Scum dots & noise from pixels with 1-2% dot density

[bluesky]

Hi, this my first post to this forum...

I have created an advertising flyer in Photoshop 7.0 which is basically photographs of our products on a white background. The source artwork was prepared at 600dpi. I provided the artwork to the printers saved as a CMYK pdf. The artwork looked great on screen and looked OK when printed out on our fairly mediocre postscript colour laser printer.

When the flyer came back from the printers I was very disappointed with the result:

1) The edges of the products where surrounded by a lot of very visible stray dots/noise. i.e. It looked like there was a swarm of midges flying around edges of the products

2) A fairly solid yellow area on one of the product pictures was also speckled with the same dots/noise.

I confronted the printer about the poor quality and after investigation they said this was because the noise existed in the source document.

When I went back to look at the source artwork I did find that there was a scattering of pixels with 1% to 2% dot density around the edge of the products (but invisible on my good quality monitor).

Also, the yellow area which was approximately 100% Y + 25% M, also contained some pixels with 1% C. i.e. 100%Y + 25%M + 1%C. But again on my monitor, as one would expect, the yellow area looked perfectly clean with smooth tonal variation.

I have done some research and understand that these 1% pixels are often referred to as scum dots. I also understand that these scum dots and noise can be introduced by the profile used for the RGB to CMYK conversions. I also appreciate that high jpeg compression ratios can introduce this type of pixel noise. (Incidentally, although the product images were originally jpegs they were of a high quality and not heavily compressed).

So, I understand basically how these 1-2% pixels can get into an image. But, the questions I really want answered are:

1) Is it correct that pixels with only 1%-2% dot density in the source image should come out in print as *very* visible speckles, when these pixels appear white on a computer monitor, and are also invsible when the image is printed on our postscript laser printer?

2) Should the yellow area of the image look similarly covered in "black" speckles just because some of the pixels are defined as 100%Y+25%M+1%C?

3) If these 1-2% pixels really can cause this level of visible speckling then whose responsibility is it to ensure the source artwork is appropriately "cleaned" so that hard copy is noise free, as seen on computer monitor? Is it my responsibility or the responsibility of the print shop? BTW, the equipment used for printing was described to me as a letterpress machine.

BTW, we have used the same product images in an A3 poster that was printed out on a professional wide format inkjet poster printer. There was no such noise visible on this A3 poster.

Any help with understanding what's happened here would be greatly appreciated.

[LoganBlade]

1 a monitor is not printing. i can make the image red on monitor if i adjust calibration on it and it wont be red if i print it.

2 you should use a page layout program then you could make sure images have clean edges from the picture boxes.

3. proofer cant mimic a 1 or 2% dot that is why it printed ok.

4 it is the designers duty to know what happens when going to press.

you now know some of the things that can happen and can fix by using proper program for best end results. Good luck next time

[zox]

I agreer with Logan but won't be as hard on you

Usually we pick up these 1-2% dots on our proofing equipment.
We flag it to customer at that stage and it is always AA, which means we fix it and customer/designer pays for it.

Process of fixing we call "sealing the background".

Yes, sometimes it gets missed and it gets printed like that and it's never our fault if it's supplied with 1% in background, you'll get it on press.

Next time be more careful and make sure either background is sealed or you make clipping path.

Good luck.

[bluesky]

Thanks for your replies. OK, I would be grateful for your advice/opinions on a few followup questions:

1) Why do 1% dots not show up on our postscript laser printer and why did they not show up on the professional wideformat inkjet poster printer?

2) Why is there not some sort of "Proofing" mode in Photoshop that mimics on screen how these 1% pixels will print out?

3) Why does the yellow area which is approx 100% Y, 25% M, 0% C (+/- 1 or 2 %) show up all speckled. I speculated that this was because whilst most pixels showed 0% C some showed 1%C. Presumably the yellow would have looked fine if it was mainly 100% Y, 25%M, 2%C with some pixels showing 1% but never 0%? i.e. what I am saying is - Presumably the problem is the non-linear jump from 0% to 1%. If the yellow area had 1% throughout with 2% or 3% in the odd pixel then this would have looked fine. Yes??

4) I did ask for a hard copy proof before requesting the print run so that I could be part of the "quality control" process. However, I was told if I wanted a hard copy proof there would be a charge (which was fairly high in my opinion). Consequently I settled for their standard digital proof. The digital proof looked just as clean as my original artwork and I was told my artwork was "fine". I did still have to pay a modest proofing charge. Now I accept that there may be colour variations between how a digital proof looks on screen and the final paper output, but surely I shouldn't have to accept such a difference in visible noise. Zox, I know you said that you "usually pick up these 1-2% dots", but surely this is something I should be able to expect from *any* good print shop's proofing process given the drastic effect it has on output!? After, all I did have to pay a proofing charge!? I don't know, maybe this is wrong forum to be looking for the "right" answer to my question :wink:

5) If I had requested a hard copy proof then this would presumably have been printed on some sort of inkjet printer, before the plates were made. Given that the proof would be printed on different equipment presumably there is no guarantee the noise would have shown up? After all, as I say, the noise didn't show up on our postscript laser printer.

6) What is the best way to fix the image? I could get rid get rid of the speckling around the edge of the images by cutting out the product images more carefully and ensuring that any subsequent RGB to CMYK conversion or jpeg compression does not reintroduce the noise. However, there are various cutouts within the product, it's not just the outer edge that would need cleaning up so there is a risk of missing somewhere given that the pixels are invisible on the screen. How would you suggest "sealing the background".

7) How do I get rid of the speckling from the yellow background. I could use the "Curves" tool to convert any pixel with less than 2% dot density to 0%. Indeed this should also get rid of the noise around the edges of the image. Is this what you would suggest?

8 ) Zox - you mention a "clipping path". This is not something with which I am familiar. Is it supported by Photoshop 7.0? Is embedding clipping paths a better approach than just recutting out the white areas from the product images?

9) Zox you mention that you usually pick up these 1-2% dots on your "proofing equipment". For my understanding, is this a piece of hardware or a software proofing tool? LoganBlade you mention that these problems can be fixed by using a "proper program for best end results". What software tools would you both recommend?

10) Would I be better off providing the artwork in RGB form and leaving CMYK conversion to the plate producer? I notice that if I convert my CMYK image back to RGB, the yellow background does not contain any B pixels - just R + G. So it does look like the 1% C pixels are created during the RGB to CMYK conversion. I believe I am using the Photoshop default profile - US Web Coated SWOP (v2). Or alternatively, should I be asking the print shop to provide me with a profile and then use this to do the RGB to CMYK conversion. Trouble is, I presume there is still a risk that the profile will create 1% cyan noise on the yellow background so I am still left with the need to "clean up" the image if this can't be left up to the print shop.

I'd be grateful for any advice you can give me.

[zox]

1. Maybe it's not that professional either of them. They have different way of producing dot as opose to screen dot on a plate and real professional equipment. Resolution could be part of reason too.

2. There is. If you put your eyedropper on your white background and it reads 1%, it will show up on plate as 1%, or at least it should be. What more do you want?

3. Possibly, hard to tell without proper test.

4. That digital proof is probably comparable with what you call Professional wide format inkjet or even worse, some digital colour laser, Pictro, ... hence the low price but you get what you pay for.
That expensive proof is probably your best bet if it's real screen dot proofer, like Kodak XP but in that case it would be pricy and I am sure it would show 1% screen.
Believe it or not this 1% we pickup on our low-end Iris43wide on 360 dpi, thta's the only thing it's good for It all depends on device and setup, YMMV.

5. Right

6. If we get it, we seal it by going to curves and pulling down 1% dot to 0.
Image is not affected overal much by 1% and it seals the bacground. This is quick and dirty trick.
I suggest you start working on your color management skills and try different setups with the same problem image until you find one that does not introduce 1% in white and color is pleasing.
It's all in ICC.

7. Look at number 6 answer.

8. Clipping paths go back as far as I can remember, maybe even from original version of Photoshop.
They can be slow method of clipping and before all this new transparency capabillity of new modern software, clipping path was the only way to safely cut out image from background and puting it on another or have the object "float". They still are very reliable method but one that I see less and less. Today's modern tools can do very good job of automatically creating clipping path. I used to do a thousands of clipping paths manually, node by node, which in itself is a skill. Zoom in 200% on image, start with first node and off you go, the key is to create great path without introducing too many nodes.

9. Look at my answer number 4.

10. If you are not experienced with color, I would advise you to leave it in RGB with instruction to your printer to pay attention to it and convert to CMYK. This way if your printer introduce the same problem (which is quite possible, believe me) they would be held accountable and be responsible for fixing, not you. If youw ant to learn, ask your printer for direction, profile, etc.. and do tons of tests. Youc an even send your tests to printer to give you proof and you can ask for deal since you will be sending "real" job so they won't charge you, etc..
Anything is possible and the key to success is to learn and talk to your printer, you can gain a lot of knowledge this way.

Good luck.

[bluesky]

Thanks a lot Zox for taking the time to answer my questions.

I have to say, I still feel responsibility should largely lie with the print shop for ensuring that the artwork I provide is going to at least reproduce acceptably on their equipment. I had to pay a compulsory "proofing charge" and to me that should include a once-over of the artwork for any obvious problems. BTW, maybe I didn't make it clear - by "digital proof" I meant a soft copy pdf - i.e. not really a proof at all! The expensive hard copy proof was something I enquired about and not something I was encouraged to take advantage of. All indications were that the artwork was fine and there was no need for a hard copy proof.

Incidentally, I believe the equipment used was a flexo press and from searching the web I see that this sort of press can have a high dot gain and the tonal jump from 0% to 1% is very big. This is illustrated in the following graphic:

http://www.esko.com/files/images/pro...-flexobump.gif

Indeed, I stumbled across a tool from Esko called FlexoFix which is designed to prepare offset-images for flexography printing:

http://www.esko.com/product.asp?id=108

Although I appreciate that the above link is a marketing spiel for Esko's product this link does provide a very good overview of flexo printing issues and does suggest that preparing an image for flexo printing has its own special challenges. From reading this spiel about Flexofix and from reading around on the web about flexo printing I do feel that the printer should take a degree of responsibility for the prepress work involved in ensuring the plates produced from the source image are going to produce a reasonable quality image on their particular equipment, especially since the plates are expensive and there is no going back once they are made. In my mind, responsibility for quality checking (+ minor corrections) to the client's artwork should be "all part of the service sir". Even if making "minor corrections" is not all part of the service they should at least be spotted and conveyed to the client.

Anyway, maybe a forum of prepress specialists is not the place to argue that the responsibility lies with them :wink:

To be fair to you Zox you did say you usually pick up these sorts of problems and flag them to the client so perhaps you do consider it all part of the service - although as you say, if you happen to miss a problem its never your fault...of course :wink:

Incidentally, the Flexofix plugin for Photoshop sounds like just the job but I don't suppose it very affordable for the end user? It seems to be aimed at the flexo printing industry, but then why would they need such a tool if the responsibility for producing "plate-ready" artwork lies with the client :roll: :wink: Anyway, enough of my sarcasm

Thanks again for having taken the time to answer my questions - it is much appreciated!

[rtwdwk]

Dont take this the wrong way but.....

I have to say, I still feel responsibility should largely lie with the print shop for ensuring that the artwork I provide is going to at least reproduce acceptably on their equipment. I had to pay a compulsory "proofing charge" and to me that should include a once-over of the artwork for any obvious problems. BTW, maybe I didn't make it clear - by "digital proof" I meant a soft copy pdf - i.e. not really a proof at all! The expensive hard copy proof was something I enquired about and not something I was encouraged to take advantage of. All indications were that the artwork was fine and there was no need for a hard copy proof.


you sent in your file with a 1% background dot.
You expected your printer to question your creative.

you should be happy that your printer can reproduce a 1% dot, there are many who can not.

If you are going to take a PDF as a soft proof, get in to the habit of going to separation preview to preview the individual CMYK plates.
USe the % tool to scan your white background to make sure there are no scum dots in the background.

[shazam]

The critical info here was 'it's a flexo job'.
A GOOD flexo printer may print a 1% dot as 10%. Many others may show as much as 20%!
For flexo, we set the RIPs to ignore anything less than 1% in the file, and for a couple of the 'other printers', ignore anything less than 2%.
The Esko plug in works well, but it's over $5000.
Watching carefully what your profiles do, and a bit of curve work in Photoshop is usually enough.The problem with JPEG introducing 'scum dots' in a normally white area goes WAY back, and is pretty well known in the trade.

[Prepressamess]

[quote="bluesky"]Incidentally, I believe the equipment used was a flexo press and from searching the web I see that this sort of press can have a high dot gain and the tonal jump from 0% to 1% is very big. This is illustrated in the following graphic: quote]

unnn how can nothing turn into something. Please explain how a 0%(nothing) can change to a 1 % (lowest possible dot).

Its like; it is a beautiful day outside and I am using a special umbrella that is known to absorb more water than any other umbrella. I am using this umbrella to catch the rain that is falling on my head, but there is no rain so hence no need for a umbrella.

[Prepressamess]

[quote="bluesky"]Incidentally, I believe the equipment used was a flexo press and from searching the web I see that this sort of press can have a high dot gain and the tonal jump from 0% to 1% is very big. This is illustrated in the following graphic: quote]

unnn how can nothing turn into something. Please explain how a 0%(nothing) can change to a 1 % (lowest possible dot).

Its like; it is a beautiful day outside and I am using a special umbrella that is known to absorb more water than any other umbrella. I am using this umbrella to catch the rain that is falling on my head, but there is no rain so hence no need for a umbrella.