| <<<Back 1 day (to 2017/09/10) | 20170911 |
celyr | Hi | 13:22.20 |
ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 13:22.20 |
celyr | any idea why screen is giving me a large output than ebook ? | 13:22.36 |
kens | No | 13:22.50 |
| Nobody can possibly answer such a question without seeing the original file | 13:23.04 |
| Why do you think screen should be smaller ? | 13:23.13 |
celyr | this is the command: gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dPDFSETTINGS=/ebook -dNOPAUSE -dQUIET -dBATCH -sOutputFile=resize.pdf file.pdf | 13:23.14 |
| well I would think it would have less dpi ? | 13:23.37 |
| The pdf file is just 3 images | 13:23.44 |
| sorry four images, one per page. | 13:23.53 |
kens | And by 'image' do you mean a bitmap ? | 13:24.06 |
celyr | Those images got acquired by a scanner | 13:24.08 |
kens | Like I said, nobody really can answer your question without seein ghte files. | 13:24.29 |
celyr | kens, I have no idea on how to check if it's a bitmap | 13:24.35 |
kens | If you want to see the various settings, they are documetned in vectordevices.htm | 13:24.42 |
celyr | I can send you the file, if needed, it's confidental anyway | 13:24.44 |
kens | Well, I'd prefer not to | 13:24.55 |
| You should review the documentation | 13:25.01 |
| If the images are gray, then the downsampling filter for screen is average, for ebook its bicubic | 13:26.31 |
| And 72 dpi for screen vs 150 for ebook | 13:26.44 |
celyr | No images are colored :) | 13:26.47 |
kens | It could just be that it compresses better | 13:26.51 |
celyr | I think that screen is just failing | 13:27.06 |
| since it's actually larger than the input | 13:27.15 |
kens | THat too is possible | 13:27.25 |
celyr | (24mb input, 24,9 screen, 8,8 ebook) | 13:27.30 |
kens | Depending on the input | 13:27.31 |
| You also haven't said which version of Ghostscript you are using, where you got it from or what platofrm you are running on | 13:28.41 |
celyr | $ gs --version - | 13:29.21 |
| 9.20 | 13:29.21 |
| running on debian stretch | 13:29.21 |
kens | Well the current version is 9.21 and 9,22 is due for release shortly | 13:29.40 |
| So it may be somethign we've fixed in the last 12 months | 13:29.49 |
| Its also entirel;y possible that Debian packaging has done something | 13:30.03 |
| I'd build from source and try that myself | 13:30.11 |
| and as I keep telling people; don't use -dPDFSETTINGS | 13:30.25 |
celyr | heh | 13:30.41 |
| it's fast and I need the job done | 13:30.46 |
kens | Choose what you want to change, select the values that you want, and alter each one individually | 13:30.48 |
| clyr, but you're complaining it doesn't 'get the job done', or so it seems to me | 13:31.08 |
celyr | I'm reading the documentation atm :) | 13:31.23 |
chrisl | There's also no claim that pdfwrite's output will be smaller than the input | 13:31.39 |
kens | The bit you want is the big table of settings | 13:31.41 |
| Yes, if the input is an indexed cmyk, and you create RGB form it, it won't be Indexed any more and will be bigger, for example | 13:32.15 |
celyr | hmm | 13:33.09 |
kens | There are other cases | 13:33.21 |
| Also an Indexed RGB which is downsampled and uses bicubic filtering will use colours that were not in the original, so the Indexing will be lost agai | 13:34.02 |
celyr | all this doesn't explain why ebook is working better than screen | 13:34.28 |
| Or it's working while screen it's not | 13:34.39 |
kens | I'd bet screen is 'working', because we've had enough complaints from people using it | 13:35.04 |
| THe answer is that the two are different | 13:35.16 |
celyr | :D | 13:35.24 |
kens | Also, as I said earlier, you are using an old version. | 13:35.34 |
| I know for sure there have been changes in this area over the last 12 months | 13:35.45 |
| (because I made them) | 13:35.55 |
celyr | Okay I'll check if with next version this will still happen | 13:36.46 |
| Forward 1 day (to 2017/09/12)>>> | |