I am using Impress to cover personal details on a scanned document. I use rectangles to cover the details. I then select-all, right-click on one of the objects, and select Convert → “To Bitmap”. It creates a single image, but it is very blurry. Everything that I found online about this problem refers to creating PDFs from Draw. Is there anyway to specify a higher resolution output for this conversion within Impress?
Select all then click File - Export, choose PNG to avoid blurring, tick the box Selection if you don’t want full page. Save with suitable name. A dialogue box will appear, choose resolution first then adjust the size for width back to original width (it changes with resolution change) and OK.
Further size reduction to the png without visible loss of quality can be made on Gimp with Image - Mode - Indexed , ok the dialogue and export to png
Note that if exporting to pdf from Writer you can Redact words that will be entirely replaced with black rectangles in the pdf
1 Like
Thank you, EarnestAI! It worked!
It did cause the image to increase from about 100KB to 400KB, but that’s with about a dozen black rectangles.
I had to first obtain the resolution of the original unredacted image using the linux/cygwin tool:
identify -verbose Original_image.png
I don’t use Gimp, but I’ve tried in the past. It’s very different from the Windows based tools that I’m used to.
As for redacting entire words at a time in Writer, since it is a scanned image, would that still be the right tool?
P.S. This forum seems to strip away my spacing between paragraphs. It makes it hard to read.
PNG is clever enough not to include unused colours but very similar colours each get their own value. Going to indexed will probably reduce size a bit. I avoid export to GIF (indexed) because I haven’t always liked the results but you could try it. I find that 150 dpi is quite readable and printable. For a standard home printer going above 225 dpi is probably unnecessary.
If you right-click on image and select Compress, you should see existing size and resolution in Apparent dimensions
I haven’t used Redact in work but did try it out when it was introduced. If called by Tools > Redact it will open the image in Draw and final result might end up at lower resolution than expected.
The original image is black and white. I can see this from the “identify” command (though I admit that much of the output is gibberish to me). The histogram data confirms this. And the Colorspace is Gray.
The redacted image uses 3 color planes and an alpha plan for transparency. The Colorspace is sRGB. The histogram data shows many levels. I find this odd, since the boxes are black (fill and border). Perhaps there are grays due to antialiasing. I used Windows image viewer to zoom in and I see grays at the black-and-white transitions, but I’m not sure if this is an artifact of the image viewer’s antialiasing for display or whether it is part of the PNG image itself. Regardless, antialiasing might explain the presence of values other than the maximum and minimum that are representable, but not the presence of multiple pixel planes.
The dots per cm is 118.11 (300dpi) to reflect the original image resolution, which is needed due to the small size text that was scanned.
The redaction in Draw doesn’t seem to be an option. I followed the description at page Redaction Feature in LibreOffice - Phase 1: Manual Redaction . In the resulting PDF, however, the rectangles are transparent gray, in contrast to what is shown in the above page. That is, the redacted details are still visible. The resolution is also too low, making it difficult to see any text, redacted or not.
I think that the only option is to accept the (3+)x file size increase.
Thanks. Despite its nonideal characteristics, it works!
P.S. Since this site removes inter-paragraph spacing, I tried indenting each new paragraph, but the text block gets interpretted as a code block
Black and white text and images when printed get fuzzy (grey) edges, scanning adds more fuzziness. If the image is greyscale then no size change can be expected going to indexed. You can reduce to B&W (1 bit) again but best done in graphics editor watching preview as you adjust the levels slider to get best balance between black and white. Sometimes, it’s not possible to get a good enough balance on scans.
You have to use the pdf redacted export in the redact toolbar, either white or black redaction. I don’t see a setting for resolution and I think it ends up at 96dpi which is really too low for comfortable reading
BTW this site use Markdown, a single return is same as a line break, two returns make a paragraph. For more
https://wiki.documentfoundation.org/Ask
Sorry, I don’t think I was clear. The PNG file that was inserted into the Impress slide is black and white. I set the scanner parameters to generate black and white, not grayscale. However, ImageMagick’s “identify” command shows the Colorspace used was Gray, and there are only two intensity levels: 0 and 255. There is no blurring or gray.
After adding the black boxes and exporting to PNG from Impress, that PNG file uses sRGB, which apparently is 3 planes plus a 4th alpha plane, which I recall is for transparency. Since the fill and border of the boxes are black, I’m not sure why this scheme was used, but it explains the 3-4x increase in file size. Furthermore, there are values in the planes that are not 0 or 255, so somewhere in the export process, there may have been antialiasing.
I’ll let go of Draw’s redaction for now, since the resolution is so low.
Regarding markdown, I tried double newlines without success. I will try again with this response message.
EarnestAI pointed that out, but the resolution is inadequate.
tdf#124377
@mikekaganski: I took a look at the example PDF at the very end if the page that you cited, but what I found in my use of the redaction feature was that it was too coarse for the small text in the image that I was redacting. I needed 300 dpi, which was the resolution of the image that I inserted into the slide.