realise that this step has become trivial these days.

I really can't say why, but the word inertia popped into my thoughts more than once.

OK, I'll admit it, it's already been postulated.

[Letrow]

It's probably a glass 1/2 full -v- 1/2 empty kind of thing. I'd take the approach "I can not see any point in NOT switching to it (ie RAW) completely". I just can't see a single down-side to a completely RAW workflow, but several advantages.

I'm just wondering how much of the traditional arguments against it are based on "historical rollover" ie "It takes a lot of time and horsepower to convert the RAWs to JPEG" when folks don't realise that this step has become trivial these days.

It might be something for a poll Colin

For me this thread still sounds a lot like an argument that has no clear-cut answer. Apparently, in a lot of cases, people can't really see the difference between the two formats, after they have been through PP. Your shadow detail might be an exception of course, that would favour the use of RAW in some cases. Even there though, I would understand the test better after I had seen it done with JPEG and RAW coming straight out of the camera.

Do we need a clear winner? Frankly, I think it is not important. It sounds a lot like the two sides are using quality as an argument, i.e. you should always go for the product that captures/has most.
I have heard this about photography, I have heard it about music, natural versus synthetic materials, horsepower versus torque, quality wine versus cheap wine, brand names versus non-brand, etc.
In most cases it would make sense to look for what you really need, and RAW and JPEG both have their separate uses.

[revi]

I can see one reason to shoot RAW+JPEG, and that's when you need some of the images NOW (as in: just after the event, with no chance to get to your software)...

[ajohnw]

For one thing shooting raw plus jpg allows images to be scanned with what ever viewer is on a machine quickly - ever so. Can't say I envy windows users on that point though, It's an awful viewer. We do have windows on a couple of lap tops. My use of them is away from home and tethered shooting.

Scanning them quickly allows crap shots to be disregarded immediately and immediate assessment made as to whether it would be best to work on the raw file and just how much work would be needed on the jpg. The jpg can also guide what is likely to be the best option for processing the raw file. On the shot posted in this thread the camera is very likely to produce a perfect jpg needing no adjustment at all. The only question on that score is precisely what compression level it uses and how much information loss that entails. On the other hand that also relates to the size the shot is eventually displayed at. These days the data loss is relatively small and nothing remotely like compacts - who suffers from inertia?

Fact - raw done properly takes longer. Grumpydriver summed it all up nicely at the start of the thread complete with images. I've added more based on some one else's image. Actually I feel these are slightly over exposed even the shot from raw. More so really see the slight white area on the right cheek. 5 bits margin. The jpg has more so for some reason turned out better. People who mention inertia really have used it the wrong way round. Often views like that suggest that they know the argument is lost anyway. I wonder if they have used cameras that could produce tif's. Raw files are down to reducing storage sizes compared with those.

But as I said each to his own. There are many similar arguments. Some shoot manual. Me? Well the camera has a number of options so I use all of them and find out what they do and pick which ones are likely to be best for what. I'm a big fan of P mode on a Pen for instance. Camera sets aperture and speed and I turn the wheel to alter the relationship if needed. That's not compensated shooting. Faster speed gives a faster aperture etc. Some say use aperture or speed priority but really there is no difference. and so it goes on.

-

[MrB]

The notion of Raw as an insurance policy is appealing. As I am getting used to knowing what the image is like from its appearance on the camera's LCD, together with the histogram and any flashing over/under exposed bits, I can record highest quality Jpegs in the camera most of the time, knowing that they can take a few tweaks in software later, and that the quality will still be very good (even if the photo is c**p!) up to the biggest I can print - A3. However, if there is a difficult shot that I am not so sure about after viewing the image on the camera's screen, all its original data from the buffer can then be recorded on the memory card as a separate Raw file, as the insurance policy for that shot.

Philip

[davidedric]

A quick comment from a sometimes amateur wildlife photographer, and a fairly novice PP'er.

For the RAW: Things happen fast, and light changes. If I were in a studio or viewing a landscape then I could hope to get the camera settings spot on, but in the field I would feel I was too likely to miss the shot if I were trying to get everything exactly right (of course, a better photographer than I might be a lot more proficient. Hence the insurance policy - it's I won't be back there again, and I certainly won't see the same thing.

For the JPEG. I have a fairly low level camera, a 600D. The buffer is none too big and none too fast. If I shoot in RAW I don't want to chance continuous shooting, because you can bet your life that the buffer will fill at the critical moment.

Just back from a trip, and I shot in RAW plus JPEG (on the basis that the JPEG adds little to the overall file size,and I might as well have it to give me an easy heads up of what I should be minimally aiming for), and for the most part just used single shot. I'll get round to posting an image or two soon, and I'd be pleased if you could let me know what you think.

[Manfred M]

Colin, I think we are into a bit of an apples and oranges discussion here. Using jpegs that come out of the camera versus creating a jpeg in PP are two different issues. I find that jgegs straight out the camera usually have fewer artefacts than ones created in PP. Camera manufacturers have done a lot of work in their hardware processing engines to produce optimal results.

The reason I advocate using jpegs, especially for use on the internet is purely related to overall efficiency. I find that while I can achieve equally good results if I use a RAW file. I can crank out an image from a jpeg that is ready for uploading to a normal website in under a minute, but to achieve a similar quality result using a RAW file, will usually take me at least 10 minutes. I know could use automation using actions to create the output, but I find I get inferior results. As I said before, the moment start serious manipulations, to get an image up to the quality level I am looking for, the RAW file is a consideration. I’m not a pro, and I don’t use Photoshop every day; I’ve used it since the release of CS, about 9 years ago and have been shooting RAW files pretty well for close to four years now, pretty well ever since I bought my D90, so while I don't look at myself as anywhere close to an expert, I am quite comfortable with the tool.

Let me try to explain my view with a bit of an analogy. One of my colleagues at work drives two vehicles, a Dodge RAM 3500 truck and a smaller Toyota Lexus. This Dodge truck may be familiar to the North American members, but perhaps not to those from other parts of the world. It is an absolutely huge vehicle with a 6.7 litre Cummins diesel engine. It has a large bed for hauling things around and has a very large towing capacity. It is not most manoeuvrable vehicle I’ve ever been in…

I have very little sympathy when my colleague complains about the high cost of refuelling the vehicle and the cost of the parts and service when the truck needs repairing or even finding a parking spot large enough to park the truck. This is a very good analogy for RAW; it can do everything, but that comes at a price. The Lexus is more like a jpeg, It is good enough for many things, is much more fuel efficient and nimble to drive, but you don’t take it down to the local building centre and use it to haul large pieces of plywood or haul a large trailer.

I have developed a workflow that works for me; so I pretty well always shoot high quality jpeg + RAW, and do know that it does cost me in terms of requiring additional in-camera and off-camera storage. I understand its advantages and its limitations and find that I have developed a workflow that gives me good results, depending on what the final use of the image will be. My tool chest in my workshop is fairly diverse. I don’t use just one tool for a job. I look at PP the same way.

I would like to leave one more thought to the RAW only shooters. When the only tool you use is a hammer, everything starts to look like a nail.

[ajohnw]

jpg doubters might want to consider these. This one is a jpg's dynamic range 5D III plus others with all curve modifiers turned off. It's curious fact that a cameras dynamic range is specified based on jpg's. I would love to see a raw one. Turning on options increases the spread of the curves - more dynamic range but only the linear part will be reproduced as we see it. The curved parts will effectively lack "contrast" and tonal range.



The vertical scale ranges form 0 to 255. What a manual raw conversion can do is give the curve any shape. eg Linear in the top right corner to give best high light reproduction. There is usually head room in the raw file over what manufacturers choose to put into the jpg as well Only problem is that the final image can only have 8 bits so including any of the extra or changing the curve to make some levels more realistic must be at the expense of other light levels in the shot if they are there. The extra head room can be used for correcting exposure either under or over by amounts that will vary according to the camera. Using it often really does just that and nothing else. The information is compressed in a jpg but it's perfectly possible to use various contrast type adjustments to make the detail in the compressed areas more apparent, For the same dynamic range the same thing would have to be done in the manual raw to jpg conversion.

There is some mileage in the resolution argument but it's a bit pointless for web shots. Original image might be 3ft by 4ft on a PC screen so they are invariably reduced dramatically. Even so the data loss depends on what compression levels the camera manufacturer chooses to allow in their jpg's. On a pen for instance there is this loads of detail. If you click on this up pops a 100% full frame view of a pen's jpg. http://www.23hq.com/ajohnw/photo/8055858/original There are other versions of that about but that one has the correct colouring.

Tests show things like this



What these are really showing is different levels of sharpening and little else. Again there is info in the jpg and it can be bought out but for web use there isn't much point. There is also a possibility that the degree of sharpening used might not suit the shot. All cameras I am aware of have different sharpening settings anyway.

I should add that the same sort of comments relate to prints. Publishers can manage 48bit colour. A typical home user can't.

-

[Colin Southern]

The reason I advocate using jpegs, especially for use on the internet is purely related to overall efficiency. I find that while I can achieve equally good results if I use a RAW file. I can crank out an image from a jpeg that is ready for uploading to a normal website in under a minute, but to achieve a similar quality result using a RAW file, will usually take me at least 10 minutes. I know could use automation using actions to create the output, but I find I get inferior results.

I'm afraid that that just doesn't seem right at all. Sounds to me like you have certain picture styles set in camera, and simply need to choose (or create) corresponding presets to be instantly applied in ACR. There's just no way it should be taking 10 minutes to achieve the same look; should be pretty much a case of selecting the right preset and being 95% of the way there.

[ajohnw]

I think you are missing both mine and Grumpy's point Collin. Perhaps we feel that it isn't worth going near raw unless we want to spend 10mins on it and it needs it. None of the modified shots I posted based on yours took that long other than an extra 5 min or maybe more spent locally applying tone mapping out of interest to see what it would do slowly applied to an area with hardly anything in it. My post raw conversion would not be like that anyway.

The final post I made used 2 presets of a sort designed to alter shots in specific ways be they jpg or raws on the way to becoming jpg's.

I'm sure that some may have read the posts and thought more about the subject in general and that hopefully will dismiss some common miss conceptions - at some point. Some will never reach that point, Ignore the posts by all means but why? They speak clearly enough.

-

[Colin Southern]

I think you are missing both mine and Grumpy's point Collin.

I get your respective points entirely Johhn ... they just don't appear logical to me I'm afraid.

Perhaps we feel that it isn't worth going near raw unless we want to spend 10mins on it and it needs it.

I think you may have missed my point though, and that is - for all intents and purposes - there doesn't have to be ANY time difference in processing a RAW capture compared to a JPEG capture - but the RAW capture offers potential advantages in other areas - with no down-side.

Ignore the posts by all means but why? They speak clearly enough.

Sorry - you've lost me on that one.

[pnodrog]

OOPS the penny dropped very harshly driving home but I was to late to withdraw my great binary maths totally incorrectly applied before it was so politely and deservingly almost ignored.

[Glenn NK]

I'm sure that some may have read the posts and thought more about the subject in general and that hopefully will dismiss some common miss conceptions - at some point. Some will never reach that point, Ignore the posts by all means but why? They speak clearly enough.

-

What are the misconceptions you allude to?

Glenn

[Manfred M]

Colin – I understand your workflow, and if I am shooting in a studio or consistently lit environment, I could see using a level of automation, but this is not generally how I shoot. I’m not looking for a whole series of images to present to a client or to sell. Most of my images are either posted to websites or are there for my own personal consumption. I am the photographer, retoucher, printer and client all rolled into one. I pick and chose the images that I will work on and save the rest in case I want to do something with them sometime in the future.

I have found two things about Photoshop:

1. ACR does not seem to do as good (by that I mean accurate) a job as View NX2 / Capture NX2 in converting RAW images out of Nikon cameras. I have noticed that several other Nikon shooters on this site have come to the same conclusion. I find this perfectly plausible, Adobe uses a different RAW conversion engine than Nikon does, and it has to be generic, by design. Adobe has to work reasonably well for all brands of cameras, whereas the Nikon software has to work well only with Nikon products. I would suspect that Nikon engineers working with Nikon software designers would get that part right, even if the resulting software is not as slick, or does not have all the bells and whistles as ACR / Photoshop. The best way to go for Nikon shooters who want to go to a level beyond jpeg is to use the Nikon supplied software to make the conversion to TIFF and then edit the files in Photoshop.

2. The jpeg files coming out of the camera are cleaner and more artifact free that what I get out of Photoshop. I noticed this a few year ago when I first started working with RAW files and tried to match the quality of jpeg images. Somehow I was getting images not as clean as straight out of the camera.

I initially thought the same as you Colin, that it did not make sense to me. Later on I was reading an article by one of the senior designers from one of the major HDRI processing software makers. The prevailing view is that one should always use RAW images because of the detail they hold, and he came out and suggested people should use jpegs for the very same reason I gave. The camera makers have invested a lot of time and effort into the jpeg output and it is very, very good.

When I think about it some more; yes a dedicated hardware image processor in a camera should be able to produce very high quality images. Perhaps more so that a generic jpeg conversion tool that you find in a general purpose software solution.

I’m certainly okay with agreeing to disagree on this one. Your experience shows you one thing, my and John’s experience shows something else. Perhaps ACR works better for Canon than for Nikon, as I haven’t heard of the RAW issues from Canon shooters.