People tell me I have to convert the drawings, or at least the text, to vector graphics.

No, no, no.

First, TIFF is a raster image format. It contains no text. It only contains pixels. So when you say your TIFF image "contains text", all you are really saying is that it has some arrangements of pixels that look like letters.

Second, vector graphics are just shapes defined mathematically. Yes, the character shapes contained in (most) fonts are vector graphics. But that's not all there is to making them text. So when one says that "text is vector", all one is really saying is that fonts contain vector paths that look like letters. But just having vector paths that are shaped like letters is not the same as having editable text.

Actual text is an object into which you can type using your keyboard to cause your operating system to retrieve and display from the font file the particular vector path that is shaped like the letter that corresponds to the key you press. So there's much more to an actual text object than just the image on your screen that looks like the characters.

There are optical character recognition (OCR) programs which try to programmatically "recognize" the arrangements of adjacent pixels in a raster image as text characters, and then create actual text objects containing those characters. OCR is far from perfect and is dependent upon good-quality raster images. Illustrator does not contain any OCR feature.

Whomever is telling you otherwise is confusing autotracing with OCR. Autotracing is another kind of program that merely tries to draw vector paths around areas of same-colored pixels. So although you can use an autotracing program to create vector paths that are similar in shape to the clumps of pixels that you are calling "text" in your image, the result will no more be actual text that you can edit than is the raster bitmap you started with.

Acrobat Pro, on the other hand, does have a built-in OCR feature. And, Acrobat Pro can also (at least try to) export a PDF as a Word file. So you may have some success with that. But don't expect miracles. How successful you are in this attempt highly depends on the quality of the original raster image. The resulting Word document is often not at all built the way (number and arrangement of objects) you would actually build it from scratch in Word. So how usefully-editable the resulting Word document is, is dependent upon the details of the specific case.

JET

You might do better with wmf or emf as your format and if you are not resizing the art once in Word then .png will give surprisingly good result.

I do not believ word supports cmyk and if you use an eps then the cmyk I believe will be converted by word to rgb. I think. So it might not be a good idea in the first place.

...the funny thing happens, when I use "Print" to PDFCreator...the .pdf I get has clear text!! ...What do you think of this?

I'd say PDF Creator ran OCR on the image, as explained in my first post. You're going all around your fingers to get to your thumb. You keyed text in Illustrator, then rasterized the text, then ran OCR on it to "convert" it back to text again.

And you're still going to end up back where you were, if you insist on publishing in Word. Word cannot import a PDF as a spot graphic.

I also tried to "Save to to Microsoft Office" as .png, but it didn´t look good.

Of course. Because now you've rasterized your text again--PNG is a raster image format.

If you insist on publishing in Word, and if you're going to erase the text in the raster image and then re-key the deleted text somewhere anyway, just import the raster image (as a TIFF, after erasing the text) into Word, and use Word's text tools to create the text.

You're doing a bunch of convoluted format translations, involving multiple programs, confusing yourself over color mode, all in a futile attempt to overcome what you cannot overcome: Word (and similar programs) is just not built for this. That's why the silly Save For Office command in Illustrator is what it is; just a repackaged PNG export.

When it comes to handling graphics, programs like Word really can't do much more than just import a simple raster image. Yes, they support the operating system metaformats (formats that can contain combinations of raster, vector, and text objects) like WMF and EMF (and somone has already suggested that you try that). But those are limited and users who don't understand the differences (i.e., the target users who would want a dedicated "Save For Office" command) stumble over them. So when Adobe built the Save For Office feature, they took the lowest-common-denominator route; they just make it export a versatile raster image format, PNG.

If you want best-practice advice, you really ought to just post a typical example of what you're working with, stating where the original image is coming from, and any other specifics of your workflow, and what your intended deployment is. (Just saying you're making an e-book isn't really telling anyone anything useful.) That way, others can see how much text is involved, how many separate text objects are called for and their arrangement, etc.

JET

I take it you mean inserting textboxes over the tiff?

Yes. Word can import a common raster image format (like TIFF or PNG). Word can create free-floating text objects. (It's extremely cumbersome, but it can do it.) As I recall, Word can "group" combinations of raster images and individual text frames in a kind of box (I forget what it's called) that can be dragged around the page or flowed in place with surrounding body text.

I read somewhere that Word often will not display an .eps properly, but will "communicate" the contained information to a Postscript printing device.

Word cannot understand or edit PostScript. It's not a PostScript program. As already explained, EPS is a format that was designed specifically to provide a "container" that can simply be imported into non-PostScript programs. When a non-PostScript program imports an EPS, it just displays the low-resolution raster preview (if any) that is contained in the EPS file's header. It otherwise simply ignores the real content of the EPS. When the file is printed to a PostScript device, the printer receives the content of the EPS in the flow of surrounding content and interprets it for printing.

PDFCreator is based on Ghostscript. Ghostscript is an emulated PostScript interpreter. So PDFCreator can interpret the content inside the the EPS container, and the resulting PDF displays the actual content, rather than merely the embedded low-resolution preview.

But again, you're still back to the same old situation if your intent is to publish from Word. Word cannot import the PDF as a free-floating graphic. Adobe Acrobat Pro can try to export the content of a PDF as individual objects that Word can understand, in Word's file format. I don't know if PDFCreator can do that.

I would not understand why the same software PDFCreator did not run OCR on all the other pdf´s...

I don't know either. Perhaps you had certain options turned on/off, or the "text" part of the original image (again, it's not text; it's just pixels) was not of sufficient resolution to be recognized successfully by the OCR routine.

JET

I don´t want to publish from Word, I want to convert my end-docx to a high quality pdf...

So finally, you get around to stating what your intended final output is to be. But you said you "have a layout for an ebook with some important black and white bitmap illustrations". So that implies that the page layouts do not consist of only the annotated illustration. So what program are you using to assemble the whole pages of the "book"?

Again: If you will just fully describe your project somone can probably recommend a workflow to minimize the convoluted steps you have been messing around with. It sounds like you are making this way too hard.

JET