What TIFF actually is
The Tagged Image File Format showed up in 1986, born out of Aldus Corporation (later absorbed into Adobe) to give desktop publishing a single container that could hold scanner output, halftone separations, and full-resolution photographs. Almost forty years later, the format is still everywhere — not because anyone loves it, but because it solved a problem that has never gone away: how to store a high-fidelity image without losing a single bit.
A TIFF file can be uncompressed, losslessly compressed (LZW, ZIP/Deflate, PackBits), or lossily compressed (JPEG-in-TIFF, though rarely used). It can hold 1-bit (black-and-white line art), 8-bit grayscale, 16-bit grayscale, 8-bit RGB, 16-bit RGB, 32-bit floating-point, CMYK, Lab, and a dozen other color models. It can hold multiple images stacked as pages. It can hold alpha channels, spot color channels, ICC profiles, and arbitrary application-specific tags.
That flexibility is exactly why the format refuses to die. Every scanner, every print shop, every GIS pipeline, every legal archivist still depends on it. And every one of those TIFF files is, on average, enormous.
Where TIFF still lives in 2026
Most consumer workflows will never produce a TIFF. The format survives in five niches where its specific properties are irreplaceable:
- Scanners. Almost every desktop and production scanner defaults to TIFF for its uncompressed output mode. A 300 DPI letter-size page scanned as 24-bit RGB TIFF is about 25 MB. A book scanner doing 600 DPI doubles that.
- Fax machines. Yes, fax is still alive — in law firms, hospitals, banking, and any industry that built compliance processes around "the fax was sent on Tuesday at 9:42 AM." Fax transmissions encode each page as a 1-bit (black-and-white) TIFF using Group 3 or Group 4 compression. The receiving machine typically saves as TIFF.
- Print shops. Offset printing still wants CMYK TIFFs at 300 DPI, sometimes 450 DPI for high-end work. A magazine cover at 300 DPI in CMYK is around 40 MB; a billboard can run into gigabytes.
- Legal and medical document archives. Courts, hospitals, and government records often require a lossless, unalterable master copy of scanned documents. TIFF — particularly uncompressed or LZW-compressed — is the de facto standard. PACS systems for medical imaging store DICOM, but the secondary "easy view" copy is often TIFF.
- GIS and satellite imagery. GeoTIFF, a TIFF variant with embedded georeference tags, is the working format for everything from Landsat scenes to municipal parcel maps. A single satellite tile can be 1–5 GB.
In other words: if you have a TIFF in your hand, it almost certainly came from a scanner, a fax, a print shop, an archive, or a map. None of those are situations where you want a 50 MB file attached to an email.
Why TIFF files are so huge
Three reasons compound:
- Lossless by default. A 24-bit RGB TIFF stores every pixel's three byte values with no discarded data. A 4000×3000 scan is exactly 4000 × 3000 × 3 = 36,000,000 bytes, plus a few hundred bytes of header. LZW compression might shrink that to 22 MB; ZIP might get it to 18 MB. Either way, you're keeping every bit.
- High bit depth. Photography TIFFs are often 16-bit per channel (48 bits per pixel) rather than the standard 8-bit. That doubles the file size again — the same 4000×3000 scan as 48-bit TIFF is 72 MB uncompressed.
- Multiple pages. A TIFF can stack images as pages in a single container. A 50-page legal contract scanned at 300 DPI becomes a single 1.2 GB TIFF. It's handy for transmission but a nightmare for everything else.
By contrast, the same 4000×3000 RGB image as a JPEG at quality 85 is around 2 MB. That is the entire appeal of converting TIFF to JPG: you trade lossless, high-bit-depth, multi-page fidelity for a file 20–50× smaller that looks identical on a screen.
The multi-page problem
Here's the first thing that trips people up: a TIFF can hold fifty images. A JPG holds one.
When you convert a multi-page TIFF, the converter has to make a decision. Three common behaviors:
- Split into N JPGs, one per page. This is the right behavior 95% of the time. A 50-page contract becomes
contract_01.jpgthroughcontract_50.jpg. Each file is independent, viewable in any browser, attachable to any email. - Export only the first page. Cheap converters do this silently. You get one JPG and lose 49 pages of content. Always verify by checking the page count of your output.
- Stack into one image. Almost never what you want, but some tools will tile pages vertically into a single tall JPG. This is a lossy, awkward layout that breaks page boundaries.
A serious TIFF-to-JPG converter tells you upfront how many pages the input has and offers to split them. The SynoxTools converter reports page count on upload and produces one JPG per page, named with a page index. For a 50-page TIFF you get 50 JPGs in a single ZIP — no surprises.
The bit-depth trade-off
TIFF commonly stores 16 bits per channel. JPG stores exactly 8 bits per channel. There is no "16-bit JPG" — the format specification caps at 8.
This is rarely a problem in practice. Here's why:
- Screens display 8 bits per channel. A modern monitor shows 16.7 million colors (8 bits × 3 channels). The extra bits in a 16-bit TIFF are invisible on any display made in the last 20 years.
- The eye can't distinguish them in static viewing. The extra bits buy you headroom — they matter when you're pulling detail out of shadows in post-processing, not when you're viewing the final image.
- Web, email, and most archival viewers expect 8-bit. Anything else just gets downconverted at view time anyway.
Where 16-bit matters: medical imaging (radiologists adjust window/level continuously), satellite imagery (analysts stretch histograms aggressively), and high-end print work that needs every tone from a raw scan. If your workflow falls in any of those buckets, keep the TIFF — JPG is the wrong tool. For everything else, 8-bit JPG is fine.
The technical conversion is a single tone-mapping step: the 65,536 levels per channel in 16-bit are quantized down to the 256 levels in 8-bit. No information is "lost" in the visible sense; you're trading dynamic range you can't see for file size you don't need.
The CMYK problem
Print shops deliver CMYK TIFFs because offset presses lay down four inks — cyan, magenta, yellow, key (black). A JPG meant for screen must be RGB because every display emits red, green, and blue light.
The conversion from CMYK to RGB is a deterministic math operation, not a lossy approximation. Every (C, M, Y, K) quad maps to a unique (R, G, B) triple using either a simple formula or an ICC profile (the profile-aware path is more accurate and is what Photoshop uses by default).
Two things to know:
- Color shift is possible. CMYK has a smaller gamut than RGB in some regions (deep blues, saturated greens) and a larger gamut in others (it can reproduce some yellows and oranges that sRGB cannot). A naïve conversion will compress the CMYK gamut into sRGB, which can flatten certain print-only colors. Profile-aware conversion minimizes this.
- Black ink gets baked in. In CMYK, "K" is its own channel — text and line art are often 100% K with no C, M, or Y. When you convert to RGB, that pure black becomes RGB(0,0,0), which is also pure black on screen. No problem. But if you later go back to print, you've lost the "this is K ink, not RGB black" information. Keep the CMYK TIFF if you ever need to reprint.
For most TIFF-to-JPG conversions, the CMYK-to-RGB step is invisible. The result looks right on screen. If you're converting a print-ready magazine cover for web preview, accept the conversion and move on.
Quality settings by scan type
Not all scans are alike. The right JPG quality depends entirely on what's in the image.
Text scans — contracts, receipts, letters, forms. Text is high-contrast: pure black ink on near-white paper. High-contrast content compresses extraordinarily well in JPEG because the DCT transform preserves sharp edges and flat fills efficiently. Quality 75 is more than enough. A 300 DPI letter-size page typically lands at 150–250 KB. Going below 70 starts to introduce ringing artifacts around character edges, which can fool OCR — so don't.
Photo scans — old photographs, prints, slides. Continuous tones, subtle gradients, fine grain. These are exactly the content JPEG was designed for, but quality needs to stay high to preserve shadow detail and avoid banding in skies. Quality 85–90 is the right range. A 600 DPI 4×6 photo scan at q85 is typically 1.5–3 MB; at q90 it's 3–5 MB.
Mixed content — magazine pages, brochures, documents with embedded photos. The text wants q75; the photos want q90. Compromise at q80, and accept that the photo regions will show very mild ringing on close inspection.
Line art and diagrams — engineering drawings, maps, blueprints. These compress terribly in JPEG because the format is designed for continuous tones, not sharp edges. If your "TIFF" is actually line art, convert to PNG instead. PNG handles 1-bit and palette-based images losslessly and often smaller than JPEG for this content.
File size: what to actually expect
Real numbers from a 50-page legal document scanned at 300 DPI, 24-bit RGB:
| Format | Size | Notes | |---|---|---| | Uncompressed multi-page TIFF | 1.18 GB | The "raw" scanner output | | LZW-compressed TIFF | 612 MB | Lossless, ~50% reduction | | JPG q75 (per page, 50 files) | 8.4 MB total | 96% reduction vs uncompressed | | JPG q85 (per page, 50 files) | 14.1 MB total | 95% reduction, near-lossless visually | | JPG q90 (per page, 50 files) | 24.8 MB total | 92% reduction, indistinguishable on screen |
For a single 4000×3000 photo scan as 48-bit TIFF:
| Format | Size | |---|---| | Uncompressed 48-bit TIFF | 72 MB | | 16-bit LZW TIFF | 41 MB | | JPG q85 (8-bit) | 2.1 MB | | JPG q90 (8-bit) | 3.8 MB |
The takeaway: a 50 MB TIFF → 2 MB JPG at q85 is a 96% reduction with no visible loss for screen viewing. That's the headline number.
OCR: do it before you flatten
If your TIFFs are scanned documents and you need searchable text — and you almost always do, eventually — run OCR before you convert to JPG.
The reason is simple. A TIFF is a structured image. OCR engines like Tesseract can read the page geometry, work on the uncompressed pixels, and produce clean text output with positional data. Once you flatten to JPG, you've added DCT compression artifacts, lost the multi-page structure, and converted bit depth. OCR still works on JPGs, but accuracy drops 2–5% on text-heavy scans — enough to turn 99% accuracy into 95%, which means one wrong character per 20-word line.
The correct archival pipeline is:
- Master TIFF — uncompressed or LZW, kept as the preservation copy.
- OCR pass on the TIFF — produces a PDF with text layer, ALTO XML, or plain text.
- JPG access copies — derived from the TIFF for web display and email.
- Searchable PDF — derived from the OCR pass for retrieval.
The SynoxTools image-to-text tool handles the OCR step in the browser using Tesseract WASM. Run it on the TIFF first, save the text, then convert the same TIFF to JPG for distribution.
Archival best practice: TIFF master, JPG access copy
Every professional digitization project follows the same two-tier model:
- Master copy: TIFF, lossless, original scan resolution, original bit depth. Stored on redundant storage (two disks plus offsite). Never modified. Never re-encoded. This is the digital equivalent of the original physical artifact — you preserve it so you can always re-derive any derivative.
- Access copy: JPG (or WebP for modern archives), 8-bit, compressed, sized for the intended use. Re-derived from the master whenever needed. If a better compression format comes along in five years, you re-derive from the master — you don't recompress the JPG.
The cost of this model is disk space. A 10,000-page archive at 600 DPI is roughly 5 TB of TIFFs. That's $100–200 of disk in 2026. The cost of not following this model is irreplaceable loss: once you've compressed the only copy to JPG, you can never get the bit depth or the original pixels back.
The rule is simple: TIFFs are masters, JPGs are access copies. Convert freely from the former to the latter; never go the other direction.
Browser-based vs. installed software
Three common paths for TIFF-to-JPG conversion:
ImageMagick (CLI) — the workhorse of every backend pipeline. The basic conversion is a one-liner:
bashconvert input.tiff output.jpg
For multi-page TIFFs, ImageMagick writes output-0.jpg, output-1.jpg, etc., automatically. For quality control:
bashconvert input.tiff -quality 85 -colorspace sRGB output_%03d.jpg
The %03d zero-pads the page index to three digits — output_001.jpg, output_002.jpg — which sorts correctly in file listings. ImageMagick also handles CMYK-to-RGB conversion with -colorspace sRGB or via ICC profiles with -profile. Powerful, scriptable, and free. Requires installing it on your machine.
Adobe Photoshop — the gold standard for color-managed conversion. Photoshop uses ICC profiles by default, handles 16-bit TIFFs natively, and lets you soft-proof the CMYK-to-RGB conversion before committing. Overkill for batch conversion; essential for one-off high-stakes conversions like a magazine cover or a fine-art reproduction.
Browser-based converters — SynoxTools and similar run entirely in your browser using WebAssembly. The TIFF is decoded locally, converted to JPG locally, and the result never leaves your machine. This matters for two reasons:
- Privacy. Legal, medical, and corporate TIFFs often contain confidential information. Uploading them to a server-side converter means trusting that server with every page. Browser-based conversion eliminates that risk entirely.
- Speed. No upload time. A 50 MB TIFF that would take 30 seconds to upload on a typical home connection converts in 2–3 seconds locally.
For one-off conversions or batches up to a few hundred pages, the browser-based path is almost always the right choice. For enterprise-scale digitization (10,000+ pages), ImageMagick on a server is more efficient — but you're trading privacy for throughput.
Batch conversion for whole archives
Digitizing a filing cabinet or a photo collection is a different problem from converting one TIFF. You're looking at hundreds or thousands of files, all needing the same treatment.
The batch workflow:
- Organize TIFFs into a single folder hierarchy (e.g.,
2024/case-files/). - Run a batch converter that processes every TIFF, splits multi-page inputs, and outputs JPGs into a mirror folder structure.
- Verify spot-checked outputs against the originals — same page count, readable text, no color shift.
- Store the master TIFFs in cold storage, the JPGs on fast storage for access.
For archives under a few thousand files, the SynoxTools batch converter handles this entirely in the browser: drop in 500 TIFFs, choose JPG output and quality, get back a ZIP with one JPG per page, named consistently. For larger archives, ImageMagick with a shell script is the standard approach:
bashfor f in *.tif; do convert "$f" -quality 85 -colorspace sRGB "${f%.tif}_%03d.jpg" done
This processes every TIFF in the current folder, splitting multi-page inputs and writing q85 JPGs. The parallel command extends this across CPU cores for 4–8× speedup on large archives.
How to convert TIFF to JPG, step by step
For most users — one TIFF, needs to become one or more JPGs — here's the entire workflow:
- Open the converter. Go to the SynoxTools converter in your browser.
- Drag in the TIFF. The tool reports the file's dimensions, bit depth, color space, and page count immediately. If it's a 50-page TIFF, you'll see "50 pages" — no surprises later.
- Choose JPG as the output format. The quality slider appears, defaulting to 85.
- Set quality based on content:
- Text scan: 75
- Photo scan: 85
- Mixed: 80
- Line art: switch output to PNG instead
- Convert. The tool decodes the TIFF (splitting multi-page inputs into individual frames), applies CMYK-to-RGB conversion if needed, tone-maps 16-bit down to 8-bit, and encodes each frame as a separate JPG.
- Download. For single-page TIFFs, one JPG. For multi-page, a ZIP containing one JPG per page, named
filename-01.jpg,filename-02.jpg, etc. - Verify by opening the result. Confirm page count matches, text is legible, and colors look right. For multi-page TIFFs, check the last page specifically — that's the one that breaks first if anything goes wrong.
The whole process takes 2–5 seconds for a single-page TIFF, 30–60 seconds for a 50-page document. Everything runs in the browser — no upload, no server, no trace of your files anywhere outside your machine.
What you keep and what you give up
Converting TIFF to JPG is a one-way operation. Here is the honest accounting of what changes:
You keep:
- The visible image, indistinguishable on any screen made in the last 20 years.
- The page count (one JPG per page).
- The colors (after CMYK-to-RGB conversion, within sRGB gamut).
- The text legibility (at q75 or above).
You give up:
- Bit depth (16-bit becomes 8-bit). Invisible on screen, gone forever in the file.
- Lossless fidelity (JPEG is lossy; you cannot recover the original pixels).
- CMYK separation (the K channel is baked into RGB; reprinting requires re-separation).
- ICC profile data (unless you explicitly embed it in the JPG).
- Multi-page container (one file becomes many).
- File size (a 96% reduction is the upside; the same 96% is gone from your master if you delete the TIFF).
The decision is straightforward. For viewing, sharing, emailing, and web display, JPG is correct and the losses are theoretical. For printing, re-editing, or long-term preservation, keep the TIFF. For everything in between — most workflows — keep both: TIFF as master, JPG as access copy. That dual-tier model has been the archival standard for 30 years and it isn't changing.