The modern trend in retouching is non-destructiveness and reversibility of every step that you take. With years of development, Photoshop introduces more and more elaborate tools supposed to help you retouch in such a way, that you don’t make any irreversible changes to images.
Imagine, for example, that you work for a client and you retouch a beautiful portrait for them, and it takes quite a lot of time. When it’s done, you flatten the image and save it as a jpeg with low quality to reduce the size in megabytes, then you send your work to the client. And – whoops – it’s not a huge success, you are asked to make some changes. Then you realize that it’s just not possible: as you haven’t saved any backup images in the process, and you can’t really work with the final image as it’s a jpeg with reduced quality, and every change brings pixel artifacts and just ruins everything. You have to start all over again!
Or imagine that you saved the image with its size in pixels reduced, you made it, for example, 1000 by 1000 pixels, and then the client tells you it must be 2000 by 2000. What can you do? Nothing, because if you try to enlarge the image, it will get all blurry. You have to start all over again!
In modern versions of Photoshop, you have all the tools and features needed to prevent yourself from messing up like that. You can use layers, adjustment layers, smart objects and many other things to be able to change anything in the image without having to start it all over again if you need to change something. You can save images uncompressed while keeping all the layers and adjustments. And there are a lot of tutorials that teach you how to organize your retouching process, how to organize everything in the image like layers and their masks, groups of layers and so on. If the client asks you to change anything – from color to shape, you can do it easily without having to redo any of the other steps.
What I'm trying to point out, is that
Why is that? There are, in fact, many reasons why. In hi-end retouching, you don’t care about file size, you can easily spend hours or even days working with a single image, and then you will most probably have to make some changes because you don’t do telepathy and clients can be a bit vague about what they want. Then your work is going to be published somewhere in high resolution or even printed, and the image quality should be high enough. Your priority number one is quality, everything else is not that important.
In catalogue retouching, it’s just the opposite. You are not going to spend days and even hours working with one image, and you will be handling many of them, sometimes hundreds in a single day. You won’t have the luxury to take your time as you’ll have to retouch in a fast manner. Dealing with batches of images, you’ll want them to be compact and not weigh hundreds of megabytes each, otherwise, your computer will have a hard time processing them. The images are going to be consistent with each other, and normally you won’t have to make many changes afterward except correcting a few mistakes. Then the images will be uploaded to some website and later, when the goods they represent are sold out, they will be removed and gone forever. Your main priorities are speed and consistency, and quality is not that important, at least not to the extent when a few blown out pixels in highlights is a tragedy.
All these differences make product image retouching a special process. Let’s see how it affects our approach to image handling. As you will have to download, upload, and handle many images, their file size is an important issue. One of the companies I worked for had to process up to 200.000 items in 3 to 4 months, and the studio was supposed to process about 1500 items daily. Every item was represented with 4 to 5 images, making it millions of images stored on servers and processed every year. To reduce file size, photographers didn’t shoot raw images as jpeg quality was sufficient for the technical demands. You cannot imagine photographers shooting jpegs for ads and fashion lookbooks, but in product image retouching, it’s very common.
When you have millions of images on a server, you don’t save back up images, you don’t use PSD or TIFF file formats, and in the end, you have just three versions of every image. You have the original file in case you have to redo it from scratch, which doesn’t happen often, but still happens. Then you have the final image saved with maximum quality, and you can use it if you need to make minor changes like removing a spot or some dust in case you've missed it. Then you have the image saved with reduced quality to reduce the file size, that’s what gets uploaded on the website.
Reduced quality images can weigh 3 to 5 times less than the maximum quality images of the same size in pixels, which is great, but if we don’t save them with maximum quality, too, every time you want to make a change, no matter how minor it is, you need to retouch once again using the original high-resolution image, otherwise, it will be all blurry and pixelated.
If we had to correct many images after finishing them, it would be wiser to have some PSD backup, to use adjustment layers and so on. But normally it’s not more than 2% of images that get altered after they are finished, and all the alterations are usually quite subtle. In this case, it’s not worth to spend terabytes of disk space on backup, as you can redo any image from scratch in mere minutes. And if there are some particularly nasty images, you're always free to save a hi-res or two just in case.
Because of all that, we usually work with 8-bit images, keeping them as flat and simple as possible. This approach inevitably leads us to the following fact: product image retouching is not about being non-destructive and it doesn't care much about reversibility. Any change you do is permanent, and you only have a limited possibility to undo what you did. So when you work, you have to be precise.
You also have to work fast. By fast, I mean that you must be able to retouch a typical image in under 5 minutes, sometimes under 3 minutes, and I've managed to organize retouching teams that spent less than 1 minute per image. It is only possible if you optimize the process and don’t do things that can be automated manually. The fact that the images have to be consistent also forces you to do the same things on all the images from the same batch, and it’s only reasonable to automate these things.
The deal is, you can't be fast if you have to work with bulky layered images. Even creating layers takes time, not to mention the time you spend on moving between them and twiddling with them and processing heavy PSD images because of them. But if I just say: “Ok, guys. No layers unless it's completely necessary, keep it flat. Remove the adjustment layer Properties tab from your workspace, you don't need it” – you might think I am some kind of a lame person without understanding how cool it is to be able to do things in a non-destructive and totally reversible way. Because that's what they try to make you think when you read and watch tutorials: that professionals use all that complicated stuff, and that's exactly what makes them professionals. Even the word “non-destructive” implies something good. Why would you want to be destructive anyway?
And that's exactly why people can't imagine how someone is able to retouch hundreds of images in a single working day. Not with this complicated workflow where you're supposed to organize layers into groups and name them (otherwise you'll forget what is what). When there are new non-destructive features in every new version of Photoshop, isn't it stupid not to use them? As a matter of fact, no. And I'll prove that. Just let this idea settle in your mind, don't throw it away as rubbish.
Now to summarize all that I’ve said.