To reduce a TIFF file's size, turn on lossless LZW or ZIP compression when you save it, which re-encodes the same pixels more efficiently with zero quality loss. That is usually the single biggest win. The full answer to how to reduce file size of TIFF image runs in order of what it costs you: compress losslessly, flatten layers, cut bit depth or color, downsample the pixels, and convert to JPEG only when a TIFF is no longer required.
TIFF is built for uncompressed, layered, archival-quality image data, which is why it earns its own guide instead of getting folded into a general upload page. My guide to compressing images for email, the web, documents, and forms maps where files get rejected in general; this page is TIFF specifically, because the format's size problem and its fixes are unlike a JPEG's. Every size figure below is arithmetic you can redo yourself, not a number I am asking you to trust.
Why are TIFF files so large?
Because a TIFF stores the full, uncompressed pixel grid by default, and every extra bit of depth or color channel multiplies straight through the file size. The math is exact, not an estimate. An 8-bit RGB image carries three color channels at one byte each, so a 6,000 by 4,000 pixel photo, a common 24-megapixel frame, comes to 6,000 times 4,000 times 3 bytes: 72 million bytes, about 72 MB with nothing compressed. Save that same frame at 16-bit color and you double the bytes in every channel, 144 MB, because 16-bit is exactly twice the data of 8-bit.
None of that is a compression algorithm working badly. It is the format doing what it was designed to do, holding every bit of the original with nothing discarded, which is also why photographers and publishers keep archival masters as TIFFs. Adobe describes TIFF as a lossless format that stays larger than most precisely so it does not lose image quality, and notes it can hold anywhere from 1 bit per pixel for pure black and white up to 32 bits with an alpha channel (Adobe: TIFF files). That range is exactly why the same photo can weigh a few megabytes or over a hundred depending on how it was saved. Understand the size as raw pixel data first, and every method below is just a different way of storing fewer bytes or fewer pixels.
How to reduce file size of TIFF image, method by method
Work from lossless to lossy, and stop as soon as the file is small enough. Every method below either stores the same pixels more efficiently, at no quality cost, or throws pixels away permanently. These are the reliable ways to reduce TIFF file size, roughly in order of how little they cost you, so you can pick the cheapest lever that gets you under your target.
| Method | Quality loss | What it changes | Best for |
|---|---|---|---|
| LZW or ZIP compression | None | Re-encodes the same pixels more efficiently | Always turn on first |
| Flatten layers | None | Drops per-layer editing data | Files still carrying Photoshop layers |
| Reduce bit depth, 16 to 8-bit | None visible | Halves the data in every channel | Working files past their editing stage |
| Convert RGB to grayscale | Color only | Cuts three channels to one | Scans and documents |
| Downsample resolution | Permanent | Shrinks the pixel grid itself | Files larger than the destination needs |
| Convert to JPEG or PNG | Format dependent | Leaves the TIFF container behind | Sharing, web, anywhere TIFF is not required |
The first row is the one you should reach for every time. The last row is where most oversized TIFFs actually end up, because a file bound for email or a web page rarely needs to stay a TIFF at all. Understanding the difference between lossless and lossy compression is what tells you how far down this table you can safely go.
Turn on lossless compression first (LZW or ZIP)
Turn on LZW or ZIP compression in the Save As dialog before you touch anything else, because both are lossless: they pack the exact same pixel data into a smaller container, and the image that comes back out is bit for bit identical. Turning on compression is the first move any time you compress a TIFF file, and it is the only one that never costs you a thing.
TIFF LZW compression, the safe default
LZW has served as the de facto standard TIFF compression for years. It is technically a TIFF extension rather than part of the base format, but its patents expired in 2003 and 2004, so it is safe to use anywhere today (Wikipedia: TIFF). LZW works by finding repeated patterns and storing each one once, so it does most on images with large flat areas and least on noisy photographs. On a screenshot or a line drawing it can cut the file substantially; on a grainy photo the gain is smaller. Because it is lossless, you never have to weigh quality against it, which is why LZW is the default worth turning on before every other lever.
ZIP and Deflate, the newer lossless option
ZIP, sometimes labeled Deflate, is the other lossless choice, added to the TIFF specification in a 2002 supplement using the same Deflate standard that powers ordinary zip files. On most photographs LZW and ZIP land within a few percent of each other, and ZIP occasionally wins on smooth gradients. The one thing to watch is compatibility: TIFF's long list of optional features earned it the joking nickname "Thousands of Incompatible File Formats," so if you are unsure a recipient's software reads ZIP-compressed TIFFs, choose LZW, which nearly every TIFF viewer recognizes without complaint.
Does flattening layers actually shrink a TIFF?
Yes, and it costs nothing once you no longer need the layers. TIFF can hold layers and multiple embedded images the same way a native Photoshop file does, and each layer carries its own pixel data inside the file. A working file with a dozen adjustment layers, masks, and duplicated elements is storing all of that on top of the base image, and none of it survives into a flattened export anyway. Flatten before you save the final version, never the working file you still plan to edit. On a heavily layered file, flattening alone can cut the size by half or more before you apply any compression at all.
How much do bit depth and color mode matter?
More than most people expect, because the relationship is pure arithmetic rather than a diminishing-returns curve. Dropping a 16-bit image to 8-bit per channel exactly halves that channel's data, since 16-bit stores two bytes per pixel per channel and 8-bit stores one. Most output, printed or on a screen, cannot show the difference, because 16-bit exists for editing headroom, not final display, so the cut is usually invisible and the question of whether that compression reduces quality answers itself in your favor.
Converting from RGB to grayscale is a bigger cut for the right file: three channels become one, so an 8-bit grayscale scan is exactly a third the size of the same image in 8-bit RGB, and a sixth the size of a 16-bit RGB original. A scanned black-and-white contract gains nothing from three color channels it never used. The chart below takes one 24-megapixel photo and shows the uncompressed size at each setting. These figures are calculated straight from the pixel dimensions, not a compression measurement, which is exactly why you can redo the arithmetic yourself.
| Setting | Channels x bytes | Uncompressed size |
|---|---|---|
| 16-bit RGB | 6000 x 4000 x 3 x 2 | 144 MB |
| 8-bit RGB | 6000 x 4000 x 3 x 1 | 72 MB |
| 8-bit grayscale | 6000 x 4000 x 1 x 1 | 24 MB |
Once a TIFF is this far down and still too big, most workflows stop editing losslessly and ask whether the file needs to stay a TIFF at all. That usually means exporting a JPEG and compressing that JPEG below your target instead of fighting the container format for another few megabytes.
When should you downsample resolution instead?
When the file's real problem is pixel count, not encoding efficiency, and you are certain you will not need those pixels back. Downsampling is a different category from everything above it: LZW, ZIP, flattening, and bit-depth changes all store the same grid more efficiently, while downsampling shrinks the grid itself and is permanently lossy, the same class of move as reducing a JPEG's quality (Wikipedia: Lossy compression). A 600-DPI scan headed only for a web page or an email is usually holding far more pixels than either destination can display, and cutting the resolution first often beats every lossless trick combined. Do this last among the reversible options, and only once you know the smaller size is genuinely enough, because you cannot get the discarded pixels back.
Is JPEG-in-TIFF compression safe to use?
No. TIFF offers an internal JPEG compression mode, and it is lossy in exactly the way a normal JPEG is, simplifying the image data and discarding detail to shrink the file. The older version of the feature, old-style JPEG-in-TIFF, is considered obsolete and should never be written by modern software, so if a viewer offers it, avoid it. A TIFF's whole reason to exist is the lossless option sitting right beside this one, so reaching for the lossy path usually means you no longer need a TIFF at all. If you want JPEG-level compression, export a real JPEG file rather than a TIFF wearing one.
When should you convert a TIFF to a lighter format?
When the destination does not require TIFF and you have run out of lossless room. TIFF is a poor fit for the web: most browsers other than Safari do not display it natively, and a heavy TIFF slows a page the same way any oversized asset does (MDN: choosing an image format). For sharing or publishing, export to JPEG when some quality loss is acceptable for a much smaller file, or PNG when you need to stay lossless and the image is not photographic. The fastest way to shrink TIFF image size for the web is often not to keep it a TIFF at all.
One honest limit: our browser tool does not accept TIFF as an input format. It works with JPG, PNG, WebP, and AVIF, so the TIFF-to-JPEG export has to happen first in Photoshop or whatever editor made the file. Once you have that JPEG, compress it rather than guessing at an export quality setting by eye, because you get a live before-and-after size as you move the slider instead of re-exporting every time you want to check.
What about CCITT Group 4 for black-and-white scans?
For pure black-and-white scans, CCITT Group 4 is the most efficient TIFF compression there is, and it is the one method the AI Overview names that none of the ranking pages actually explain. Group 4, also called T.6 fax encoding, is a lossless scheme built specifically for 1-bit images, black or white with nothing in between, and on a text document it can shrink a scan to a fraction of what LZW manages. The catch is the 1-bit requirement: it only works once the image is genuinely bitonal, so a grayscale or color scan has to be reduced to black and white first. That is fine for a signed contract or a form and wrong for a photograph. If your large TIFFs are scanned paperwork, converting them to bitonal and saving with Group 4 is usually the single biggest win available, more than LZW or ZIP on the same file.
Step by step: compress a TIFF in Photoshop
In Photoshop, flatten any layers you no longer need, then choose File, Save As, pick TIFF, and select LZW or ZIP in the compression section of the TIFF Options dialog (Adobe Community: reduce the size of TIFF files).
- If the file has layers you no longer need, choose Layer, then Flatten Image first, since flattening is a separate step before the save dialog opens.
- To cut bit depth, go to Image, then Mode, and pick 8 Bits/Channel; to drop color entirely, choose Grayscale in the same menu.
- To downsample, open Image, then Image Size, and lower the pixel dimensions to what the destination actually needs.
- Choose File, then Save As, and select TIFF as the format.
- In the TIFF Options dialog, pick LZW or ZIP under Image Compression. Both are lossless, so either is a safe default.
- Save, then compare the new size against the original in Finder or Explorer to confirm the drop.
When should you leave a TIFF alone?
When it is an archival master or a working file you are still editing. The whole point of TIFF is holding image data losslessly so it can be reopened, re-edited, and re-saved indefinitely without the generational quality loss a lossy format accumulates. Flattening a file you still need layers in, or downsampling a scan before you know its final destination, trades away flexibility you cannot recover for a size cut you may not have needed. Keep the master TIFF as it is, and generate a smaller lossy export, JPEG or otherwise, only for the specific destination that needs one.