Workflow of RAW files in Affinity Photo

[SeanZ]

For the photo of 14bits Raw format,  it enters directly to Development persona when I open it and displays RGBA/32(HDR), which can be adjusted for exposure, contrast, saturation, etc.. Then I click the button to enter Photo Persona,  it becomes RGBA/16. I can also make some adjustments in the studios. If I come back to the Photo Persona again, the picture is RGBA/16.

So, I have some questions:
1. Will the photo quality be reduced When entering into Photo Persona from Development persona？
2. Is there any difference between adjusting in Development persona and Photo Persona? 
3. What workflow should I use when processing raw photos？

[James Ritson]

Hi @SeanZ, here's a quick breakdown of the pixel formats and what happens:

As you've observed, when you open a RAW file in Affinity Photo it's processed in a working pixel format of 32-bit float, which is unbounded. This prevents values outside the range of 0-1 from being clipped and discarded, and allows for highlight recovery. Colour operations are processed in ROMM RGB (ProPhoto), which helps colour fidelity even if the intended output colour space is sRGB. You are essentially working in the highest quality that is reasonable.

When you click Develop, by default, the pixel format is converted to 16-bit integer, which is not unbounded. Any values outside the range of 0-1 will be clipped and rounded, so you should ensure you are happy with the tones before you proceed (e.g. using highlight recovery, changing black point etc). The colour space is also converted to whichever option you have chosen—by default, this is sRGB, but you can change this by clicking the Profiles option on the right hand side and choosing another output profile like ROMM RGB or Adobe RGB.

I say by default, because you can change the output format to 32-bit HDR on the develop assistant (the tuxedo/suit icon on the top toolbar). Be very mindful, however, that 32-bit in Affinity Photo is a linear compositing format as opposed to nonlinear. Adjustments will behave differently (typically they will seem more sensitive) and blend modes may produce unexpected results. I would avoid working in 32-bit unless you either want to author HDR/EDR content—this is not the same as HDR merging and tone mapping—or you need the precision for more esoteric genres like deep sky astrophotography. A lot of customers think working in 32-bit is going to offer them the best quality possible. Whilst this is technically true, there are many caveats and I would seriously advise against it unless you have a specific requirement for this format.

To answer your questions:

10 hours ago, SeanZ said:

1. Will the photo quality be reduced When entering into Photo Persona from Development persona？

Technically, the answer is yes, but it's like I mentioned above: unless you have a very specific requirement to work in 32-bit, any loss in quality will be negligible. 16-bit nonlinear precision is more than enough for 99% of image editing use cases. Here's an example of my workflow: I will often choose a wider colour profile like ROMM RGB, then make my image look as flat as possible in the Develop Persona, usually by removing the tone curve and bringing in the highlights slightly if they are clipping. I'll then develop to 16-bit and do all of my tonal adjustments in the main Photo Persona. I have yet to complain about any loss in quality 

10 hours ago, SeanZ said:

2. Is there any difference between adjusting in Development persona and Photo Persona? 

The functionality is more or less the same, but the Develop Persona has more intuitive sliders. In fact, in demos I will frequently explain to people that if you want to do some simple destructive edits to an image, you can simply go into that persona and use a few sliders rather than have to build up a layer stack. One key difference will be the white balance slider: when developing a RAW file, it has access to the initial white balance metadata and the slider is measured in Kelvin. Once an image is developed, however, this slider then becomes percentage based and takes on an arbitrary scale.

10 hours ago, SeanZ said:

3. What workflow should I use when processing raw photos？

Whatever works best for you, I think. My approach, which I explained above, is pretty fool proof, but you can do as little or as much work during the initial development as you feel is appropriate, e.g. you might want to add luminance noise reduction during the development stage, perform defringing, change white balance etc. Just be aware that if you perform certain steps like noise reduction during the initial development, you can't undo them. With noise reduction, I prefer to add a live Denoise layer in the Photo Persona, then mask out the areas I want to keep as detailed as possible. Again, though, it's down to personal choice.

Hope that helps!

[SeanZ]

17 hours ago, James Ritson said:

Hi @SeanZ, here's a quick breakdown of the pixel formats and what happens:

As you've observed, when you open a RAW file in Affinity Photo it's processed in a working pixel format of 32-bit float, which is unbounded. This prevents values outside the range of 0-1 from being clipped and discarded, and allows for highlight recovery. Colour operations are processed in ROMM RGB (ProPhoto), which helps colour fidelity even if the intended output colour space is sRGB. You are essentially working in the highest quality that is reasonable.

When you click Develop, by default, the pixel format is converted to 16-bit integer, which is not unbounded. Any values outside the range of 0-1 will be clipped and rounded, so you should ensure you are happy with the tones before you proceed (e.g. using highlight recovery, changing black point etc). The colour space is also converted to whichever option you have chosen—by default, this is sRGB, but you can change this by clicking the Profiles option on the right hand side and choosing another output profile like ROMM RGB or Adobe RGB.

I say by default, because you can change the output format to 32-bit HDR on the develop assistant (the tuxedo/suit icon on the top toolbar). Be very mindful, however, that 32-bit in Affinity Photo is a linear compositing format as opposed to nonlinear. Adjustments will behave differently (typically they will seem more sensitive) and blend modes may produce unexpected results. I would avoid working in 32-bit unless you either want to author HDR/EDR content—this is not the same as HDR merging and tone mapping—or you need the precision for more esoteric genres like deep sky astrophotography. A lot of customers think working in 32-bit is going to offer them the best quality possible. Whilst this is technically true, there are many caveats and I would seriously advise against it unless you have a specific requirement for this format.

To answer your questions:

Technically, the answer is yes, but it's like I mentioned above: unless you have a very specific requirement to work in 32-bit, any loss in quality will be negligible. 16-bit nonlinear precision is more than enough for 99% of image editing use cases. Here's an example of my workflow: I will often choose a wider colour profile like ROMM RGB, then make my image look as flat as possible in the Develop Persona, usually by removing the tone curve and bringing in the highlights slightly if they are clipping. I'll then develop to 16-bit and do all of my tonal adjustments in the main Photo Persona. I have yet to complain about any loss in quality 

The functionality is more or less the same, but the Develop Persona has more intuitive sliders. In fact, in demos I will frequently explain to people that if you want to do some simple destructive edits to an image, you can simply go into that persona and use a few sliders rather than have to build up a layer stack. One key difference will be the white balance slider: when developing a RAW file, it has access to the initial white balance metadata and the slider is measured in Kelvin. Once an image is developed, however, this slider then becomes percentage based and takes on an arbitrary scale.

Whatever works best for you, I think. My approach, which I explained above, is pretty fool proof, but you can do as little or as much work during the initial development as you feel is appropriate, e.g. you might want to add luminance noise reduction during the development stage, perform defringing, change white balance etc. Just be aware that if you perform certain steps like noise reduction during the initial development, you can't undo them. With noise reduction, I prefer to add a live Denoise layer in the Photo Persona, then mask out the areas I want to keep as detailed as possible. Again, though, it's down to personal choice.

Hope that helps!

Thanks a lot, @James Ritson. It's really clear.

[Cecil]

On 12/29/2019 at 8:57 PM, SeanZ said:

For the photo of 14bits Raw format,  it enters directly to Development persona when I open it and displays RGBA/32(HDR), which can be adjusted for exposure, contrast, saturation, etc.. Then I click the button to enter Photo Persona,  it becomes RGBA/16. I can also make some adjustments in the studios. If I come back to the Photo Persona again, the picture is RGBA/16.

So, I have some questions:
1. Will the photo quality be reduced When entering into Photo Persona from Development persona？
2. Is there any difference between adjusting in Development persona and Photo Persona? 
3. What workflow should I use when processing raw photos？

I think you for your OP.  James Ritson response is an outstanding example, of Affinity’s forum value.  

[Picts]

I'm new to this forum, but I think this is where I might ask a question about the RAW developer.  Does it work only on RAW files, or can its features be used on JPG photos as well?