Fundamental Photoshop Question

[blansky]

On my initial foray into Photoshop about a year ago I don't think I ever got clear on the sizing issue. To be safe I always cropped the image to 20x24 immediately then began to retouch it.

I usually print 20x24, 24x36, and they come into photoshop from the card at about 57 meg. The part the escapes me or never really got settled is, at what size does one start to manipulate and deal with retouching of the image.

What I've been doing is sizing the image to say 20x24 or the largest I will print it and then start to work on it. When I'm done, this is what I call my "master". It is often over a gigabite file. Then if I want an 8x10 say, I'll down size it to that, then print it. Same with say 16x20. I often keep all these size files in a folder which is pretty large. I keep them all because when printing, the adjustments are usually the same except for the exposure.

My questions is:

Is it the same thing to, when bringing the file into photoshop just leave it at the 57 meg and do all the adjusting and retouching at that size then call that the "master" (which will obviously be a lot smaller file than what I've been doing,) and then upsize to the 20x24 etc when deciding the sizes I want to print, or should I upsize immediately when bringing the file into photoshop.

The disconnect in my mind seems to be that: is it better to work on large files of an image or not. Will the image have the same quality if retouched at 8x10 and upsized to 16x20 as it will be to be retouched at 16x20 and saved at that.

Thanks in advance.

[Schewe]

The retouching should ALWAYS be done at the highest resolution (dimensions & PPI) you are EVERY likely to need...retouching at a smaller size and then up-sampling will up-sample the retouching which will usually be less good (more obvious) than retouching on the largest size and down-sampling.

Tone & color correction can be done on lower rez images before up-sampling but the accuracy issue would still be true if you used a lot of adjustment layers with layer masks that used masking to localize the correction. The masks would have to be up-sampled just like an actual image.

You should also note that the physical size in Photoshop is really only ever determined by the pixel dimensions, not the units in inches & PPI. The dimensions are derived based upon the ratio of dimensions vs PPI. So, you can have a large dimension image but it'll have less PPI (pixels per inch) than the same pixel dimensions would than a smaller images size at higher PPI.

[blansky]

OK. Please be patient.

With my Canon 5d an image comes into photoshop through ACR at 72.8 MP and 2912 x 4368 (360) showing a document size of 8.089 x 12.133 at 360.

If I increase the size without resampling to 20x30 the resolution becomes 145.6.

My workflow is to bring the image in, crop to 20x24 with the resample checked and get a file: pixel dimensions 7200 x 8640 and document size as 20 x 24 at 360 resolution. The pixel dimensions now is 356 MP.

So you are saying that my workflow is correct and retouching at this size is better than the initial 72.8 MP.

I gues my question is I just upsized the file which means that I interpolated or added (copied) some pixels so why is the retouching better here than at the original (real) size.

Please remember that I only used my workflow this way because I wanted to be on the safe side but really didn't know what the hell I was doing. The problem of course is that my files are now one to two gigs in size when saving a couple of different image sizes that I printed.

Thanks again.

Michael

[01af]

To be safe I always cropped the image to 20x24 immediately then began to retouch it. [...] and they come into Photoshop from the card at about 57 meg.[a href=\"index.php?act=findpost&pid=196275\"][{POST_SNAPBACK}][/a]

To be safe from what? 20 × 24 what? 57 meg what?

The part the escapes me or never really got settled is, at what size does one start to manipulate and deal with retouching of the image.[a href=\"index.php?act=findpost&pid=196275\"][{POST_SNAPBACK}][/a]

At the size of the image file.

-- Olaf

[MichaelEzra]

below are the steps I would recommend:

1. create a master file at camera's resolution and minimal sharpening
2. do most of the retouching, preserve textture
3. save master file
4. create a copy of the master file
5. upsample to resolution needed for printing
6. finish retouching
7. save master-print file
8. create a copy of master print file and covert it to output color profile
9. adjust image to specific limitations of the output profile.
10. save and print

[Mike Louw]

With my Canon 5d an image comes into photoshop through ACR at 72.8 MP and 2912 x 4368 (360) showing a document size of 8.089 x 12.133 at 360.

I would also appreciate some advice, as another 5D user.

To produce the best possible print quality, on an Epson 3800, from a Canon 5D file, would it in general be better to:

1. Reduce the ppi to 240 to produce a larger image, or

2. Leave the ppi at 360 and "up-sample" the image (i.e.: interpolation)?

Thanks

Mike

[Wayne Fox]

I would also appreciate some advice, as another 5D user.

To produce the best possible print quality, on an Epson 3800, from a Canon 5D file, would it in general be better to:

1. Reduce the ppi to 240 to produce a larger image, or

2. Leave the ppi at 360 and "up-sample" the image (i.e.: interpolation)?

Thanks

Mike
[a href=\"index.php?act=findpost&pid=196308\"][{POST_SNAPBACK}][/a]

Do not resample the image.  Simply size it _without_ resampling to the final output size, leaving it at it's native resolution, and send that to the printer.  Be sure to perform output sharpening after you have set it's output dimensions.

If the listed output DPI in the printer driver goes above 480, you may get better results re-sampling the file by 50%, and if it get's below 180 or so, you may get better results re-sampling by 200%.  Using even increments allows for optimum results with bicubic sharpening. After re-sampling, set the output size again if needed without resampling, perform output sharpening and send the new native size to the printer.

There is a very good section in "From Camera to Print" that discusses this workflow and how to use it.  I was a little skeptical at first, but have found it is better than my previous workflows, including master files resampled to specific DPI.

[Nill Toulme]

Or just give Qimage a try.  It changed my life.  ;-)

Nill
~~
www.toulme.net

[AJSJones]

I usually print 20x24, 24x36, and they come into photoshop from the card at about 57 meg. The part the escapes me or never really got settled is, at what size does one start to manipulate and deal with retouching of the image.


The disconnect in my mind seems to be that: is it better to work on large files of an image or not. Will the image have the same quality if retouched at 8x10 and upsized to 16x20 as it will be to be retouched at 16x20 and saved at that.

Thanks in advance.
[a href=\"index.php?act=findpost&pid=196275\"][{POST_SNAPBACK}][/a]

The best answer to disconnect you from your current way of thinking is for you to say to yourself " my image has no physical size until just before I print it; it only has dimensions measured in pixels".  Keep the original pixels (don't create or throw away any), do all your postprocessing except "output sharpening" , save the master and THEN do the final crop.  Only now do you decide what size print you want.  Qimage seems to be highly recommended for adjusting image pixels to output pixels (i.e. scaling to larger output)  (not for Mac alasso I have no advice on sharpening using Qimage).  On my images I'll go down to 240 ppi  (original pixels) - if that's not a big enough print, only then will I scale up - then apply output sharpening for my paper and ppi combination and print (PKSharpener).  The scale-up (uprez) method is a personal choice and to some extent subject matter dependent.

Andy