I'm tired of posting screenshots and all that anymore, but here's a(nother) bug that came up lately in Studio:

After converting an editable PDF in Studio, some or all the images turn to black squares or some weird graphical artifacts (looking like a damaged scanner or something). The funny part is the problem is solved after restarting Studio (NOT Windows) - but not always 100%. Recently I had a situation where a file was messed up completely, I then reset Studio, and after that only three or so images were corrupted, with the rest being ok.

I've seen stuff like that exactly a year ago already (see the OP), but now it seems to have reoccurred without a reason. I run the horribly unpredictable and buggy Win 10 on Surface (it has a special build), but this smells like SDL might have some say here nonetheless. I'm not sure what this might be related to. Any suggestions or people with the same problem?

I did activate IRIS recently, but honestly it doesn't seem to have made any difference in conversion whatsoever in terms of reconstruction of formatting, handling different document objects and such. I'm not sure it even works at all. I'll try to disable it and see then, but this exposes me to the risk of a certain DeepL pre-translation bug coming up again…

Paul, I'll gladly provide you with a test file (which may pretty much be any PDF with images of my choice), but we both know it's going to work perfectly on your end - for some reason.

I'm tired of posting screenshots and all that anymore, but here's a(nother) bug that came up lately in Studio:

After converting an editable PDF in Studio, some or all the images turn to black squares or some weird graphical artifacts (looking like a damaged scanner or something). The funny part is the problem is solved after restarting Studio (NOT Windows) - but not always 100%. Recently I had a situation where a file was messed up completely, I then reset Studio, and after that only three or so images were corrupted, with the rest being ok.

I've seen stuff like that exactly a year ago already (see the OP), but now it seems to have reoccurred without a reason. I run the horribly unpredictable and buggy Win 10 on Surface (it has a special build), but this smells like SDL might have some say here nonetheless. I'm not sure what this might be related to. Any suggestions or people with the same problem?

I did activate IRIS recently, but honestly it doesn't seem to have made any difference in conversion whatsoever in terms of reconstruction of formatting, handling different document objects and such. I'm not sure it even works at all. I'll try to disable it and see then, but this exposes me to the risk of a certain DeepL pre-translation bug coming up again…

Paul, I'll gladly provide you with a test file (which may pretty much be any PDF with images of my choice), but we both know it's going to work perfectly on your end - for some reason.

Hi Adrian,

you'll get the best results in "Soliding" PDF created by MS Word. If you use Acrobat or Adobe Reader then use CTRL+D to open the properties window. Check the first tab to see the creator (software) of the PDF. As Paul said - the source (not file extension but a source of the PDF file) is important for later processes. If it will be MS Word of Office then you can expect quite good results. That doesn't mean the results will be perfect, without tabs or multiple spaces between words or ghost tags... but the best possible using Solid.

Keep in mind that Solid won't provide good results with PDFs from InDesign or Framemaker. PDFs from TeX/LaTeX, Quicksilver, and other generators/systems are always unexpectable - the test drive is needed in every case.

You will receive bad results using Solid to convert PDFs from Corel, Quark (depends on layout and Quark version) and software like Photoshop and Illustrator (and so on...).

The perfect software which can handle all type of PDFs is ABBYY Fine Reader but you have to know how to operate the software and adapt the workflow to get these high-quality results. It's not a simple 3-step converter but very complexed software with many options and possibilities which user has to know and be able to use where/when needed.

Otherwise - if we've had any good auto-converter - the DTP specialists in LSPs would be not necessary, right? ;)

"it often attempts to "scan" things instead of just converting them, from my experience" - starting from version 14 it is not scanning instead of converting. As far as I know, if the PDF is editable it reads the content like you'd copy it from the PDF (unless there are some issues with special characters) - that depends on the app settings. We're back to the same thing - user must know what he/it (I mean both - user and the source file) needs and how to configure the software's settings for a specific project/case.

Anyway, without the screenshots or file, it's impossible to tell more as every project is different and needs an individual approach.

Regards,

Rafal

Thanks Rafał, this was elaborate (and, admittedly, not much I didn't know already) - but the point here is that the problem can be remedied by resetting Studio, which obviously points to another filthy bug ignored/not hunted properly by SDL. So it's not a performance issue, but a stability/reliability one.

And converting things manually (which is the only sensible way indeed) with Finereader takes ages, as we all know, so it's certainly not a good option if you value your time and money, unless you happen to have a slave to do the work for you.

Hehe, you can always hire a freelance slave for approx. 1 EUR per A4 page with 10 pages/h productivity.