Raw vs Jpeg vs Tiff

This is one of those episodes I wanted to make a long time before this blog ever existed. However, since I am doing my best to publish according to a logical thread I’ve always found myself coming up with other areas of focus that took priority; at least until now. Photography is a passionating art which brings people to discover new horizons and often argue virulently on numerous topics such as which camera manufacturer is better between Canon and Nikon. The output format produced by Digital Cameras has had undoubtedly its fair share of blog discussions.  So why do I bother publishing a video and a blog post on this well-covered topic you may ask. Well, I have often heard and read people are starting they argumentation by saying RAW is THE format you must use but without really giving you much insight and this is what I am attempting here.

How does a Digital camera work?

The photon wells

Your sensor is responsible for converting the optical information (information coming from the light through the glass elements) into a Digital one that can be rendered as an image. The way it works is that it collects photoelectrons from the light photons that collide with the sensor photovoltaic surface. What makes the surface of the sensor sensitive to light is quite complex but for the sake of understanding, you can visualise it as a field with millions of wells whose sole purpose is to capture the photons (information from the light). The number of these photon wells depends on of the size of the sensor, and in most Digital Cameras, you will have one photon well per pixel. All photon wells have the same depth making them capacity limited. However an another valuable aspect is the width of the well itself; the wider/larger the photon well, the more information is being captured. Interestingly if you take two sensors of the same physical size but one has 18MB, and the other has 12MB, the second sensor will most likely produce a better image at a higher ISO (we will see this further down below).

The ISO

We already covered the ISO, but this was mostly focusing on its role within the Exposure Triangle. I had even written: “Each time you increase the ISO value by one full stop, you increase the camera sensor sensitivity, so it captures twice more information from the light.” However, this is not quite accurate. The ISO will not increase the sensitivity of the sensor but rather work as an amplifier. Once the exposure time is up, the photon wells have been filled but to read the Wells’ content, the sensor needs to amplify it, and that is what the ISO does. Analogy: Imagine a 10m deep well with a pump that is just 3m below the ground level. It’d mean that until you have at least 7m of water, your pump won’t be able to suck anything. The ISO is there to amplify the amount of information to make it worth analysing further. Going back to what we saw above when the photon wells are less numerous but wider; as you can imagine, the larger the photon well, the more information is being collected. Hence you do not need to amplify it by much to make it readable. Therefore, sensors of equivalent physical size with less numerous but wider wells will offer better result in higher ISO.

Raw

Raw or RAW is not an acronym. It only refers to the Raw data (NEF data for Nikon) that is produced by the sensor and read out from the Photon Wells (once the ISO setting has been applied). It is quite similar to the Negative in Film Photography. Any digital camera does produce Raw data, even though only DSLR and few Point&Shoots offer it as an output format. The RAW (or NEF ) file can be open by some post-production tools such as the basic one that came with your DSLR or more advanced like Adobe Photoshop Camera RAW or Adobe Lightroom.

Pixels

I have often said that Photography is the art of capturing light and lately we even saw that light is colourful (ref The White Balance). We know that the millions of colours you can achieve with your Digital camera are based on combinations of the three channels: Red, Green and Blue. Your sensor is made of millions of pixels (i.e The Canon 60D has 17915904 pixels arranged by 5184×3456). While the role of the pixels is to absorb the light information (colours), individually, each pixel is responding to a specific channel: meaning that some pixels will react to the Red information only, others to Blue exclusively and the rest to Green.  This is the Bayer sensor which is the most common sensor found in Digital Cameras today. Note: Since our eye is more receptive to the variations Green, the number of Green pixel is more important than of Red or Blue.

Raw data processing

Once the Raw data is produced, it needs to be processed to obtain the pixel colour information, and this is known as the Bayer Interpolation. Once the colours have been defined, the image is taking shape, but other settings need to be applied such as the White Balance, the contrast, the sharpness, saturation… Those settings are usually defined through your camera menus.

JPEG compression

At this stage, the image is done, and the camera needs to generate an output file that will be, storable and readable by most programs running on phones, computers or even directly handled by modern printers. JPEG is an acronym: Joint Photographic Expert Group and is a lossy compression format for Digital image. The compression level is adjustable. However, there is one important fact to know: each time you open a JPEG image in software, edit it and then save the result, the lossy compression takes place leading you to an unavoidable quality loss.

TIFF

Tagged Image File Format is a lossless format for Digital image that is highly praised by publishing and photography companies. It was invented by Aldus which was acquired by Adobe who since 2009 hold the copyright and its specification. The amazing thing about TIFF is that while one can compress it (8-bit or 16-bit), one does not loose any data unlike with JPEG. To understand the way it works, imagine data that is identified as array collections and the definition of the array takes less space than its content. Some Point&Shoot cameras enable you to choose TIFF as the output format.

Here is the image creation workflow:

Why do serious photographers praise RAW over JEPG?

As you can see in the diagram above, RAW file is generated as soon as the information from the light has been converted into data. That data is not yet an image, and it needs to be interpreted by a program to generate a visual image, whether it is in the camera or in a post-production software such as the basic one that was shipped with your camera or more advanced ones like Adobe Photoshop Camera Raw or Adobe Lightroom. Camera manufacturers provide their RAW/NEF specifications to the software editors so they can produce proper interpreter. It is important to acknowledge that Raw/NEF data specification improve all the time, and so do the interpreters, hence should you take a RAW from 5 years ago and compare the resulting image you got then with the one you would with today software, you are bound to see massive improvement in the interpretation of the data and what can be done to it. Therefore I would rather rely on software to generate the final image than the camera itself. Once the ISO readout is performed, and the RAW data is generated, no interpolation, no image analysis has yet been performed giving you a quasi-total freedom to define in post-production the White Balance, the contrast, the Saturation, the sharpness and many more alterations. The generation of JPEG is taking place after all those alterations. Without going through all settings, you can easily understand that when you set your White Balance or the saturation (increasing the colour intensity in the image) you are overwriting data and generating a JPEG file after that stage will provide you with, no matter what compression level you choose, a file with far less data than the RAW file. When editing a JPEG image in post-production is like if you were going to a photo shop and handing them a printed image, asking them to retouch it (alter the exposure, change the colour saturation, the skin tone, etc.). The shop technician can still do some alteration but he is undoubtedly more limited than if you had handed over to him the Negative. In previous episodes, we saw the difference between a Correct Exposure and a Correct Creative Exposure. The main difference is the word “Creative” which means your resulting Photography needs to bring some artistic value by its content of course but also by its composition, how the subject is shown. Post-production is part of that process, and we will see in future episodes what to expect from post-production tools.

If RAW is the synonym of Freedom, why most cameras out there are limited to JPEG?

The answer is simple: one cannot look at a Raw file since it is not yet an image. JPEG is an image, hence can be shared instantly with all via phone, memory stick, the internet and viewed or even printed just as quickly as it is received. The file size is rather small from a few kilobytes to less than 10 Megabytes depending on its compression whereas a RAW file is usually around 24MB and above depending on the sensor size and pixel count. This means that on the same SD card or other memory media, one will be able to store in an average of 1.4 times more JPEG than RAW files. So while the price of the storage media is getting cheaper every day, it still represents a cost and trusts me this is getting even truer when looking at storage once the files are on the computer. All these points make the JPEG the most consumer friendly format, at first.

Note 1

Cameras that produce the Raw file as an optional output format, often offer you the freedom to produce both Raw and JPEG for the same shot, giving you two output files. One could say this is giving your the best of both worlds, the ready-to-eat format and the post-production format. As far as I am concerned, I only shoot JPEG when making shots for this blog I know I do not want to edit or when I shoot with my phone to capture and share an instant image.

Note 2

What about TIFF file then? Well TIFF that was produced by the camera does contain the setting choices that were applied after the Bayer Interpolation, so the file does not contain all the Raw information.