From: INTERNET:rhansen@2540dpi.f2s.com, INTERNET:rhansen@2540dpi.f2s.com
Date: Fri, Jan 4, 2002, 10:25 AM
RE: [colortheory] resaving jpeg
I've heard conflicting information on whether opening, editing and resaving a jpeg file adds more loss to the file. Some say every time you resave a jpeg the lossy compression causes more loss. Others have said this only occurs if you do a Save As and make a new file. Just saving the original doesn't cause any more loss. Does anyone have an official answer to this?Thanks,
RJay
http://www.2540dpi.f2s.com
From: Gordon Pritchard
Date: Fri, Jan 4, 2002, 12:54 PM
RE: RE: [colortheory] resaving jpeg
RJay wrote:
> I've heard conflicting information on whether opening, editing and
> resaving a jpeg file adds more loss to the file.
Hi, Here is a technique for visually evaluating the compression artifacts of CT images. (thanks to Richard Zaleski ex Sells Printing Company LLC and Murray Oles - Treasure Chest)Gordon Pritchard
Commercial Print Avatar
CreoScitex
Vancouver Canada
T: 604.451.2700 ext 2870
C: 604.351.2437
http://www.creoscitex.com>Print - dot's what it's about!<
(1) Open both the compressed and non-compressed photos in Photoshop 5.02.
Have the non-compressed photo selected.(2) Select the Calculations dialog from the Image pull down menu.
(3) For Source 1, choose your non-compressed image, background layer, Cyan channel. For Source 2, choose your compressed image, background layer, Cyan channel. For Blending, Select Subtract, 100% opacity, offset 0, scale 1. Leave the mask checkpoint deselected. For the result, select new channel. Then click on the "OK" button.
(4) This creates a new channel in your non-compressed image. Target this channel in your new channels dialog, rename it "subtract C" or some other intuitive name. Use the Image~Adjust~Invert command on this channel.
(5) Perform a "Edit~Select All" on the "subtract C" channel. Next, perform an "Edit~Copy".
(6) Now perform a "File~New" command to create a new document. Photoshop senses what is on the clipboard and automatically puts in the pixel dimensions of what is on the clipboard. Select the mode as CMYK.
(7) Target the Cyan channel of this new document and perform an "Edit~Paste Into".
(8) Repeat steps 2 through 7 for the magenta, yellow and black channels. The resulting new image is an accurate visual representation of what areas changed between the compressed and non-compressed images.
You might want to use your "Levels" command to increase the contrast in the final cmyk image to make the "differences" in the image more clear.
A caveat: although you can see differences in the file -- they may never end up being visible in the final image on press since the halftoning process, due to its lack of resolution, will effectively "screen" out some of the artifacts. You might want to also to take a representative image (includes flat screen areas, gradients, and fine detail) at one or two basic input resolutions and perform JPEGging at three or four scales of compression/quality. Impose the JPEGged images along with the original and output halftone contract proofs at 2 or 3 typical LPIs (say 133, 175, 200, FM). Then use that as your real-world reference hard copy proof visual guide to the effects of different levels of JPEG.
The calculations method does, however, show you how the JPEG algorithm works on images which can, in turn, help guide you as to where to look for artifacts on other images.
From: Joe Butts
Date: Sat, Jan 5, 2002, 7:41 PM
RE: [colortheory] resaving jpeg
In a message dated 1/5/02 7:44:48 AM Pacific Standard Time, Gordon Pritchard<
gordon_pritchard@creoscitex.com>
writes:> (1) Open both the compressed and non-compressed photos in Photoshop 5.02.
Specifically 5.02? Who's still on 5.02? Will 6.0.1 work? Why not?
Joe Butts
From: Dan Margulis
Date: Tue, Jan 8, 2002, 10:55 AM
RE: [colortheory] resaving jpeg
RJay writes,>>I've heard conflicting information on whether opening, editing and resaving a jpeg file adds more loss to the file. Some say every time you resave a jpeg the lossy compression causes more loss. Others have said this only occurs if you do a Save As and make a new file. Just saving the original doesn't cause any more loss. Does anyone have an official answer to this?>>
I suspect that nobody knows, because nobody has taken the time to save and resave a bunch of JPEGs 20 times in different ways and investigate what happens. The reason nobody has done that is there would be a widespread consensus that any workflow that caused the same image to be JPEGged over and over is not a very good one. Storage space is very cheap now. Files that you believe are likely to get revised shouldn't be JPEGged--if you need a JPEG for the web or whatever, make one from your working file, but don't keep compressing the same one as you make edits.
Similarly, I have no problem with archiving images as JPEG when you really think that its just an archive that's unlikely to be touched later.
Dan Margulis
From: steven moreno
Date: Tue, Jan 8, 2002, 11:30 AM
RE: Re: [colortheory] resaving jpeg
Dan, you would lose quality if your JPEG were already compressed and you were to compress it again. The JPEG compression scheme reduces file size by discarding information. Always use an uncompressed JPEG file from the start for a quality JPEG.
From: RJayHansen
Date: Tue, Jan 8, 2002, 1:41 PM
RE: Re: [colortheory] resaving jpeg
Quoting Dan Margulis:> I suspect that nobody knows, because nobody has taken the time to save and
> resave a bunch of JPEGs 20 times in different ways and investigate what
> happens. The reason nobody has done that is there would be a widespread
> consensus that any workflow that caused the same image to be JPEGged over
> and over is not a very good one. Storage space is very cheap now. Files
> that you believe are likely to get revised shouldn't be JPEGged--if you
> need a JPEG for the web or whatever, make one from your working file, but
> don't keep compressing the same one as you make edits.Thanks to everyone who replied.
This issue arises for us (a print shop) when receiving files from customers with some images saved as jpeg. If we have to go in and edit any of their jpeg files, then resaving as tiff or eps can make updating the page layout file more difficult, ie. QuarkXpress will assume a file extension indicates the correct format of the file. If I give the resaved file a .tif extension Quark won't see it as a modified version of the original placed file since it has a different name, but if I save as tiff and leave the file name the same (with a .jpg extension), Quark will say it's the wrong type of file.
My practice before has been to save as jpeg with maximum quality/minimum compression, and from most of the answers I've received here, this still seems like the best solution to this situation.
By the way we do ask our customers to submit image files as tiff or eps, but as anyone who's worked in the prepress industry knows, you can't control what the customers give you.
RJay Hansen
http://www.2540dpi.f2s.com
From: Les De Moss
Date: Tue, Jan 8, 2002, 8:07 PM
RE: Re: [colortheory] resaving jpeg
As a practical matter, in our pro-lab and studio we open jpg's we receive (or create in our studio), immediately save as a tiff, and archive the "original" jpg. All subsequent work is performed and saved as an uncompressed tiff. Only when the final edit is approved, do we resave as a jpg - if it is necessary for the intended use. The final tiff file is always archived, and considered the "original" in the event that other uses surface. IOW: jpg format is only employed for a specific use, always created from a working tiff file.-Les De Moss
Gerards Photo Lab, Inc.
Fort Collins, Colorado
www.gerardsphoto.com
From: William Garabrant
Date: Tue, Jan 8, 2002, 6:29 PM
RE: Re: [colortheory] resaving jpeg
From: Dan Margulis
> I suspect that nobody knows, because nobody has taken the time to save and
> resave a bunch of JPEGs 20 times in different ways and investigate what
> happens.I have... But I used layers set to difference mode to check the degradation. The damage is negligible (when saving out of PS6 at quality 12 with a hi-res photo). At 100% view, after 10 re-saves, there's barely anything to see. Zoomed to 400%, I could see little bits of "difference". After 20 re-saves, there was visible artifacting at 100% view, but really nothing traumatic.
We're sometimes forced by databank and clipping path restrictions to use eps/jpg. We're also forced sometimes by annoying customers to make a lot of late changes to files. Sometimes that means re-jpgging, and sometimes the degredation is apparent. Depends on the original file, of course.
But, what gets me upset is that the jpg encoding in Photoshop's eps files is NOT the same quality as regular jpgs. A file saved as eps/jpg max will always be smaller and show more artifacting than one saved as jpg/12...
William Garabrant
Kulmbach, Germany
From: Karl Snyder
Date: Tue, Jan 8, 2002, 6:28 PM
RE: Re: [colortheory] resaving jpeg
On Tue, 8 Jan 2002 16:35:20 -0000, your pet rodent escaped it's cage, climbed the computer table, ran across the keyboard and accidently hit keys that wrote:>When I was taking my MSc a colleague was required to investigate and report
>on the JPEG format; he tested the resaving of various files at various
>compression rates. The outcome was that no noticeable loss occurred to a
>file that was resaved up to 4 times (using the higher compression settings);
>however, the file size was not reduced very much in size after the first 2
>saves. A quality fall off was quite noticeable after the 5th save. The
>general opinion was to open the original JPEG file, modify, then save using
>a lossless compression; thus not incurring any further damage.
>
>But Dan indicated that any workflow that uses JPEG again and again would be
>seriously wrong; JPEG being a fantastic format for many things but not
>archiving your original files!
Craig,When most save as JPEG, they do so to save space thus use a high compression factor. I've never tried to resave at 100% to see what the successive generation would be like. I've just never found a reason to do so. I know taking a 20% file, editing it and saving it again at 20% results in a very bad looking image.
Just curious as why you do not accept in PSD format? Is it because Quack does not accept that format 'cause it is from Adobe?
Miss spellings of proper names intended!
Karl S.
Karl Snyder
Boulder, Colorado
http://www.RockyMountainNationalPark.Info
http://www.EstesParkOnLine.Com/
http://www.MtEvans.Com/
Moderator of the Yahoo OutdoorPhotographers group
To join:
OutdoorPhotographers-subscribe@yahoogroups.com
From: Pylant, Brian
Date: Tue, Jan 8, 2002, 6:28 PM
RE: RE: [colortheory] resaving jpeg
> If I give the resaved file a .tif
> extension Quark won't see
> it as a modified version of the original placed file since it
> has a different
> name,Delete or move the original .jpg and Quark will report that it cannot find it, at which point you can relink to the .tif and Quark will retain your cropping and positioning.
Brian
From: Chris Murphy
Date: Tue, Jan 8, 2002, 8:07 PM
RE: Re: [colortheory] resaving jpeg
Karl Snyder writes:>Just curious as why you do not accept in PSD format? Is it
>because Quack does not accept that format 'cause it is from >Adobe?I think Quack doesn't accept PSD because PSD is a proprietary format. I think ALAP is the only company that makes a product allowing another product (QuarkXPress XTension for QuarkXPress) to use PSD files. That likely has to do with a "special relationship" between Adobe and ALAP that allowed that to occur.
That said, I don't know if Graphic Converter even deals with the PSD format. I do know that the Mac OS X Finder is able to generate previews from PSD files - it literally reads the whole PSD document, layers and everything, and builds a preview based on it. Depending on the complexity of the document, this can take some time, but it will do it.
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: Craig Auckland
Date: Tue, Jan 8, 2002, 1:39 PM
RE: [colortheory] resaving jpeg
When I was taking my MSc a colleague was required to investigate and report on the JPEG format; he tested the resaving of various files at various compression rates. The outcome was that no noticeable loss occurred to a file that was resaved up to 4 times (using the higher compression settings); however, the file size was not reduced very much in size after the first 2 saves. A quality fall off was quite noticeable after the 5th save. The general opinion was to open the original JPEG file, modify, then save using a lossless compression; thus not incurring any further damage.But Dan indicated that any workflow that uses JPEG again and again would be seriously wrong; JPEG being a fantastic format for many things but not archiving your original files!
Regards
CraigCraig Auckland | Photographer
Telephone: 07930 337 226
Facsimile: 07931 607 428
From: Stephen Marsh
Date: Tue, Jan 8, 2002, 8:07 PM
RE: [colortheory] Re: resaving jpeg
RJay Hansen writes:> By the way we do ask our customers to submit image files as tiff or eps, but as > anyone who's worked in the prepress industry knows, you can't control what the > customers give you.
Chuckling...too true.
There is one suggestion I can think of - if your customers are proactive.
Suggest to your clients that they disable the JPEG import xtension in Quark - stop the problem at the source. Designers are often lazy or are forced to cut corners and a verbal 'dont use JPG' may not be enough. But if the artist building the layout is forced to import a TIFF or EPS from the start with no option of JPEG then this would be the ideal solution (well, ideal for you[g]). If your workflow has issues with LZW compressed TIFF, then you can also disable the xtension in Quark as well.
If the file is starting off in JPEG, then it could probably use some artifact removal in a duped colour blend layer or in AB channels of LAB - but this would be at the start of the process before edits and channel blends make any artifacts worse. The halftone screen will usually cancel out any minor artifacts caused from normal amounts of JPEG compression and artifact removal does not seem to make a better final composite print at this late point - only cleaner film or plate separations.
Regards,
Stephen Marsh.
From: Chris Murphy
Date: Tue, Jan 8, 2002, 9:04 PM
RE: Re: [colortheory] resaving jpeg
Dan Margulis writes:>Dan, you would lose quality if your JPEG were already
>compressed and you were to compress it again.
>The JPEG compression scheme reduces file size by
>discarding information. Always use an uncompressed
>JPEG file from the start for a quality JPEG.I was unaware there was such a thing as uncompressed JPEG. I thought by definition you got at least some compression whenever using JPEG. I know there is supposedly a lossless JPEG in the works, but I don't know the status of it.
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: Kevin Connery
Date: Tue, Jan 8, 2002, 11:16 PM
RE: [colortheory] RE: resaving jpeg
William Garabrant noted:
>>But, what gets me upset is that the jpg encoding in Photoshop's eps files is NOT the same quality as regular jpgs. A file saved as eps/jpg max will always be smaller and show more artifacting than one saved as jpg/12...>>
The JPEG specification doesn't mandate any specific choice for how the color quantization is done, nor does it mandate what the luminosity breakdown is; that's left to each implementation. As long as it can be DEcoded, the way it gets there--and the level of compression/color quantization/etc. that's done--is not an issue to the standard.
That's part of the reason why resaving can be more problematic for some than others; different tools operate differently. For example, the now-antique STORM PicturePress JPEG tool was deterministic; if you took an image and compressed it at a given level--ANY given level, high or low quality--re-opened it, edited one pixel, saved it using the identical compression level, edited that same pixel, resaved, etc., you could repeat as many times as you'd like, and the ONLY extra damage would be confined to the 64 pixels (8x8 block) containing the changed pixel. All the rest would be unchanged. I ran an AppleScript for this for over 500 open/saves without the editing, and there was absolutely no changes whatsoever. That was a 1991 or 1992 implementation.
Very few JPEG tools are deterministic in that same way, and resaving WILL cause all the cells to be recalculated each time, adding more quantization error at every save for the entire file.
Thus, even if the guy down the hall doesn't have a problem with resaving, you might--if you're using a different tool, or a different compression level, or whatever.
--kdc
From: Chris Murphy
Date: Wed, Jan 9, 2002, 9:14 AM
RE: Re: [colortheory] resaving jpeg
Karl Snyder writes:>If you do a "Save for Web..." you have the option in JPEG format
>to have 0 to 100 quality option. A 2.1 MB 8 BITs/channel RGB file
>would be saved as a 386 KB JPEG file...so there will be some loss
>at 100 "Quality"!What I'm responding to is the statement by someone Dan posted for regarding "always use an uncompressed JPEG file" and I'm saying I've never heard of such a thing. I've heard of lossless JPEG compression, but that is still compression. The idea of have a 3MB image saved as a JPEG that would be a 3MB file is a foreign concept to me. Whether you have lossy or lossless JPEG compression isn't the issue, but the idea there is such a thing as a NO compression JPEG. I don't think such a thing exists, unless it's called TIFF (with no compression).
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: Karl Snyder
Date: Tue, Jan 8, 2002, 11:17 PM
RE: Re: [colortheory] resaving jpeg
On Tue, 8 Jan 2002 18:16:14 -0700, your pet rodent escaped it's cage, climbed the computer table, ran across the keyboard and accidently hit keys that wrote:>I was unaware there was such a thing as uncompressed JPEG. I thought by
>definition you got at least some compression whenever using JPEG. I know
>there is supposedly a lossless JPEG in the works, but I don't know the
>status of it.
Hi Chris...also in The Peoples Republic of Bldr!
If you do a "Save for Web..." you have the option in JPEG format to have 0 to 100 quality option. A 2.1 MB 8 BITs/channel RGB file would be saved as a 386 KB JPEG file...so there will be some loss at 100 "Quality"!
Karl Snyder
Boulder, Colorado
http://www.RockyMountainNationalPark.Info
http://www.EstesParkOnLine.Com/
http://www.MtEvans.Com/
From: RJay Hansen
Date: Wed, Jan 9, 2002, 11:11 AM
RE: Re: [colortheory] resaving jpeg
Quoting Karl Snyder:> Just curious as why you do not accept in PSD format? Is it
> because Quack does not accept that format 'cause it is from
> Adobe?In our experience, tiff or eps formats are the most reliable for printing to an imagesetter. They separate correctly and don't cause PostScript errors (barring the occasional corrupt file).
Also, as to Brian's advice for deleting or moving the original, William Alexander emailed me offlist with the same suggestion. I should have thought of that myself. I've used the same technique before with other graphics files.
Sorry this subject ended up being rather off-topic (probably more appropriate for a prepress list), but it obviously generated a lot of interest and I appreciate everyone's input.
Thanks to all,
RJay Hansenhttp://www.2540dpi.f2s.com/scaletron.html
From: Dan Margulis
Date: Wed, Jan 9, 2002, 11:11 AM
RE: Re: [colortheory] resaving jpeg
Chris Murphy writes,>>I think Quack doesn't accept PSD because PSD is a proprietary format.<<
Adobe doesn't care if other people use the format. Corel products support it as do Macromedia's. My understanding is that Quark's is a deliberate workflow-aiding decision. An awful lot of people, myself included, do all their preliminary work in PSD format and only convert to TIFF once the job is ready to print. Thus, a TIFF on my system indicates a final file and a PSD is something that isn't ready for prime time.
In this scenario, we don't WANT Quark to be able to place a PSD.
>>That said, I don't know if Graphic Converter even deals with the PSD format.>>
It supports it fully, including writing.
Dan Margulis
From: William Garabrant
Date: Wed, Jan 9, 2002, 5:04 PM
RE: Re: [colortheory] Re: resaving jpeg
> Suggest to your clients that they disable the JPEG import xtension in
> Quark - stop the problem at the source. Designers are often lazy or are
> forced to cut corners and a verbal 'dont use JPG' may not be enough. But
> if the artist building the layout is forced to import a TIFF or EPS from
> the start with no option of JPEG then this would be the ideal solution
> (well, ideal for you). If your workflow has issues with LZW
> compressed TIFF, then you can also disable the xtension in Quark as
> well.But... jpg is an option with eps files. And unfortunately, often our only option :-(
And in the hands of the unwitting, perhaps a dangerous option, now that I think about it.Did you catch my statement about normal jpg files and jpg-encoded eps's, Stephen? Because of clipping paths, we use eps and because of a lousy, expensive, new databank
system, they mostly have to be jpg-encoded. Because of the sheer numbers of image files we store, keeping everything doubled up as tif and eps would be difficult (I can just see my boss's eyes if I were to suggest it, lol).
So... We have a lot of files that only exist as eps/jpg and we have some goofy customers who like to change things around (numerous times) before they've made up their minds. So, we sometimes find ourselves re-jpgging images two or three times. If the images weren't of the highest quality to begin with, we have a problem... But in reality it's those customers who generally don't know what they're doing and really don't appreciate quality, if you get my drift... :o)
William Garabrant
Kulmbach, Germany
From: William Garabrant
Date: Wed, Jan 9, 2002, 5:03 PM
RE: Re: [colortheory] resaving jpeg
> But, what gets me upset is that the jpg encoding in Photoshop's eps files is
> NOT the same quality as regular jpgs. A file saved as eps/jpg max will
> always be smaller and show more artifacting than one saved as jpg/12...
>
> Thus, even if the guy down the hall doesn't have a problem with resaving,
> you might--if you're using a different tool, or a different compression
> level, or whatever.That was an interesting reply, thanks. But I was talking about the difference in file quality and size when saving either as normal .jpg with quality 12 or eps/jpg max quality (from the same computer and the same program, Photoshop 6). The normal jpg files are bigger and don't degrade after multiple re-saves the same as eps/jpg... why?
Thanks,
William Garabrant
Kulmbach, Germany
From: Chris Murphy
Date: Thu, Jan 10, 2002, 8:08 AM
RE: Re: [colortheory] RE: resaving jpeg
Kevin Connery writes:>But, what gets me upset is that the jpg encoding in Photoshop's eps files is
>NOT the same quality as regular jpgs. A file saved as eps/jpg max will
>always be smaller and show more artifacting than one saved as jpg/12...I'm looking for confirmation from Adobe on on this, but if you go into a JPEG save as dialog, you will notice that if you click on the pop-up menu and select maximum that the numeric compression factor is not 12, but 10. I think that what's probably happening is that EPS JPEG encoding set to maximum is using a compression factor of 10.
The question I'm posing is, if this is the case, why is it that EPS JPEG compression is limited to a quality factor of 10, and JPEG and TIFF/JPEG compression both have a quality factor of 12 available to them?
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: William Alexander
Date: Thu, Jan 10, 2002, 10:23 AM
RE: Re: [colortheory] Re: resaving jpeg
This mention of eps/jpg encoded is the first I've ever heard of it, even though I've been working in PhotoShop since 1994 and have on many occasions noticed it in the eps save dialog box. I think the reason I've not acknowledged it is that I have been fortunate enough to never even consider it. I work at a small company, but storage space in regards to images has never once been an issue with my publisher, especially if the alternative is a threat to quality. Isn't there a combination of a couple hundred gigs of hard drive space and a cd burner that will prevent you from ever having to click jpeg at all? How many images could you possibly need available at your fingertips at one time? A cd archive is always only a disk drive away...William Alexander
Art Director
Leisure Publishing Company
540-989-6138
From: William Alexander
Date: Thu, Jan 10, 2002, 12:47 PM
RE: Re: [colortheory] Re: resaving jpeg
>>The JPEG option for EPS is only for preview data an doesn1t affect the output. If you need a clipping path, you need to save the file as an EPS so the option of picking JPEG only affects the on screen preview.>>Are you entirely certain of that Andrew? It seems to me that it refers to the encoding of the entire file... not in the "preview" submenu, but in the "encoding" submenu.
From: jrscratch
Date: Thu, Jan 10, 2002, 12:48 PM
RE: [colortheory] Re: resaving jpeg
--- In colortheory@y..., William Alexander wrote:
>Isn't there a combination of a couple hundred gigs of hard drive space
>and a cd burner that will prevent you from ever having to click jpeg
>at all? How many images could you possibly need available at your
>fingertips at one time? A cd archive is always only a disk drive
>away...I felt it necesary to point out, at least in my case... That the attractiveness of using jpeg compression is for file transmision--rather than a storage concern. As for Quality loss, That is another discussion.
Ron Scratch
Quality Analyst
RR Donnelley
From: Andrew Rodney
Date: Thu, Jan 10, 2002, 12:47 PM
RE: Re: [colortheory] Re: resaving jpeg
on 1/10/02 7:02 AM, William Alexander wrote:> This mention of eps/jpg encoded is the first I've ever heard of it, even
> though I've been working in PhotoShop since 1994 and have on many occasions
> noticed it in the eps save dialog box. I think the reason I've not
> acknowledged it is that I have been fortunate enough to never even consider
> it.
>
> But... jpg is an option with eps files. And unfortunately, often our only
> option :-(
> And in the hands of the unwitting, perhaps a dangerous option, now that I
> think about it.
The JPEG option for EPS is only for preview data an doesn1t affect the output. If you need a clipping path, you need to save the file as an EPS so the option of picking JPEG only affects the on screen preview.Andrew Rodney
From: Les De Moss
Date: Thu, Jan 10, 2002, 12:47 PM
RE: Re: [colortheory] Re: resaving jpeg
In a static, closed workflow you certainly have this option. In an open workflow that includes receipt of files from others, and internet transfer of large files between locations, compression becomes an issue. IOW: It's not a simply a question of storage availability within your particular workflow.Les De Moss
Gerards Photo Lab, Inc.
Fort Collins, Co
www.gerardsphoto.com> This mention of eps/jpg encoded is the first I've ever heard of it, even
> though I've been working in PhotoShop since 1994 and have on many occasions
> noticed it in the eps save dialog box. I think the reason I've not
> acknowledged it is that I have been fortunate enough to never even consider
> it. I work at a small company, but storage space in regards to images has
> never once been an issue with my publisher, especially if the alternative is
> a threat to quality. Isn't there a combination of a couple hundred gigs of
> hard drive space and a cd burner that will prevent you from ever having to
> click jpeg at all?
From: William Alexander
Date: Thu, Jan 10, 2002, 12:51 PM
RE: Re: [colortheory] Re: resaving jpeg
Even in an open workflow including huge quantities of huge files, it seems like using jpeg compression is defeating the purpose of the huge file size? If your #1 goal is quality, why risk it? Storage space is cheap... and if ftp transfer etc. convenience over-rules quality then quality must not be that high of a priority.> In a static, closed workflow you certainly have this option. In an open
> workflow that includes receipt of files from others, and internet transfer
> of large files between locations, compression becomes an issue. IOW: It's
> not a simply a question of storage availability within your particular
> workflow.
From: Mike Russell
Date: Thu, Jan 10, 2002, 2:48 PM
RE: [colortheory] Re: resaving jpeg
Just a note re jpeg re-saving.There is very little, though still some degradation, if a jpeg is re-saved using exactly the same numeric settings as for the original save.
This is very close to a lossless save, though small comfort for those who see no reason for any quality compromise. As a practical matter images are easier to produce and store in jpeg. So for those who are forced to re-save to jpeg there may be cold comfort in the fact that re-saving is not the anathema to quality that is the conventional wisdom here.
Photoshop's often misapplied histogram tool has a valid application in measuring image degradation. Subtracting images and noting the average and standard deviation of the resulting near-black image is a good quantitative measure of quality loss. I sympathize with those who feel that any change is bad, but as a strictly practical matter, how far off can the image be if the max pixel value change is 1, and only a few percent of pixels are changed at all?
Chris's mention of the differing range numbers of the jpeg quality settings reminds me of the movie Spinal Tap, where an amplifier was deemed superior by its owner because it had a max volume knob setting of 11 instead of 10:
http://www.nature.com/nsu/000113/000113-2.htmlhttp://www.zocalo.net/~mgr
http://geigy.2y.net
From: Chris Murphy
Date: Thu, Jan 10, 2002, 2:50 PM
RE: Re: [colortheory] Re: resaving jpeg
William Alexander writes:>Isn't there a combination of a couple hundred gigs of
>hard drive space and a cd burner that will prevent you from ever having to
>click jpeg at all?If you have a catalog with 50,000 items in it, EPS + JPEG might make sense. Since QuarkXPress can place a regular JPEG image without it needed to be EPS, I would think that would be a better way to go to get good compression at a higher quality setting than is available for EPS + JPEG.
Even *better* would be for Quark to get off their butts and support TIFF + ZIP compression, which has been around for a while. This is a good compression scheme, and is totally lossless. But QuarkXPress 4.11 has no idea what it is. It's not good enough to add support in QuarkXpress 5. They need to make an XTension to add support into 4.11 as well. InDesign 1.0 even supported TIFF + ZIP.
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: Chris Murphy
Date: Thu, Jan 10, 2002, 2:50 PM
RE: Re: [colortheory] Re: resaving jpeg
Andrew Rodney writes:>The JPEG option for EPS is only for preview data an doesn1t affect the
>output. If you need a clipping path, you need to save the file as an EPS so
>the option of picking JPEG only affects the on screen preview.There is a Preview pop-up menu which has one JPEG option. This does affect only the preview, which is used for placing images in page layout applications, and JPEG is your best bet there.
However, we are talking about the Encoding pop-up menu, where there are four JPEG quality options. This affects the entire contents of a Photoshop EPS file, and does use lossy JPEG compression to do it's job.
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: Les De Moss
Date: Thu, Jan 10, 2002, 2:50 PM
RE: Re: [colortheory] Re: resaving jpeg
Note that an open workflow includes receipt of file formats from outside sources over which we have little if any control, as well as output to jpg format when requested/required by outside users (for whatever their reasons may be). My point is that while quality over storage space may be an internal priority within your work flow -as it is in ours- it cannot be controlled in all aspects. Since jpg files are a fact of life, the issue becomes how we handle them internally.... and as I previously posted, and you seem to agree, employing uncompressed tiff files within our internal workflow is the obvious choice when image quality is the priority.As a side note- I have found it vastly interesting that the initial thread on this subject: "does resaving (Save as) a jpg file degrade quality, as opposed to simply resaving. As of yet, I have not seen a definitive answer to this *seemingly* simple question.
-Les De Moss
From: Les De Moss
Date: Thu, Jan 10, 2002, 2:50 PM
RE: Re: [colortheory] Re: resaving jpeg
Note that an open workflow includes receipt of file formats from outside sources over which we have little if any control, as well as output to jpg format when requested/required by outside users (for whatever their reasons may be). My point is that while quality over storage space may be an internal priority within your work flow -as it is in ours- it cannot be controlled in all aspects. Since jpg files are a fact of life, the issue becomes how we handle them internally.... and as I previously posted, and you seem to agree, employing uncompressed tiff files -or any appropriate uncompressed file format- within our internal workflow is the obvious choice when image quality is the priority.-Les De Moss
> Even in an open workflow including huge quantities of huge files, it seems
> like using jpeg compression is defeating the purpose of the huge file size?
> If your #1 goal is quality, why risk it? Storage space is cheap... and if
> ftp transfer etc. convenience over-rules quality then quality must not be
> that high of a priority
From: Andrew Rodney\
Date: Thu, Jan 10, 2002, 2:51 PM
RE: Re: [colortheory] Re: resaving jpeg
on 1/10/02 8:44 AM, William Alexander wrote:> Are you entirely certain of that Andrew? It seems to me that it refers to
> the encoding of the entire file... not in the "preview" submenu, but in the
> "encoding" submenu.Yup. Well that is, you have the option of saving the preview as JPEG. For Encoding, yes, you can go either way (I wasn1t clear sorry). Saving a JPEG preview will make the file smaller.
According to Real World Photoshop, it1s not a good idea to save JPEG encoding of data for inclusion into Quark (they say it prints a black image). Apparently there ARE applications that can save a loseless EPS using compression but not Photoshop.
I really wish that more people would support MrSid. I find it does a MUCH better job of compression with no ill effect (up to a point). A 10:1 JPEG doesn1t look as good as a 20:1 MrSid file. There is supposed to be a newer version coming that produces true loseless compression at 10:1 or less (current version is visually loseless). Compressing files as .Sid don1t seem to take much longer than JPEG. JPEG 2000 doesn1t seem like it1s ever going to make it to market (and it1s big deal is progress JPEG which is kind of cool for the web but not all that interesting anywhere else).
Andrew Rodney
From: Kevin Connery
Date: Thu, Jan 10, 2002, 4:13 PM
RE: RE: [colortheory] Re: resaving jpeg
Les De Moss wrote:
> As a side note- I have found it vastly interesting that the initial thread
> on this subject: "does resaving (Save as) a jpg file degrade quality, as
> opposed to simply resaving. As of yet, I have not seen a definitive answer
> to this *seemingly* simple question.The definitive answer is, sadly, "it depends".
The current JPEG specifications do not mandate how this is performed, and each application does it the way their development team felt was best for them/their customers/whatever code they had already.
SOME applications can open, save, open, resave, open, resave with NO impact whatsoever if the settings are unchanged. Storm's PicturePress was one of those. Ads for some current add-on JPEG tools include such claims but I've not tested it--resaving isn't an issue for me now.
MOST applications WILL do extra "damage" each time it's saved. Some applications are less damaging than others for each resave, but extra damage is done on each resave in those applications. The settings in a given application will have an impact on how much extra artifacting is done.
ALL applications will add extra compression artifacts when different settings are used and the image resaved.
There simply isn't a yes or no answer to this.
--kdc
From: Kevin Connery
Date: Thu, Jan 10, 2002, 2:51 PM
RE: RE: [colortheory] RE: resaving jpeg
Actually, the quote was from me quoting someone else...That said, no implementation of JPEG is guaranteed to use the same 'values'
for any compression level. Thus, Photoshop may have 1-12, 1-10, 1-100, or whatever, and Photodoodler could have the same 1-12, 1-10, or whatever, with completely different results. There's not even a guarantee--as Chris notes--that the same product will use the same compression in different parts of the same application, even if it uses the same numeric settings.There have been products where the numbers were reversed, with 10 being the most compressed and not the most quality.
The standard for this ... isn't.
--kdc
From: Chris Murphy
Date: Thu, Jan 10, 2002, 4:14 PM
RE: Re: [colortheory] Re: resaving jpeg
Mike Russell writes:>There is very little, though still some degradation, if a jpeg is re-saved
>using exactly the same numeric settings as for the original save.
>
>This is very close to a lossless save, though small comfort for those who
>see no reason for any quality compromise.I disagree. If I use a quality factor of 1, save the image, reopen that image and then save again as quality factor of 1, I see significant and noticeable additional artifacting.
>Chris's mention of the differing range numbers of the jpeg quality settings
>reminds me of the movie Spinal Tap, where an amplifier was deemed superior
>by its owner because it had a max volume knob setting of 11 instead of 10.Those numbers are arbitrary. It's entirely possible for a volume of 10 on one amplifier to supply more juice than a volume setting of 11 on another amplifier. The difference between JPEG quality 10 and 12 is not arbitrary and it is noticable. Is it noticable if you take two JPEGs, one at 10 the other at 12, and output them at 100% on a press? I would be surprised if there is any noticable difference. But if you needed to enlarge the image significantly, or you combined some enlargement and output to a higher resolution output device, yes you can see the difference. Each individual's tolerances will determine if the difference is a passing curiousity or a problem.
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: Les De Moss
Date: Thu, Jan 10, 2002, 6:51 PM
RE: Re: [colortheory] Re: resaving jpeg
Kevin- Understanding even that much helps provide a save-strategy based on the priority of a particular use. At best it points to testing in order to gain a deeper understanding of how each application codes it's jpg.... a difference I am sure many folks - including myself- did not appreciate.Thanks,
Les De Moss> The definitive answer is, sadly, "it depends".
>
> The current JPEG specifications do not mandate how this is performed, and
> each application does it the way their development team felt was best for
> them/their customers/whatever code they had already.
>
> SOME applications can open, save, open, resave, open, resave with NO
impact
> whatsoever if the settings are unchanged. Storm's PicturePress was one of
> those. Ads for some current add-on JPEG tools include such claims but I've
> not tested it--resaving isn't an issue for me now.
>
> MOST applications WILL do extra "damage" each time it's saved. Some
> applications are less damaging than others for each resave, but extra damage
> is done on each resave in those applications.
From: Jerry L. P'Simer
Date: Thu, Jan 10, 2002, 10:16 PM
RE: Re: [colortheory] Re: resaving jpeg
Andrew Rodney wrote:> The JPEG option for EPS is only for preview data an doesn1t affect the
> output. If you need a clipping path, you need to save the file as an EPS so
> the option of picking JPEG only affects the on screen preview.This is not true at all. I stopped using the eps format in my work flow with the release of QuarkExpress4 which allows the use of clipping paths with tiff files. Also, the only reason that I will use the eps format is because of jpeg compression allowing huge files to be transferred cross country via the internet. Providing this has only occurred once, there is no noticeable loss of quality for litho purposes. This format generates a smaller file than the jpeg format with like settings. Why this is I do not know.
I recall that a few years ago some of the operators I work with did use the eps format with jpeg compression for saving all files. About 60 hours of work was hosed as a result of files being resaved multiple times after a series of edits. As a result, the use of lossy compression was banned for use within our workflow.
Jerry P'Simer
From: William Alexander\
Date: Fri, Jan 11, 2002, 9:05 AM
RE: Re: [colortheory] TIFF/EPS (was resaving jpeg)
Hope this question isn't too basic, but why is Tiff better than EPS? (assuming that neither format is using jpg compression) I've always used EPS rather than TIFF but have no reason for it, and would switch in a heartbeat if there were advantages to using TIFF instead.William Alexander
From: RJay Hansen
Date: Fri, Jan 11, 2002, 11:40 AM
RE: Re: [colortheory] TIFF/EPS (was resaving jpeg)
Quoting William Alexander:> Hope this question isn't too basic, but why is Tiff better than EPS?
> (assuming that neither format is using jpg compression) I've always used
> EPS rather than TIFF but have no reason for it, and would switch in a heartbeat
> if there were advantages to using TIFF instead.
The joke in the prepress industry is EPS stands for Evil PostScript. Generally speaking, in our experience, TIFF is the most reliable, trouble-free format for PostScript printing. With EPS, you are nesting PostScript objects inside of more PostScript. As I indicated earlier in this thread, we ask our customers to submit their graphic files as TIFF or EPS. I also tell them to favor TIFF unless the file has to be saved as EPS (ie. file contains vector data or is in duotone mode or something).
That being said, there are good arguments for using EPS instead. For a good, quick comparison of some of the pros and cons of each format see this URL:
http://users.belgacom.net/prepresspanic/formats/tiffversuseps.htm
Incidentally, this guy's web site is a great source of information on various aspects of prepress and PostScript workflows.
RJay Hansen
http://www.2540dpi.f2s.com
From: Dan Margulis
Date: Fri, Jan 11, 2002, 11:40 AM
Re: [colortheory] TIFF/EPS (was resaving jpeg)
William writes,>>Hope this question isn't too basic, but why is Tiff better than EPS? (assuming that neither format is using jpg compression) I've always used EPS rather than TIFF but have no reason for it, and would switch in a heartbeat if there were advantages to using TIFF instead.>>
TIFF (at least the basic TIFF, not the misnamed "Advanced" TIFF introducedas an option in PS 6) is more bulletproof, particularly if it's being used by a client.
TIFFs are a lot kinder to the RIP, in that parts of the TIFF that are cropped out in the page layout aren't processed, as is unnecessary resolution. EPSs have to be processed in their totality. On a weaker RIP, such as those found on many desktop printers, this can add up to a huge amount of time if the EPSs are at the high resolution needed for press.
Also, many novice users inadvertently embed screens or transfer curves into EPSs. These are invisible to the next person, until the job implodes down the road.
Dan Margulis
From: mact-net
Date: Fri, Jan 11, 2002, 11:39 AM
RE: Re: [colortheory] TIFF/EPS (was resaving jpeg)
A tiff is almost invariably smaller, sometimes dramatically so
I just tried several variations on a cmyk tiff.
tiff = 28Meg.
eps (jpeg max encoding)= 12Meg
eps (binary encoding)=35 M
eps (ascii encoding)=73M
So a jpeg encoded file is smaller.
But quark or PageMaker cannot separate it. I just tried it again: Print separations and distill the result. the photo shows up, in color, on the K plate only.
So, jpeg encoding might be smaller but each one will have to be opened and resaved so it is much slower overall.
Mac Townsend,
Adcom Graphics, Fairfield, CA:
Electronic Prepress
www.adcomgraphics.com
From: Chris Brown\
Date: Fri, Jan 11, 2002, 1:07 PM
RE: [colortheory] TIFF/EPS
> Hope this question isn't too basic, but why is Tiff better than EPS?
> (assuming that neither format is using jpg compression) I've always used EPS
> rather than TIFF but have no reason for it, and would switch in a heartbeat
> if there were advantages to using TIFF instead.I prefer TIFF because it supports multiple alpha channels and is a smaller file size then EPS in uncompressed form.
Chris Brown
From: William Garabrant
Date: Fri, Jan 11, 2002, 6:10 PM
RE: [colortheory] Re: resaving jpeg - Goodnight, please
<rant>This thread has become an absolute mess with people spouting opinions on things they know nothing about and making judgement calls where none were asked for.
FWIW, we've got a few terabytes worth of image data archived and about 200Gb in process at any given time; we've got a database workflow that for whatever damn reason only likes smaller file sizes; we use clipping paths for probably 90% of the work; we print everything conceivable, from expensive calendars and artwork to free weekly supermarket flyers (I just spent 8 weeks prepping a very high quality and expensive catalog that'll be printed in 8 languages and have a run of close to 500,000), so, "quality" isn't always TOP priority, but it's always a main ingredient...
Anyway, I asked what I thought was a straight-forward question, i.e., why are the compression algorithms apparently different when saving as eps/jpg max vs. straight jpg with quality 12, and nobody seemed capable of just saying... "I don't know".
If anybody really cares, when saving as eps out of Photoshop, you can choose to ENCODE the image data INSIDE the postscript file (eps stands for Encapsulated PostScript) as JPEG. You save a LOT of storage space and more importantly in our case RIP TIME by doing this. Not every workflow can support the file format, though, so maybe it isn't option for all of you. And, yes, if you find yourself being forced to re-save the images a GOOD NUMBER of times you'll see degradation in the files. FWIW, this does not happen THAT often with us. Just once in a while we get a bone-headed customer who changes his/her mind a few times AFTER the final corrections have been confirmed. And in those RARE instances, we may re-scan or somebody like me will just have to spend some time in Photoshop making sure the image will look acceptable when printed...
</rant>
William Garabrant
Teamsatz u. Litho, GmbH
Neudrossenfeld, Germany
From: Héctor Antonio Roldán Catalán
Date: Sat, Jan 12, 2002, 12:40 AM
RE: [colortheory] cds, jpgs, pre press...
> This thread has become an absolute mess with people
> spouting opinions on things they know nothing about and
> making judgement calls where **none were asked for**.About this big thread, try to see the good side : )
Sometimes the threads get very big (as with ICC, profiles...) but always with good info. I would say: "Take the best and leave the rest..." If you already know something, please note that some people on the list may not know it. Iv'e seen multiple postings and "eternal ones" but always comes and end.
Yes, there have been answers to questions nobody asked, but filled with good info.
I try to keep quality on the prints, JPEG causes data loss so I try to avoid it. I've had problems with EPS encoded with JPG (When they get to the RIP) and I must say that only hapened with some RIPS. TIFFS are great but some RIPS just hate TIFFs with alpha channels . At the end I would say that in a perfect world it would be nice to have a file format that provide us small files with no data loss, but it seems that the gods are having troubles preparing this for us because we are still waiting.
I'm a photographer too and I started to create a big digital colection with all my work, I don't like JPGS but is not a big solution to burn cds with only 90 or 150 files. Ok, cds are cheap now but you cannot burn and burn just for fun. When you got a huge collection you'll see this on your side and on the client side you'll see that most of the people just hate to browse a big photo catalog to choose 17 photos and discover that just for bad luck, each photo is on a diferent cd!! I had to experiment to get the best quality with the best file size knowing that i was going to loose some quality.
I have worked on digital prepress receiving arts from the customers and is useful to know as much as you can get about each file format. Sometimes you have to experiment but sometimes there are folks that have walked on the same road and have interesting suggestions.
Too much JPEG info?? When I worked on digital prepress very often the clients asked the company to send somebody to give them a short course so they would prepare a better final digital art and I found that the user (yes, the person that sometimes sends you RGB files, TIFFs with alpha channels and even zipped files) where very open to my suggestions because I gave them *when needed* very big explanations. Instead, other companies used to send them somebody to say JPEGS just don't work with our rip and we are still finding out why.
Sometimes you need good answers to convince.
About the "I don't know" answers, I wouldn't like to be on a list that when somebody post a thread and you get 25 "I don't know" answers... Some people here answer more that you asked for. This is the way I've been learning a lot and I thank them all.
Most of what I would have say about JPGS have been said... :)
Héctor Antonio Roldán Catalán
Celular. (502) 815.4424
Tel.: (502) 254.04.15 / Telefax: (502) 254.02.48
ICQ: 69835949
From: Les De Moss
Date: Fri, Jan 11, 2002, 7:54 PM
RE: Re: [colortheory] Re: resaving jpeg - Goodnight, please
> This thread has become an absolute mess with people
> spouting opinions on things they know nothing about and
> making judgement calls where
**none were asked for**.With all due respect, if you seek only one opinion - and only the "correct one" at that, an open exchange provided by this group will likely disappoint. I believe most folks choosing to participate in this list do so with the understanding that once started, every thread becomes the property and interest of all who read/respond.
The confusion indicated in this thread is a bit of an education in itself about the vagaries of the subject... most of us benefit from the exchange, even when a forthright answer is, well , not forthcoming<g>.
Les De Moss
Ft. Collins, Co.
From: Ron Kelly
Date: Sat, Jan 12, 2002, 8:21 AM
RE: Re: [colortheory] cds, jpgs, pre press...
Dear Héctor:
I think you expressed yourself eloquently and I agree wholeheartedly that "excess information" is not a problem for me. I've learned a lot by lurking, which is what I do mostly on this list. I'm going to finish now before we get a missive from the massive about list etiquette and efficiency. How about them profiles?Cheers,
Ron KellyHéctor Antonio Roldán Catalán wrote:
> > This thread has become an absolute mess with people
> > spouting opinions on things they know nothing about and
> > making judgement calls where **none were asked for**.
>
> About this big thread, try to see the good side : )
From: Chris Murphy
Date: Sat, Jan 12, 2002, 8:20 AM
RE: Re: [colortheory] TIFF/EPS (was resaving jpeg)
William Alexander writes:>Hope this question isn't too basic, but why is Tiff better than EPS?
>(assuming that neither format is using jpg compression) I've always used EPS
>rather than TIFF but have no reason for it, and would switch in a heartbeat
>if there were advantages to using TIFF instead.TIFF is smaller, even with no compression. I'm pretty sure one advantage of EPS in a QuarkXpress workflow is that the previews are not added to the QuarkXpress document, whereas with TIFFs, the previews generated by QuarkXPress are saved in the QXP document. So for more complex layouts, the QXP document size can become much larger than if EPS files were used.
Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932
From: Gary Owens
Date: Sat, Jan 12, 2002, 1:03 PM
RE: [colortheory] resaving jpg
Hi group,Whether you use JPEG, TIFF, or whatever is like a lot of other things discussed on this list. You have to consider first your output then such factors as storage; repurposing; etc.
I use JPEG because the photolab can output either JPEG or TIFF and preffers the smallest file possible. I've been Jpging 24Mb files down to 11Mb which works great.
I then store my work file and the JPEG file (I consider that luxary to be able to store files that big). I had my storage cost down to 4 cents per meg with my MAG Optical machine; but cheap MAG disks are no more. Such is life in this fast paced tech world.
Gary Owens
From: Gary Owens
Date: Sat, Jan 12, 2002, 1:03 PM
RE: [colortheory] Re: resaving jpeg
> Specifically 5.02? Who's still on 5.02? Will 6.0.1 work? Why not?Hi,
This question should be rephrased to ask How many people are still using MAC II's with Photoshop 2.5 to do ther work. It would surprise many to find out how much antique hardware and software is still being used. And the output is just as good as if it came out of PS 6.0.
Gary Owens
Adobe Photoshop training classes are taught in the US by Sterling Ledet & Associates, Inc.