Perfect Images, Lower Costs: How Mediacorp Solved Both

ABOUT

Singapore’s Largest Media Network

Mediacorp is Singapore’s national media network and largest content creator. Its purpose is to create engaging and trusted content, as well as to connect communities and inspire people. Mediacorp engages over 3 million people in Singapore daily across four languages on its digital platforms including mewatch, melisten, and CNA.asia, six TV channels, and 11 radio stations. Beyond Singapore, Mediacorp also has a growing international audience through CNA and content distributed across markets.The company is the recipient of several industry accolades including New York Festivals, World Media Festivals, PromaxBDA (World and Asia), and Asian Academy Creative Awards.

THE CHALLENGE

Hitting the Limits of Storage and Traffic

Working in four languages across six TV channels, 11 radio stations, several digital news and podcast platforms, and associated viewer newsletters, Mediacorp works tirelessly to keep over 3 million Singaporean residents informed and entertained every day.

Key to this mission is imagery, which Mediacorp manages on a massive scale. For years images had been managed in a third-party commercial content management system (CMS), but the system was starting to strain as usage scaled, according to Mediacorp’s Solutions Architect, Siva Subramanian.

“The main challenge we started to have was with storage — in terms of the physical amount we were accumulating and also the cost of serving images,” says Subramanian. By storing a copy of every image used on its own and third-party servers, Mediacorp was starting to hit the buffers on what its partner could store and incurring mounting traffic costs.

In 2012, Subramanian thought he could solve the problem by switching to the popular open-source CMS, Drupal. But over time, the model proved less sustainable as costs continued to rise. “We had been hosting all the images files, and realized that the images were now at 1,000 GB, but the code base was less than one GB; so 1,000 times bigger than our code base,” Subramanian points out.

Another issue that began to surface was feedback from editorial colleagues and photographers about the published image quality across the expanded set of channels.“ Despite our best efforts using existing tools, there remained quality issues with the images in things like sharpness and accurate skin tones,” says Subramanian. “We needed to remedy this situation for good.”

THE CLOUDINARY SOLUTION

‘The Name That Kept Coming Up’

Subramanian led a thorough search for a DAM that could efficiently serve images at massive scale, ultimately choosing Cloudinary. To evaluate potential solutions, his team sought recommendations from colleagues at other media companies. “Cloudinary was the name that kept coming up again and again.”

His team really valued Cloudinary’s approach for optimizing image format size and quality, with a commitment to the AV1 Image File Format (AVIF) format. Cloudinary’s transformations that optimize images for format and quality (f_auto and q_auto) were key to solving the two main concerns, saving on storage costs and improving image quality. Crucially, they were able to do this on Medicorp’s massive scale. Cloudinary’s transformations also helped Mediacorp successfully transition from JPEG to AVIF, a process that had previously stalled due to quality issues.

Subramanian says a combination of consistent use of AVIF over JPEG and Cloudinary transformations changed the whole Mediacorp image delivery situation overnight: Images loaded faster, file sizes were redacted, and image quality improved.

Tight integration between Drupal and Cloudinary has made image hosting service much more straightforward and cost-effective. Subramanian explains how this works in practice: “Our regular CMS users, 30 or so editors, log into our CMS, upload images, and then push them directly into Cloudinary instead of saving them in Drupal. By doing so we avoid all the costs that had been such a problem previously.”

THE RESULTS

Solving for Quantity, Team Freed Up to Boost Quality

With the high-volume image workflow now automated and running at peak efficiency, and storage costs reduced, the web team has more capacity to focus on enhancing quality and service. Pre-sets are in place so that upon upload every image is dynamically transformed based on requirements. Team members very rarely even need to log into the Cloudinary console, so smooth is the process now.

A major boon has been a dramatic reduction in image quality-related complaints. As Subramanian explains, “The few [complaints] we get now are from super-proficient photographers who spot flaws that consumers are unlikely to detect. And with Cloudinary we’re able to address these quickly by simply changing the parameter in the URL.”

Another benefit has been how easily the right images are served on an important part of Mediacorp’s offering: its multi-lingual e-newsletter channels. “Now, all our newsletters’ image links are from Cloudinary, which again makes life much easier.”

Looking ahead, Subramanian is excited to explore Cloudinary’s advanced AI features to help further restore and enhance image quality.