A pdf can hide fat photos inside. You look at two pages on the screen and think this should be 400 kb. The file is 14 mb. The pages are photos of paper, not typed text. Each photo is a beast. The pdf is only a wrapper. Wrapping a beast does not make it small. I learned this after “compress” on a random site gave me 13 mb back and a lecture in my own head.
FileOnDesk is not Adobe. Browser cannot crush a scan like Acrobat. We do not fully compress pdf like Adobe. I will keep saying it because the search is “compress pdf” and the honest answer here is a different path. Photos first. Compress Image. Then JPG to PDF. New pdf from skinny photos. That is the diet that works in a tab.
A pdf can hide fat photos inside
Typed pdf from Word is often small. A 10 page letter, real text, maybe 200 kb. Scan pdf from HP is a stack of pictures. 10 pictures at 600 dpi colour is a brick. If you merge two bricks, you get a bigger brick. Merge does not diet. Split does not diet. Rotate does not diet. They shuffle pages. The fat stays.
People drop a scan into anything named compress and hope. Some online tools send the file to a server and run Ghostscript or similar. Sometimes they get small. Sometimes they get a blur. Sometimes they get a privacy problem they did not think about. Our files stay in the browser. The trade is, we cannot run a heavy Acrobat-style job on your 28 mb scan. RAM, time, honesty.
I check size with File Size Checker before I waste hope. 14 mb, two pages, I already know there are photos inside. I do not need a detective. If it is 300 kb and still “too big” for a portal that wants 200 kb, then you are in a different fight, maybe they want a image not a pdf, or they want 100 kb. Then compress the image more, or crop.
Screenshots in a pdf. Same thing. A WhatsApp screenshot is a image. Four screenshots in one pdf from a phone share can be several mb. Looks like text. Is not text. Compress each screenshot as image, then jpg to pdf.
Duplicate pages. Scan both sides and also a extra blank. Blank is still a full page photo of white. White is not free if it is a picture of white. Split out the blank if you can, but the remaining pages stay fat until you remake them from compressed jpgs.
Browser tools are not Adobe, we say that on purpose
Adobe Reduce File Size, or PDF Optimizer, can downsample images, change quality, flatten. A person who knows the menus can take 28 mb to 2 mb and still read the stamp. That person has a licence and a PC that is not a 2014 office box with three toolbars. Many of us do not.
So the workflow on this site is not “press compress pdf and pray”. It is:
Take photos of the pages, or export pages as images if you already have a scan and a way to screenshot each page. Screenshot of a scan is ugly but it works in a pinch. Better to photo the paper.
Run each photo through Compress Image. Slider. Preview. If the stamp dies, raise quality. If the file is still huge, crop extra table, or lower a bit more.
Drop the compressed jpgs into JPG to PDF. Order them. Download.
Check the new pdf size. If it is still big, one of the jpgs was still big. Find which page. Compress that one again. Replace. Make pdf again. Do not hunt for a pdf compress button that we did not promise.
If someone labelled a tool compress pdf on another website, read whether it uploads. Read whether they keep the file. I cannot police them. I can tell you what we do not do.
20 mb limit. If the original is 40 mb, this tab will not swallow it. Photo the paper. You do not need the 40 mb file to make a sendable copy.
Quality. A compressed jpg at 70 looks fine for a typed letter. A licence with a small stamp may need 80. A selfie for a form can go lower. Different article. Same slider.
I had a contractor send a “compressed” pdf that was 21 mb. He had printed to pdf from a Word that had full size photos of a site visit, then he zipped it. Zip of a pdf is not how Gmail works. You attach the pdf, not the zip, because the other office does not unzip on a phone. Compress the photos in Word before you make the pdf, or pull the photos out, compress here, build pdf here.
Photos first, then jpg to pdf
This is the part I write on a sticky note near the HP.
Do not scan at 600 colour for a letter. If you must use HP, grey 300. If the default is locked, walk away, use the phone.
Phone photos, straight, daylight.
Compress image, each.
Jpg to pdf, one file.
Size checker, confirm.
Send.
If you already only have the fat pdf and the paper is gone, you are stuck with screenshots of each page or a real Acrobat. Screenshot each page on a laptop, save, compress those screenshots, jpg to pdf. The quality will drop. For a emergency email it is better than not sending. For a court bundle, go find the paper.
Merge after this workflow, not before. Merge two fat files and you doubled the fat. Merge two light files and you still have a light file. Merge PDF at the end if you have two light pdfs to join.
I will not write that we almost compress like Adobe. We do not. The slider is on the photo. The pdf is a stack of those photos. That is why a “compressed pdf” from a wrong method still looks big. The photos inside never went on a diet. Put them on a diet. Then wrap them again.
One more office story. A tender pdf came back from the other party “already compressed”. It was 19 mb. Two pages of typed terms, rest was a colour scan of a cancelled cheque at full camera size. I pulled that cheque photo out by photographing the screen like a uncivilised person, ran Compress Image to around 70, built JPG to PDF for that page, merged with the typed pages. Final was under 2 mb. Cheque number still readable. That is the honest compress. Not a magic pdf button.