How to convert an RGB file to a grayscale file using only one channel

Just to be clear: I have asked two different questions, but responders keep getting them mixed up. One is about grayscale differences; the other (this thread) is asking can PS preserve a Channel Mixer setting when converting RGB to grayscale – and the answer is "It can't".

Something I forgot to mention previously because I am dealing here with only B&W negatives: when I use Channel Mixer, one of the settings I tick is "Monochrome", so I am only dealing with grayscale images within an RGB file. Why do that…?

The workflow I have been using stems from dealing with scanned slides. I was assuming I could use the same workflow, just add a Channel Mixer layer to select the best channel from the B&W neg scans. However, since PS can't keep my selected channel when converting to grasycale mode, I'll have to put together a new Action specially for negs.

Re the comments:

In addition, you have adjustment layers that have no meaning in grayscale, like Channel Mixer. How did you expect that to be translated to grayscale? Stop and think about it.

Agreed that Channel Mixer has no meaning in Grayscale, but I was using it in RGB, and when I selected the Green channel, PS showed me a grayscale image of the green channel. And a grayscale image is perfectly acceptable in a grayscale colour space. I still can't see any reason why PS couldn't ask me, when I want to change modes, when I have ticked the monochrome option: "Do you want to turn the Channel Mixer layer into a new grayscale layer, or do you want to delete it?"

Maybe my understanding is wrong, but I see a Channel Mixer layer as just another image layer that has been mixed together from underlying channels. Under what circumstances, given a "monochrome" setting, would problems occur if a Channel Mixer layer was converted to a new grayscale layer in grayscale mode?

But you cannot keep adjustment layers! You have to flatten. Again, this has nothing to do with Photoshop.

I disagree. I have just changed modes from RGB to grayscale, kept all the adjustment layers (except the Channel Mixer layer which was substituted with the Green channel), and the images are identical. Five layers were involved:

• three Curve layers (for contrast adjustment)
• a Clone layer (non-destructive cloning)
• and a Flare layer, to counteract a small problem when using a Nikon Coolscan to scan high-contrast slides.

The images before and after the mode change were identical.

If you use VueScan  you can scan in grayscale, and choose to use only the green channel for the scan.

VueScan supports most scanners on the market, but not Imacon/Hasselblad.

Problem is: sometimes the red or blue channels give better results, so I always scan in RGB and select the best channel. One example I recently came across was part-discolouration of the negative, either during original processing in the 1951, or through age. Choosing the red channel made the discolouration almost invisible.

Anyway, thanks for all the feedback. I've learnt quite a bit, some rather esoteric, but all most important for my dealings with B&W negatives.

1. PS doesn't allow a Channel Mixer setting to survive a mode change.
2. To retain the look of an image when changing modes, the original and destination modes must have the same gamma.
3. Channels use the gamma of the Working Gray, which must be the same as the image gamma if the same look is to be retained when copying and pasting a Channel to a image layer.

Thanks for the detailed response, but I would rather not flatten. The reason is: I am scanning about 300 historically significant negatives from the 1950s – photos taken by two men of bushwalking in Tasmania. One of the fellows (now deceased) has just had a feature named after him, and the other has first ascents to his credit. Images will end up as part of a Blu-ray. So until the AV is finalised, I want to retain all layers in case I need to re-edit. I save a flattened 1080P, 8-bit version for working with inside Premiere, because Premiere will choke if I give it large tiff files (100-200MB). But for archiving, I'd rather retain all adjustment layers.

These are not well exposed photos in general. Over exposure is the main problem, and for a negative I find that difficult to correct. I work in 16-bit because of the extreme corrections, but that makes the file sizes huge, not helped by the files being in RGB, with two of the channels superfluous. A typical neg is below.

Normally I do not need to keep the red and blue channels: they are not as sharp as the green. But that is not always the case. Occasionally the red or blue channel has a better look.

To begin an edit, I first check out which channel to use by viewing at 100% for sharpness. I then apply an Action which adds six layers, one of which is a Channel Mixer so I can select the channel to use (usually green, but sometimes red or blue). Sometimes during the edit, it becomes obvious that the Green channel, even though sharper, has other imperfections not visible in the other channels. So I change channels. When the edit is complete, I'd like to save the file as grayscale with only the chosen channel. To save space.

That's why I'd like PS to be respect my chosen channel – otherwise I'll have to manually bring the chosen channel into play at the end of the edit and then delete the background. But it seems PS wasn't set up for that type of manipulation.

So, is there a way in PS for retaining my chosen channel and deleting the others when I convert to grayscale?

Or better yet provide an exemplary file?