QImage versus Lightroom for Printing

[enduser]

In Qimage, you can "Print to File" and then view the processed image in whatever software you use. Doing this allows all kinds of comparisons without actually doing a print.

[Robert Ardill]

Hi! Robert you can also go to the site below and get some advice, there are some very experienced and helpful users that offer great tips. http://ddisoftware.com/tech/qimage-ultimate/
Also some learning videos here including one on DFS.
http://www.ddisoftware.com/qimage-u/learn.htm

Many thanks Denis - I've taken the plunge and bought QU (I hope I get more use out of it than I did from QI Pro!).  The main issue for me was the change in workflow, but I think it's going to work out.

Robert

[Bart_van_der_Wolf]

In Qimage, you can "Print to File" and then view the processed image in whatever software you use. Doing this allows all kinds of comparisons without actually doing a print.

Exactly the way I did my 3x magnified example.

One then looks at the print-flie output at the appropriate zoom setting on display for an impression of how it would look in print (within the usual limitations of comparing such different output media). Use the Photoshop View Print size menu, or calculate: zoom% = (display PPI / Print PPI) for an exact displayed size, or use the Qimage Image Examiner (<Ctrl>X while hovering over a thumbnail of the printfile), and zoom out to the approx. required level.

Cheers,
Bart

[Denis de Gannes]

Many thanks Denis - I've taken the plunge and bought QU (I hope I get more use out of it than I did from QI Pro!).  The main issue for me was the change in workflow, but I think it's going to work out.

Robert

Hope you enjoy it as much as I do.

[robgo2]

I have been using Qimage as my go to printing software since 2003 and have never had the temptation to move to Photoshop or Lightroom. Lightroom is my go to software for processing my raw files and I have used since the initial Beta version. I have Photoshop CS 6. To match the resizing options in Qimage you would have to use the likes of Perfect Resize 8 (formally Genuine Fractals) as a plugin for PS CS or Lightroom at a cost of $150.

Ah, yes, but you can buy Pefect Resize 8 standalone version for only $50.  If you want a PS plugin, I would advise stepping up to the Perfect Photo Suite 8.5 for $180, which is money well spent, IMO.  It contains Perfect Resize as well as seven other modules that can be very useful.  I use the standalone version, and It has become my main photo editor these days.  I think that CS6 is starting to feel unwanted.  However, PPS does not have a printing module.  For that, one can use PS or LR.

I have no experience with Qimage, as I am on a Mac, and I print with ImagePrint.  None of what I wrote above is meant as a comparison of Perfect Resize and Qimage with regard to resizing and sharpening quality.

Rob

[Denis de Gannes]

I am really only focusing on the original thread topic "Qimage versus Lightroom for printing". I have both and IMO I prefer Qimage to Lightroom for quality prints. I am well aware of the quality resizing options of Perfect Resize, the point I was making is you would need something like this along with Lightroom to match Qimage. I already have Qimage so I am not necessarily interested in also purchasing additional software.
There are lots of options in the market so there are choices, the final test of what is best for you and your taste rests with you.
Print quality can only be assessed by actually viewing the final print. In the final analysis you would need to make a print using Lightroom and one using Qimage and decide for your self.   

[Robert Ardill]

Hi,

I've done a fair amount of testing (and used a lot of paper!!), so I thought I would report back with my conclusions.

PRINT RESOLUTION

First of all, I checked out the effect of print resolution, from the link that Bart gave me: http://www.luminous-landscape.com/forum/index.php?topic=54798.msg447163#msg447163.

I did the testing on my HPZ3100, which has the following print options:
Render: 300ppi Print 1200x1200
Render: 600ppi Print 1200x1200
Render: 600ppi Print 2400x1200

I printed on a high gloss paper after having calibrated the printer/paper and profiled the paper using i1Profiler and an i1Pro2.

DITHERING SMOOTH COLORS (I checked 3 different shades).
First of all I eliminated the 2400x1200 print resolution (the printer does extra passes to achieve this) as the dithering was LESS smooth than at 1200x1200. I could see this by scanning the print at 4800dpi.  There may be some advantage for diagonal lines, for example, but I see no advantage for photography.  

The smoothest dithering was on the 600ppi 1200x1200 (when viewed using a 4800dpi scan).  However there is no visible difference to the eye between this and the print at 300ppi 1200x1200. There may be a barely noticeable difference using an 8x loupe. So, from a tone/color point of view there seems to be no discernible advantage in going to 600ppi. BTW: the printer matches the 300ppi dithering when printing at 600ppi by doubling the number of passes; the spot sizes are identical: so my conclusions here may not be valid for an Epson as the Epson print head is capable of changing the spot size.

DOWNSIZED PRINT
With the print downsized for 300ppi but at print resolution for 600ppi, there is no visible difference between the prints (by eye or with an 8x loupe).

UPSIZED PRINT
With the print at print resolution for 300ppi (but upsized x2 for 600ppi), there is no visible difference between the prints (by eye or with an 8x loupe).
With the print upsized by 2x for 300ppi (and 4x for 600ppi), the 300ppi print is better, as expected (although the difference is surprisingly small).

All of the resizing was done using Photoshop with resampling set to Automatic.  I used Nik Sharpener 3 for the final output sharpening, after resize, same amount in all cases (the only other sharpening was a small amount of raw sharpening in Lightroom).

CONCLUSION
My conclusion is that there is no advantage (with the Z3100 at any rate) in printing at 600ppi. IMO, the best setting is 300ppi for the print, 300ppi rendering resolution in the print driver and 1200x1200dpi printing.  My view is that upsizing to 600ppi will only damaging the image.  However, if the image resolution is at 600ppi or higher (in other words, a very small print – about 24x16cm for a 1DS3 21MP image), then there may be a very slight advantage in printing at that resolution.  

I haven’t tested on matte paper, but I expect the same results – but with less sharpness overall because of the ink diffusion on the matte paper.

QIMAGE v. PHOTOSHOP/LIGHTROOM
As far a printing goes, there is no difference between QImage and Photoshop (I haven’t tested against Lightroom as I expect that LR and PS will be the same or very similar).  It comes down to sharpening and resizing, and there is quite a difference there, as one would expect.

SHARPENING
I’m new to QImage (having just purchased it) so I may not be using it optimally.  The test was a 3x upsized image printed at 300ppi on both Photoshop and QImage.

QIMAGE DFS
At first sight, the QImage Deep Focus Sharpening appears to be really excellent. It sharpens the image with no visible haloes. However, what it does is to lighten and darken pixels, not just at the edges, but over a much wider area, so that it’s a bit like adding local contrast.  This certainly gives the impression of sharpness, but at the expense of tonal change in the image and desaturation of the lighter colors.  For example a light colored leaf or reed may easily become leached to white or to a much paler color.

On the other hand, what is good is that the DFS appears to be applied AFTER resizing.  If used carefully it gives a good result.

SHARPENING AND UPSIZING
I made 3 comparisons:
1. Upsize and DFS sharpen in QImage (effectively one operation)
2. Upsize in Photoshop; Smart Sharpen; Output Sharpen using Nik Sharpener Pro 3
3. Upsize in Photoshop; Smart Sharpen; Output Sharpen + Local Contrast using Nik Sharpener Pro 3

With 2 and 3 I used a raw smart object with the Smart Sharpen filter below the Nik Sharpener filter (so that the Smart Sharpen is applied first).

1 appeared sharper than 2.  On close examination I could see that this was due to the increased local contrast in the QImage print, so I added a small amount of Local Contrast in the Nik Sharpener filter (the filter can apply Output Sharpening, Structure, Local Contrast and Focus at the same time) and reprinted the image as 3.  At this stage QImage and Photoshop prints appeared equally sharp.

There is really nothing to separate the two prints 1 and 3 viewed at normal distances.  However it is clear that the QImage print, 1, has a loss of detail as you can see in the scanned image below.  There’s a sort of posterization effect, so that the rock, for example, appears to have gray patches where there should be a variation in tone, resulting in a loss of structure to the surface.  Also, there is a leaching of color in the reeds for example (not visible in the scanned image below).

CONCLUSION
My conclusion is that QImage does a superb job and does so almost automatically.  I do think that it is possible to achieve a better result in Photoshop, and certainly a much more controllable one: so I will continue to use Photoshop for critical prints, but I do intend to use QImage where ‘pretty good’ is good enough.

Nik Sharpener Pro 3 (which I had but wasn’t using) really does a very good job of output sharpening, local contrast and ‘structure’ and it has a great feature in that the image can be soft-proofed to show how it should look on print (from a sharpening point of view).  This soft-proofing (sharpening) feature would be a fantastic addition to QImage.  However I prefer Photoshop Smart Sharpen for normal sharpening to the Nik Focus sharpen: Smart Sharpen is very controllable and it’s easy to cut out haloes, damp down shadow noise etc. The Nik Focus sharpen has no controls except for Strength.



You can see a larger version of this image here: http://www.irelandupclose.com/customer/LL/QUvPS.jpg

I hope this has been of interest and with any luck we’ll have an interesting discussion to follow!

Robert

[Denis de Gannes]

Hi! Robert see you have been doing some testing. This thread started out comparing Qimage versus Photoshop/ Lightroom and appears you need Nik Sharpener Pro to try and even out the contest. Sorry I would not be able to comment further since I have never had a reason to use Nik Sharpener Pro.

[Robert Ardill]

Hi! Robert see you have been doing some testing. This thread started out comparing Qimage versus Photoshop/ Lightroom and appears you need Nik Sharpener Pro to try and even out the contest. Sorry I would not be able to comment further since I have never had a reason to use Nik Sharpener Pro.

Hi Denis,

I just used Nik Sharpener but you could equally well use Smart Sharpen for output sharpening.  In Lightroom you would use the Print Sharpening.  I'll give these a go and see if there is a significant difference.

I didn't use Nik Sharpener to even out the contest ... as I mentioned, I've purchased QImage Ultimate, so I have a vested interest in QU being better!

Robert

[Bart_van_der_Wolf]

Hi,

I've done a fair amount of testing (and used a lot of paper!!), so I thought I would report back with my conclusions.

Hi Robert,

Thanks for the feedback. I don't have time to address all tests you did, but allow me to add a few observations before I turn in for the night.

I assume you used Fusion interpolation. While pretty good, there are other methods as well, so depending on image content you may want to try a few of the others as well. Hybrid SE e.g. is free of ringing/halos, and can be sharpened more without creating too much halo.

I'm also not completely familiar with the specifics of the HP printer used, so there may be settings that could have an influence. It e.g. took a while for Epson printer users to discover the 'Finest detail' option, without which the printer downsamples to a lower resolution. I'm not saying such a thing happens with the HP, just that I don't know. Make sure to check the PPI indicator at the top right above the layout preview, it will show what the printer will do.

As for dithering, it makes a difference whether you send 8-bit/channel data to Qimage, or 16-b/ch. The latter gets dithered upon import, and (if selected) after sharpening and conversion to the output profile for your media.

As for downsampling, Qimage uses an adjustable amount of anti-aliasing. Depending on subject structure, you may be able to switch it to a lower setting than what you used. 'Medium' is default, so 'Lower' or even 'Off' may create sharper results, but potentially also some clearly aliasing artifacts in e.g. brick structures or similar repetitive structures.

As for upsampling, it is possible to use special sharpening techniques that are better than what Qimage can do, but at the expense of more work. That makes it a choice for more or less convenience. However, if you do want to take more time, then you can also do that with Qimage. When you save the upsampled print output file without Smart DFS sharpening, you can also use the image editor and roll your own level (radius and amount, and even targeted by color range) of sharpening,  and save that as a print filter for automatic future use. So you only create such an optimization once, and re-use it for multiple files by checking the (P.filt.) check-box at the bottom right of the interface.

That's it for the moment.

Cheers,
Bart

[Robert Ardill]

Hi Bart,

Thanks for taking the time to reply.

I assume you used Fusion interpolation. While pretty good, there are other methods as well, so depending on image content you may want to try a few of the others as well. Hybrid SE e.g. is free of ringing/halos, and can be sharpened more without creating too much halo.

Yes, I did use Fusion with a setting of 5.  As I mentioned, I'm very new to QImage (even though I've had QImage Pro for years), and so I'm quite likely not using the optimal settings.  Unfortunately (as far as I can see) the full image editor preview and other previews don't show the resized image with (the resizing) sharpening, or the effect of the different resizing algorithms, so it's only possible to gauge this by saving the file and then opening in Photoshop.  I've no doubt that getting the right resizing/sharpening is critical.  Perhaps you could convince Mike Chaney to add a soft-proof like the one in Nik Sharpener ... it's not a feature I've seen elsewhere and it's really tremendous.

I'm also not completely familiar with the specifics of the HP printer used, so there may be settings that could have an influence. It e.g. took a while for Epson printer users to discover the 'Finest detail' option, without which the printer downsamples to a lower resolution. I'm not saying such a thing happens with the HP, just that I don't know. Make sure to check the PPI indicator at the top right above the layout preview, it will show what the printer will do.

I have an Epson 4800 which is currently out of commission (and I had a 4000), so I'm quite familiar with these printers (as good as the HP, but no better IMO). The settings on the HPZ3100 is clearer than on the Epson because the print driver shows the Rendering PPI and the Print DPI (so for example, you can choose a rendering intent of 600ppi and a print resolution of 1200x1200dpi.  If you then send the printer a file at anything but 600ppi, the driver will resample to 600ppi before printing.  You can also select rendering at 300ppi and print at 1200x1200 and the printer will lay down exactly the same number of dots as it would if the file rendering was at 600ppi and print at 1200x1200.  You can also select more passes and more detail (2400x1200), but apart from the print taking more time, the print quality difference is not visible to the eye, IMO, and only barely with a loupe.  I didn't align the print heads before doing the tests and this could have an effect.

As for dithering, it makes a difference whether you send 8-bit/channel data to Qimage, or 16-b/ch. The latter gets dithered upon import, and (if selected) after sharpening and conversion to the output profile for your media.

Does QImage work entirely in 8-bit? Seems like it from http://forums.steves-digicams.com/printing-qimage-support-forum/54299-does-qimage-support-16-bit-files.html, where Mike Chaney says: “Dither is only used when you have pallete driven color like 256 color images. When converting from 16 bits/channel to 8 bits/channel, the information is simply remapped to 0-255 instead of 0-65535. All image programs do the same thing when converting from 48-24 or 24-48. None of them use dithering because it isn't necessary.”

If so, that’s not so great as I would certainly prefer to do things like upsizing and sharpening in 16 bits.

As for downsampling, Qimage uses an adjustable amount of anti-aliasing. Depending on subject structure, you may be able to switch it to a lower setting than what you used. 'Medium' is default, so 'Lower' or even 'Off' may create sharper results, but potentially also some clearly aliasing artifacts in e.g. brick structures or similar repetitive structures.

 

Thanks for that – so far I’ve not had to downsample much and haven’t had an aliasing problem, but I’ve stored it in memory for the future.  That’s certainly a nice feature!

As for upsampling, it is possible to use special sharpening techniques that are better than what Qimage can do, but at the expense of more work. That makes it a choice for more or less convenience. However, if you do want to take more time, then you can also do that with Qimage. When you save the upsampled print output file without Smart DFS sharpening, you can also use the image editor and roll your own level (radius and amount, and even targeted by color range) of sharpening,  and save that as a print filter for automatic future use. So you only create such an optimization once, and re-use it for multiple files by checking the (P.filt.) check-box at the bottom right of the interface.

What I’ve found so far is that DFS is OK but only with a radius of 1 … or maybe 2, max.  It seems better (on screen) to use a higher amount with a radius of 1 than a lower amount with a radius of 2.  The example I give above used a radius of 2. 

As I mentioned, it’s very difficult to gauge the overall sharpening because the full-screen editor doesn’t show the effect of the resize sharpening (or I don’t know how to get it to show that).

The printer filter is a very nice feature – however again the issue I would have with it is that its effect can only be seen on the saved file or on the print, not on screen. 

What I didn’t mention is that QImage is much better than Lightroom from the point of view of print layout, print borders, viewing the print with adjustments on the page etc.  I’m focusing more on the negatives than the positives of QImage at the moment … but I really should say that in my opinion it is an excellent print program.  I don’t think I would use it for raw development because programs like Capture One and Lightroom have a lot of extra features compared to QImage for raw development.  However, if I was printing loads of raw images and if I wasn’t looking for optimum prints then I would certainly consider QImage from raw to print.

My dilemma is keeping my overall workflow simple but effective for low-volume printing, but quite high volume image selection and development.  Lightroom is very good for image management, developing, map module, web, slideshows etc; it’s just OK for printing IMO.  As I always use Photoshop in between development and print/web, I do have the option of proper resizing, sharpening etc., in Photoshop, and then using Lightroom just for print output without any resizing or sharpening.  Introducing QImage somewhat complicates the issue for me as QImage does have a lot more control over print processing than does Lightroom. That’s why I’m so interested in finding out if QImage can do as good as (or even nearly as good as) a job as I can do in Photoshop in terms of resizing and output sharpening, because I would then use QImage for that purpose.  But that also means that the sharpening has to be as good in QImage as it is in Photoshop (whether I use Smart Sharpen, Nik Sharpener, edge masks/no edge masks … or whatever), because I do not want to do the output sharpening before the resizing, clearly.

If QImage had better visual previews than it currently does it would make the decision much easier for me: I’ve spent ages getting my system set up so that what I see on my monitor is really, really close to what I get on print … and QImage doesn’t quite work in that setup right now.

Robert

[Denis de Gannes]

Have you used the soft proof or image examiner feature in Qimage. See screen capture.
 

[mchaney]

Robert,

I fear (as I think you did) that you were not achieving the best results in Qimage because you are circumventing some of the features that help it deliver maximum quality automatically.  You mention a 3x upsampled image printed at 300 PPI.  You also mention using upsampling and DFS in Qimage.  I think what you did, in effect, forced Qimage to upsample and sharpen twice, which is certainly not optimal.  In Qimage, you don't need all the gyrations that you do in other programs: you'll achieve the best results by just printing the original file at the size you want.  Maybe I'm not 100% clear on exactly what steps you performed when, but if you had a 100 PPI image that you upsampled to 300 PPI, sharpened, and then printed... stop doing that.    Just drag the original 100 PPI image over to the page and print it at the size you want and then compare that result.  It'll automatically upsample, sharpen (based on the size you chose), and print.  If you need more sharpening, just set your output sharpening higher than the default (5).

I'm also more than a bit surprised to hear that you think you saw DFS overbrightening.  That's exactly what it doesn't do, as it was designed to avoid those artifacts.  For example, with USM, if you sharpen an image of sky shot through tree leaves, some of the small gaps/holes that should show sky between the leaves will actually turn from blue (sky) to stark white.  That's because USM over-drives the contrast.  DFS cannot and does not do that: where you got white specular dots through the leaves with USM, you'll get the true blue with DFS because it has a limiter that doesn't allow it to blow past the actual color of the objects being sharpened.  In fact, I don't see that in your print either.

Regards,
Mike

[Ernst Dinkla]

There has been a Z3100 driver/firmware combination that asks for 1200 PPI input resolution in soms driver settings. Not the most recent ones. The Z3200 drivers never went above 600 PPI input resolution.

On processing + printing time: the antialiasing filter in QU's downsampling can take more time in processing than the printing with 600 PPI input takes. So with big image files you may go that route even if the media is not showing more detail. Gradations usually improve with the 600 PPI input settings too, it is not just the details you should check.

Ernst Dinkla

[Robert Ardill]

Robert,

I fear (as I think you did) that you were not achieving the best results in Qimage because you are circumventing some of the features that help it deliver maximum quality automatically.  You mention a 3x upsampled image printed at 300 PPI.  You also mention using upsampling and DFS in Qimage.  I think what you did, in effect, forced Qimage to upsample and sharpen twice, which is certainly not optimal.  In Qimage, you don't need all the gyrations that you do in other programs: you'll achieve the best results by just printing the original file at the size you want.  Maybe I'm not 100% clear on exactly what steps you performed when, but if you had a 100 PPI image that you upsampled to 300 PPI, sharpened, and then printed... stop doing that.    Just drag the original 100 PPI image over to the page and print it at the size you want and then compare that result.  It'll automatically upsample, sharpen (based on the size you chose), and print.  If you need more sharpening, just set your output sharpening higher than the default (5).

I'm also more than a bit surprised to hear that you think you saw DFS overbrightening.  That's exactly what it doesn't do, as it was designed to avoid those artifacts.  For example, with USM, if you sharpen an image of sky shot through tree leaves, some of the small gaps/holes that should show sky between the leaves will actually turn from blue (sky) to stark white.  That's because USM over-drives the contrast.  DFS cannot and does not do that: where you got white specular dots through the leaves with USM, you'll get the true blue with DFS because it has a limiter that doesn't allow it to blow past the actual color of the objects being sharpened.  In fact, I don't see that in your print either.

Regards,
Mike

Hi Mike,

Many thanks for your comments - and I'm sure I'm not using QImage to its best advantage.  You're right, what I did was to upsize by 3x (as a test) using Fusion with a sharpening of 5; I then applied a DFS sharpening with a radius of 2 and amount of 100.  The reason for doing that is that the image looked very soft on the previews (all of them) ... and the reason for that as far as I can see is that QU does not show the sharpening that has been applied during resize (which I assumed, when I did the test, that it did).  I will try again without the second sharpening and hopefully the results will be better.

As I've mentioned, the issue with this approach is that I will either have to print to file and examine the file to see how much sharpening has been applied during resizing; or I will have to print, possibly several times, until I get the sharpening I want.  I don't much like either of these options, especially the second one.

It would be very useful if the QU previews showed the image with the resize sharpening.  And it would be even more useful if it could show a soft-proof, as does Nik Sharpener Pro, showing how the sharpening is likely to appear on the particular paper type selected (at the print size).

Regarding DFS: what I see rather than haloes is a sort of posterisation. If you look at the image I posted you should be able to see this, if not you can download a 25% downsized zip of the scan here: http://www.irelandupclose.com/customer/LL/QUvPS-L.zip.

This image below is interesting to look at:



The image is resized to show the pixels.

On the left is a gradient going from light gray to middle gray with a darker gray background (made in Photoshop).  On the right is the same image with DFS 3/200 applied.  As you can see, where there is a transition, the lighter pixels have been lightened and the darker ones darkened – which is similar to what the equivalent USM would do, but less aggressively, as you can see here:



However, what DFS also does is to lighten and darken a much wider band than the radius (of 3 in this case), whereas USM limits the sharpening to the radius.  The effect of that can be (as I’ve shown in the images) to flatten, or give a posterized look, to small areas (which you won't get with USM or some variant like Smart Sharpen).  

In the DFS pixel-sized test above you can also see that the pixels are spread in a semi-random pattern, which may give an impression of texture, but can also result in flattening because light and dark pixels are spread next to each other over a large area.

It’s not that I think DFS is bad (I’m on the fence at the moment) … but I think it needs to be used with a small radius (not more than 2) and low amount (not more than 200), otherwise there is a strong risk of this sort of posterized look.

Anyway … that’s just my (current) opinion, and as I’ve said I am a total QU newbie!

Robert

[jrsforums]

Robert,

No amount of screen proofing will ever be able to tell you if sharpening is correct....only printing will.  Based on experience you may get close, but it is essentially experience guessing.

Qimage allows you to test paper/printer combinations, adjusting the amount of DFS sharpening applied.  Then, assuming use of same interpolation, the results will be consistent for essentially whatever size you decide to print to.

My starting point into Qimage is a print that looks good on the screen.  I then export an 8 bit Tiff in PrinterRGB, which I then print what ever size I need, cropping to aspect ratio in Qimage.

John

[Robert Ardill]

There has been a Z3100 driver/firmware combination that asks for 1200 PPI input resolution in soms driver settings. Not the most recent ones. The Z3200 drivers never went above 600 PPI input resolution.

On processing + printing time: the antialiasing filter in QU's downsampling can take more time in processing than the printing with 600 PPI input takes. So with big image files you may go that route even if the media is not showing more detail. Gradations usually improve with the 600 PPI input settings too, it is not just the details you should check.

Ernst Dinkla

Hi Ernst, the driver I have only allows 600ppi max (I can't imagine how the driver would handle 1200ppi as that would be equivalent to 1 pixel per dot: it would have to downsample or throw away pixels!).

I have checked gradients at 300ppi and 600ppi and I really don't see it on print.  If I scan the print at high resolution I can see that there is a difference in the dithering, with the 600ppi being smoother ... but even on a high gloss paper it's invisible even with a magnifying glass (and on this test I score 100% http://www.xrite.com/online-color-test-challenge, so I must have fairly good color vision .  I personally would never upsample to 600ppi unless the image resolution was already very close to it.

Robert

[Robert Ardill]

Robert,

No amount of screen proofing will ever be able to tell you if sharpening is correct....only printing will.  Based on experience you may get close, but it is essentially experience guessing.

Qimage allows you to test paper/printer combinations, adjusting the amount of DFS sharpening applied.  Then, assuming use of same interpolation, the results will be consistent for essentially whatever size you decide to print to.

My starting point into Qimage is a print that looks good on the screen.  I then export an 8 bit Tiff in PrinterRGB, which I then print what ever size I need, cropping to aspect ratio in Qimage.

John

Yes, well of course that's true ... up to a point.  However I do think that if you stick to a limited range of papers and to a limited size of prints and to certain types of images that it's possible to get a very accurate idea of what the print will look like ... with experience and testing, as you say.  To be honest, I've only just come across the Nik sharpen proof preview and I'm not at all sold on it (although I'm keeping an open mind).  I think it's better to stick to the screen at print size and to judge the sharpening visually, taking into account the softening on the print due to ink diffusion and dithering (so sharpen a bit more, but not too much more).

Perhaps it's a lack of trust on my part, but I much prefer to judge these things visually rather than rely on an automated system (even if this has been tuned for paper and size).

The problem with your workflow, to my mind, is that if the image looks good on the screen, you will already have applied sharpening to it.  If you then resize it you will be resizing the sharpening too, so that it will be too much if you are upsizing and too little if you are downsizing ... and then you will be applying sharpening with the resizing presumably, so sharpening on top of sharpening.  I think it's much better to resize and then to sharpen for output - and that may mean more than one type of sharpening (for example local contrast and edge-sharpening).

Robert

[Robert Ardill]

Have you used the soft proof or image examiner feature in Qimage. See screen capture.
 

Hi Denis,

Yes, I have (both) ... but they don't show the resize sharpening (to my knowledge).

Robert

[jrsforums]

In the Fraser/Schewe three steps of sharpening, I appliy the first two steps to the native image until it looks the way I want it on the screen.

The final step, output sharpening, is applied by Qimage based on output size, and any adjustments for specific paper.  I do not have to remember the differences by paper.  Once I have tested and arrived at what I consider the correct settings, Qimage remembers and will use the same settings whenever I select that paper.