Opens in a new tab
EST. 2007
ADMISSIONS OPEN
Apply Now
← Creative Hut Library
Photography basics · Image files

The Digital Image Pipeline

A Guide for Photographers and Cinematographers

Written by Abin Alex and Anu Joseph.

Creative Hut Academy Library · Last updated 31 October 2026

Abstract

The digital image pipeline turns light into a file in four steps: sensor readout, analogue-to-digital conversion, processing and saving. This guide explains each step and the meaning of bit depth. It distinguishes the RAW file, which keeps the sensor data with little processing, from the JPEG file, which is processed and compressed in the camera, and it outlines other file types and how to choose between them. It closes with the changes AI brings to the pipeline. It is written for photographers and cinematographers.

Keywords: digital image pipeline; RAW; JPEG; bit depth; analogue-to-digital conversion

What happens between the sensor and the file?

Short answer. The sensor reads out an analogue signal, a converter turns it into numbers, the processor builds an image and the camera saves a file (Nakamura, 2006).

  • Sensor readout: each photosite holds an electrical charge in proportion to the light it received (Nakamura, 2006).
  • Analogue-to-digital conversion (ADC): the charge is converted into a number. The number of bits decides how many tone levels are possible (Nakamura, 2006).
  • Image processor: demosaicing builds full colour from the Bayer mosaic, then noise reduction, sharpening and tone mapping follow (Nakamura, 2006) (Bayer, 1976).
  • File: the result is saved as RAW, JPEG or HEIF.
  • Colour theory such as white balance and colour spaces is covered in Unit III of the course.
StepWhat happensOutput
1 ReadoutCharge is read from each pixelAnalogue signal
2 ADCSignal is turned into numbersRaw digital data
3 ProcessorDemosaic, denoise, sharpen, tone mapImage data
4 SaveData is encoded and storedRAW, JPEG or HEIF file

Table 1. The four steps of the digital image pipeline (Nakamura, 2006).

In photography
Your camera does all four steps in a fraction of a second after you press the shutter button.
In cinematography
Cinema cameras follow the same pipeline, but they record the data in formats built for grading, such as RAW video.
[Lecture example slot: a flow diagram of the pipeline]

Cham tipLearn the four words: readout, convert, process, save.

Cham taskWrite the four steps on a card and add one setting that affects each step.

Check yourself: What does the ADC do?

It converts the analogue sensor signal into digital numbers.

What is bit depth?

Short answer. Bit depth is the number of bits used to store each colour channel. More bits mean more tone levels and smoother gradients.

  • An 8-bit JPEG has 256 tone levels per channel, which gives about 16.8 million colours (256 × 256 × 256).
  • A 12-bit RAW file has 4096 levels per channel, which gives about 68.7 billion colours.
  • A 14-bit RAW file has 16 384 levels per channel, which gives about 4.4 trillion colours.
  • More levels give finer control in editing, especially in skies and shadows, where banding can appear in 8-bit files.
Bit depthLevels per channelTotal colours
8-bit25616.8 million
12-bit409668.7 billion
14-bit16 3844.4 trillion

Table 2. Bit depth and colours. Totals are levels cubed (calculated).

In photography
If you plan to edit heavily, shoot a higher bit depth.
In cinematography
Log and RAW video recording use higher bit depth so that colourists have room to grade.
[Lecture example slot: a gradient with banding]

Cham tipBit depth is about smoothness of tone, not about how many colours the eye can name.

Cham taskCalculate the colours for 10-bit video: 1024 × 1024 × 1024.

Check yourself: How many levels does 12-bit give?

4096 levels per channel.

What is the difference between RAW and JPEG?

Short answer. A RAW file holds unprocessed or lightly processed sensor data and needs software before use. A JPEG is processed, compressed and ready to share.

  • RAW is called a digital negative. It must be processed in software before it is printed or shared.
  • RAW has a wider dynamic range and colour gamut, finer editing control and the option to change the colour space after capture.
  • JPEG opens in almost any program, uses compression and records 8 bits per channel, so it has less room for editing.
  • RAW files are usually larger than JPEG files because they hold more data. RAW formats differ by manufacturer, so each needs compatible software.
  • Many cameras can record both at once, so you can share the JPEG and keep the RAW.
PointRAWJPEG
ProcessingDone later in softwareDone in the camera
Bit depth12 to 14 bit common8 bit
File sizeLargerSmaller
Editing roomHighLimited
Ready to shareNoYes

Table 3. RAW and JPEG compared.

In photography
Shoot RAW plus JPEG while you learn. You get a quick file and a safe master.
In cinematography
Cinema RAW and log recording follow the same idea: keep the most data and decide the look later.
[Lecture example slot: a RAW file and its JPEG]

Cham tipUse RAW when the light is difficult and JPEG when speed matters more than editing.

Cham taskShoot the same scene in RAW and JPEG. Lift the shadows by two stops on both and compare.

Check yourself: Which file type is usually larger, RAW or JPEG?

RAW.

What other image file types are there?

Short answer. Image file types differ in how they compress data. Some lose data and some keep it exactly.

  • TIFF is flexible and usually lossless, so files are large.
  • PNG is lossless and exactly reversible.
  • GIF uses a table of up to 256 colours.
  • JPEG is optimised for photographs and reaches high compression.
  • BMP is an uncompressed format created by Microsoft.
  • A raster image is a grid of pixels. A vector image is stored as lines and curves and can be scaled to any size.
TypeCompressionBest for
JPEGLossyPhotographs for web and sharing
PNGLosslessGraphics, screenshots
TIFFUsually losslessPrint and archive
GIF256 coloursSimple animations
Vector (SVG, AI)Not pixelsLogos, illustrations

Table 4. Common image file types.

In photography
Save a TIFF or the original RAW as your master, and export JPEGs from it for sharing.
In cinematography
Cinema pipelines use image sequences such as DPX or EXR for effects work.
[Lecture example slot: a folder of different file types]

Cham tipLossless means you can recover the original exactly. Lossy means some data is gone for good.

Cham taskSave one photo as JPEG at three quality levels. Compare file sizes and quality.

Check yourself: Is PNG lossy or lossless?

Lossless.

How does AI change the pipeline?

Short answer. AI now helps with noise reduction, upscaling and subject recognition inside the pipeline, and content credentials can sign the result (Lukac, 2011).

  • Computational methods can merge frames, reduce noise and recognise subjects before the file is saved (Lukac, 2011).
  • AI upscaling creates detail by prediction. It is a guess, not a measurement.
  • C2PA content credentials are a technical standard that lets a camera or editor sign the capture and edit history cryptographically.
  • Credentials do not prove that a picture is true. They record where the file came from and what was done to it.
  • Decide what your work needs: an honest record, or the best-looking result.
In photography
Keep the RAW original of any picture you edit heavily with AI tools.
In cinematography
Keep the original camera files of any footage you edit heavily, so you can show what was recorded.
[Lecture example slot: a content credentials label on a photo]

Cham tipKnow what your camera does before the file is written.

Cham taskCheck whether your camera or phone offers content credentials. Write what it records.

Check yourself: Does AI upscaling add real detail?

No. It predicts detail, so the added detail is a guess.

Questions beginners ask

What is the digital image pipeline?

It is the chain from sensor readout and analogue-to-digital conversion to processing and saving the file.

What is bit depth?

Bit depth is the number of bits used for each colour channel. 8-bit has 256 levels, 12-bit has 4096 and 14-bit has 16 384.

What is a RAW file?

A RAW file holds unprocessed or lightly processed sensor data. It needs software to process it before use.

Is RAW better than JPEG?

RAW gives more editing room. JPEG is smaller and ready to share. The better choice depends on the job.

Are RAW files bigger than JPEG?

Usually yes, because RAW holds more data and JPEG is compressed.

What are C2PA content credentials?

They are a technical standard that records and signs where a file came from and how it was edited.

References

  • Bayer, B.E. (1976) Color imaging array. US Patent 3,971,065. Washington, DC: US Patent and Trademark Office.
  • Lukac, R. (ed.) (2011) Computational Photography: Methods and Applications. Boca Raton: CRC Press.
  • Nakamura, J. (ed.) (2006) Image Sensors and Signal Processing for Digital Still Cameras. Boca Raton: CRC Press.

How to cite this guide. Alex, A., & Joseph, A. (2026). Digital image pipeline: A guide for photographers and cinematographers. Creative Hut Academy. https://creativehut.org/library/digital-image-pipeline/

© 2026 Creative Hut Academy. Written by Abin Alex and Anu Joseph. Drafted with AI assistance. Last updated 31 October 2026. This guide is free to read. To copy, republish, translate or adapt any part of it, including the diagrams, you need written permission. Ask through the contact page. The books, standards and data sheets listed above are sources, not copies. Found a mistake? Write to director@creativehut.org.
More guides in the Library →
Learn this hands on

Make this your craft at Creative Hut Academy

Creative Hut Academy in Mattakkara, Kottayam, Kerala teaches photography, cinematography, editing, sound and AI filmmaking in one-year residential diplomas. Students come from across India.

Want to talk? Message us on WhatsApp, or ask Cham, our AI mentor, a study doubt.