PNG

[Pehpsi]

I started exporting my full size pics to PNG today for the first time just to have a play, and they look great compared to JPG's. Some of the JPG's had faint lines through them even at full size/quality, but the PNG's didn't.

So, why have i been using JPG all this time? Is it ok to switch from now on and use PNG?

Also i choose sRGB when exporting from Aperture, but when i 'get info' the file says:

colour space: RGB
profile name: sRGB profile

Is that right?

Hope you can help, Cheers.

[Killakoala]

PNG is excellent and also allows for layers to be kept.

I don't anything about Aperture, but I prefer to keep my images in Adobe RGB as sRGB is for viewing on monitors.

For printing, best not to use sRGB.

But you can convert back to CYMK or aRGB if you are going to print.

[stubbsy]

Historically png was developed as an open source gif replacement because of patent and licencing issues with the gif compression algorithms. So it has similarly functionality to gifs (but NOT animation). As for PNG vs JPG the short answer is that both are fine but jpeg is lossy and png is not. the only problems you'd have switching from jpeg to png would be if you wanted to use software or upload your pic to somewhere that didn't support png. I store all my lossless image versions as dng and my lossy ones as jpeg since at the compression levels I use both jpeg and png formats are adequate and jpeg is more standard.

Edit : Corrected factual errors

[Alex]

I tend to save my master files (those that have been edited but not sharpened) as PSD. Yes, PSDs are huge in size and they are photoshop files, but they keep layers and are lossless. I'd say TIFFs are also good, but you need to make sure that you choose to keep the layers and while doing so the file sizes are normally larger than the corresponding psd.

Alex

[rooboy]

I don't anything about Aperture, but I prefer to keep my images in Adobe RGB as sRGB is for viewing on monitors.

For printing, best not to use sRGB.

But you can convert back to CYMK or aRGB if you are going to print.

Wrong, wrong, technically correct I'm not meaning to start an argument, but I think this advice is a bit misleading, especially in the beginner's section.

I don't anything about Aperture, but I prefer to keep my images in Adobe RGB as sRGB is for viewing on monitors.

Unless you are using a $10k Eizo LCD or a very, very high quality CRT monitor, you can't see the full aRGB colour gamut. Most LCDs can't event display the full sRGB gamut. If you are working in aRGB, you are guessing at the colour being used, as your monitor can only display its nearest colour to the aRGB numerical value.

I agree that sRGB if for viewing on the internet though.

For printing, best not to use sRGB.

This really depends how you are printing. Most printers can't match the full gamut of sRGB, let alone aRGB. There is a great example here, which compares 2 common printer types (high end inkjet and pro Fuji lab). The different printers overlap both colour spaces in several places. The real answer is to know how you want to print, and choose a colour space accordingly. If you're like me and get all your prints done in a lab, you can see there is almost no benefit to working in aRGB (according to that test, only a tiny amount of extra yellows in the highlights). If you are printing at home on an excellent inkjet, aRGB may be the way to go.

Do bear in mind that you won't actually be able to see those extra colours on your screen though (only in the final print). Colours you can't see on a printer that can't print them

But you can convert back to CYMK or aRGB if you are going to print.

Agree with you here.

Pehpsi, your colour space info looks fine. My advice would be that unless you really, really know what you are doing and know that you have a specific need for aRGB, stick with sRGB.

[sirhc55]

Having worked for a number of years in producing pics for printing (litho) I NEVER use sRGB unless they are for the web and then I duplicate and down convert to sRGB.

Converting backwards and forwards between colour spaces is not recommended.

IMO the best workflow is:

Shoot RAW in Adobe RGB (you could also shoot RAW+JPEG). The RAW+JPEG allows for a ”negative” and JPEG for sRGB.

If you are shooting for litho, do all of your PP’ing in RGB and only when satisfied with the result, convert to CYMK.

Prior to todays wonderful colour sync, I would produce artwork on the computer and have the resulting file proofed by the printer I was going to use, preferably on the paper to be used for the job.

I would then adjust the monitor visually to the hard proof to match all pics for the final print run.

[Alex]

Having worked for a number of years in producing pics for printing (litho) I NEVER use sRGB unless they are for the web and then I duplicate and down convert to sRGB.

Converting backwards and forwards between colour spaces is not recommended.

IMO the best workflow is:

Shoot RAW in Adobe RGB (you could also shoot RAW+JPEG). The RAW+JPEG allows for a ”negative” and JPEG for sRGB.

If you are shooting for litho, do all of your PP’ing in RGB and only when satisfied with the result, convert to CYMK.

Prior to todays wonderful colour sync, I would produce artwork on the computer and have the resulting file proofed by the printer I was going to use, preferably on the paper to be used for the job.

I would then adjust the monitor visually to the hard proof to match all pics for the final print run.

Chris,

But if you are shooting raw, I believe it matters not whether you select aRGB in-camera or in raw processing software later on. At least Capture NX allows you to pick the colourspace you want in pping.

Alex

[Matt. K]

Isn't it wonderful? Everyone has their own way of getting to where they want to go!

[sirhc55]

But if you are shooting raw, I believe it matters not whether you select aRGB in-camera or in raw processing software later on. At least Capture NX allows you to pick the colourspace you want in pping.

Alex

A very good point Alex. I would assume that because you can select Adobe RGB OR sRGB in camera there must be a difference in captured colour space. RGB has a greater gamut than sRGB so it beggars the question - how do you add back information when converting sRGB to aRGB. The answer IMO is interpolation which is not good - mind you, I could be totally wrong

[Alex]

But if you are shooting raw, I believe it matters not whether you select aRGB in-camera or in raw processing software later on. At least Capture NX allows you to pick the colourspace you want in pping.

Alex

A very good point Alex. I would assume that because you can select Adobe RGB OR sRGB in camera there must be a difference in captured colour space. RGB has a greater gamut than sRGB so it beggars the question - how do you add back information when converting sRGB to aRGB. The answer IMO is interpolation which is not good - mind you, I could be totally wrong

Thanks Chris. You are probably right. I guess it boils down to whether RAW captures data in the wider gamut by default and then RAW conversion locks the colour space or whether if you shoot sRGB then you interpolate in the raw converter, i.e. leading to worse quality.

Alex

[Pehpsi]

Thanks for all the info guys. Calibration and colour spaces confuse me the most in photography!

Someone mentioned that PNG is lossy, but up on Wiki it says:

"PNG (Portable Network Graphics) is a bitmapped image format that employs lossless data compression"

I don't do any printing, and am mainly trying to get the best quality for upload to RedBubble.com. They say JPG or PNG are preferred. I thought PNG was better because it won't compress?

[Killakoala]

http://www.redbubble.com/forums/2/topics/1

It does say it prefers JPG although they accept both JPG and PNG. T-Shirt designs must be in PNG.

Other than that i will stick to what i know. This isn't it.

[stubbsy]

... on Wiki it says:

"PNG (Portable Network Graphics) is a bitmapped image format that employs lossless data compression"

I stand corrected - always thought it was lossy. I've just edited my post to make it correct. I'll now slink off into the corner.

[Matt. K]

Chris

I would assume that because you can select Adobe RGB OR sRGB in camera there must be a difference in captured colour space.

RAW image is always captured as aRGB, the option to preset simply being the same as option of WB or sharpening etc. In my experiance it is near impossible to pick the colour space of a printed image up to about 50cm x 40cm provided care was taken with the exposure and post processing. For those who are not having their work printed commercially I recommend they capture, process and print using sRGB colour space for consistency of results. If anyone can demonstrate to me, with prints, that aRGB offers signficant advantages then I would be very interested in seeing those results.

[Steffen]

I would assume that because you can select Adobe RGB OR sRGB in camera there must be a difference in captured colour space.

Thanks Chris. You are probably right. I guess it boils down to whether RAW captures data in the wider gamut by default and then RAW conversion locks the colour space or whether if you shoot sRGB then you interpolate in the raw converter, i.e. leading to worse quality.

The gong goes to Alex in this case, I'm afraid

The RAW image is independent of the selected camera colour space, and has a very different (and wider) gamut than either sRGB or aRGB. The embedded JPEG thumbnail is adjusted for the selected colour space - I believe (don't quote me on that).

On (some?) Nikons (maybe others?), the RAW image isn't totally independent of the selected white balance, though, since that value is also used to tweak the analog gain of the sensor signal amplifiers.

Colour spaces come into play when real-world, device-independent colour (as defined by CIE-XYZ or -LAB) needs to be reproduced by devices like monitors and printers, or fictitious devices (like sRGB or aRGB) if the actual device reproducing the image isn't known in advance.

The biggest leap of faith in the process of RAW conversion is deciding on the set of colours the image should contain, given that

a) we can't know for sure which colours the sensor saw, only piece it together from RAW data plus camera curves and recorded white point
b) we judge the colour by looking at the image through a pretty limited device with its own reproduction characteristics - the monitor.

Hence, the best we can do is get some of the corner stones subjectively right (for example white point, skin tones, etc) from the monitor rendition of the image, and hope for a lot of glorious colour to come out when the image is printed on professional printing equipment (the colour gamut of prosumer ink-jets isn't that much different from that of a good monitor).

As for Aperture, you don't decide on the output colour space until you produce output for a particular device and for a particular purpose. In Aperture (and other RAW converters of the non-destructive kind, like Lightroom and Bibble), you keep the RAW (the "negative") and your list of image adjustments. Exported versions like JPEGs for web or PNGs for t-shirts are merely prints that are made with a specific target audience and a specific set of circumstances in mind (such as, what monitors are your web images typically be viewed on, or what illuminant are your gallery prints going to be viewed under).

If you have to take a detour through external image editors this simplicity is lost, however, and you have to settle for one or another colour space for the image exchanged between the two (or more) programs. This can incur loss of colour information, which one tries to counter by using the widest practical colour space understood by the involved softwares, and as many bits per pixel per channel as possible (at least 16).

Cheers
Steffen.

[Matt. K]

Steffen.
That's what I said.