Or sometimes also PDF? Here is where you say that every journal you’ve looked at supports more, but in my experience these two or three are what are wanted by every journal I’ve looked at submitting to.
EPS and PDF can be vector format, but they don’t specify that your figures should not be raster. As far as I know TIFF doesn’t have anything special about it. I understand why they don’t want JPEG, but why not e.g. PNG (all I found is this)? TIFF and EPS are rather dated formats; PDF is at least somewhat current.
Back when publishers started to embrace electronic submissions, those were the only formats that were widely used for academic papers, and had good enough image quality for publication.
TIFF is uncompressed and EPS will work with legacy applications (and also the vector stuff).
They are dated but in some respects better. "Dated"in this respect seems to mean not compressed. They tend to be large files but because of that they are not “lossy” formats and preserve the original data. Something that I could see being important for academic purposes.
PNG files can’t be saved in CMYK mode (used for offset printing and safer for quality inkjet printing.) There are certainly work arounds, but they involve extra steps.
To some degree, asking for .eps and TIFF is an attempt to get people preparing submissions to consult someone who knows about graphics, and who has the software to export those formats successfully. There’s no format that guarantees good reproduction, but in practice requesting certain formats will increase the average quality of submissions.
Similarly, asking for photos saved at 300ppi doesn’t guarantee a good image, and if someone just converts a lo-rez image to 300ppi it doesn’t help the image at all. Asking for 300ppi photos is just a way to improve overall submission quality without writing out a paragraph of detailed specifications.
If a publisher hopes for vector art submissions they are more likely to get pure vector art if they ask for .eps than if they request a PDF. Ideally they should, again, provide exact specification, but often space, time and design conspire against that.
Beat me to it. While quality illustrations of both kinds can be submitted in almost any format, forcing contributors through the narrow gate of these two pro-only formats means they won’t get shaky phone-cam photos, BMP exports from Excel and web graphics.
Also academic resistance to change - the studies on this “jay-peg” thing aren’t complete yet.
The .tif image format is recognized as the standard image archive format by the National Archives. Perhaps journals want to preserve their electronic data in the standard archive format.
That’s almost certainly a driving factor, but probably more from parallel lines of reasoning and need than a hierarchical decision. TIFFis standardized, flexible, has no licensing issues and most importantly, non-lossy.
Its biggest drawbacks are that it’s a bulkier file format than even some other non-lossy formats, and has a 4GB size limit.
Since PDF is the evolved form of EPS, I suspect it will replace EPS and possibly TIFF as a standalone format in the future. Right now, though, its standards are implemented in fairly slipshod fashion and it’s hard to distinguish one created to archival standards from one pinched off by low-end software.
TIFF can certainly be compressed. And they’d better take my compression (13 mb > 300k, in my experience they have crappy upload routines as well). It’s simple line art figures; it doesn’t need to be rich.
Good point, although I’ll note that this is also true for online-only journals.
TIFF is an image wrapper system. The actual format of the image can be of several different types, only a few of which are popular. Technically you can invent your own encoding.
But the image can be compressed (run length encoding was popular in the very earliest days of the format), lossy, whatever. You can have TIFFs encoded with the JPEG compression system and on and on.
This catch-all variety of encodings makes TIFF a lousy standard, unless you carefully specify which encodings are accepted.
EPS is an extended form of PostScript, which contains a remarkably sophisticated programming language. It was common for images like GIFs to be encoded as a hex stream with a small program in PS to decode the stream. There is no reason to suppose that other lossy formats could not also be programed. However memory (in particular stack space) was limited and you could overflow the device’s engine and crash it.
One thing both formats have in common, if you specify things right, is that they both scale well. I.e., the author might be looking at it at 300-600 dpi while the final printing might be 1200-2400 dpi and you don’t get raster effects. But, again, you have to use the right encoding.
I know next to nothing about the technical aspects of the different formats, but I do know that at least in my field, some journals are beginning to routinely run algorithms on images of raw data (blot images, micrographs, that sort of thing) to look for indications that the images have been manipulated to misrepresent or mislead. That may play into this as well. If nothing else, I’d imagine it’s easier to write these algorithms to examine a limited number of formats.
