Author Archives:

Media3 1.10 is out

Posted by Andrew Lewis, Software Engineer



Media3 1.10 is out!

Media3 1.10 includes new features, bug fixes and feature improvements, including Material3-based playback widgets, expanded format support in ExoPlayer and improved speed adjustment when exporting media with Transformer. Read on to find out more, and check out the full release notes for a comprehensive list of changes.

Playback UI and Compose

We are continuing to expand the media3-ui-compose-material3 module to help you build Compose UIs for playback.

We've added a new Player Composable that combines a ContentFrame with customizable playback controls, giving you an out-of-the-box player widget with a modern UI.

This release also adds a ProgressSlider Composable for displaying player progress and performing seeks using dragging and tapping gestures. For playback speed management, a new PlaybackSpeedControl is available in the base media3-ui-compose module, alongside a styled PlaybackSpeedToggleButton in the Material 3 module.

We'll continue working on new additions like track selection utils, subtitle support and more customization options in the upcoming Media3 releases. We're eager to hear your feedback so please share your thoughts on the project issue tracker.

 Player Composable in the Media3 Compose demo app

Playback feature enhancements

Media3 1.10 includes a variety of additions and improvements across the playback modules:
  • Format support: ExoPlayer now supports extracting Dolby Vision Profile 10 and Versatile Video Coding (VVC) tracks in MP4 containers, and we've introduced MPEG-H UI manager support in the decoder_mpeghextension. The IAMF extension now seamlessly supports binaural output, either through the decoder viaiamf_tools or through the Android OS Spatializer, with new logic to match the output layout of the speakers.

  • Ad playback: Improvements to reliability, improved HLS interstitial support forX-PLAYOUT-LIMIT  and X-SNAP, and with the latest IMA SDK dependency you can control whether ad click-through URLs open in custom tabs with setEnableCustomTabs.

  • HLS: ExoPlayer now allows location fallback upon encountering load errors if redundant streams from different locations are available.
  • Session: MediaSessionService now extends LifecycleService, allowing apps to access the lifecycle scoping of the service.

One of our key focus areas this year is on playback efficiency and performance. Media3 1.10 includes experimental support for scheduling the core playback loop in a more efficient way. You can try this out by enabling experimentalSetDynamicSchedulingEnabled() via the ExoPlayer.Builder. We plan to make further improvements in future releases so stay tuned!

Media editing and Transformer

For developers building media editing experiences, we've made speed adjustments more robust. EditedMediaItem.Builder.setFrameRate()can now set a maximum output frame rate for video. This is particularly helpful for controlling output size and maintaining performance when increasing media speed with setSpeed().

New modules for frame extraction and applying Lottie effects

In this release we've split some functionality into new modules to reduce the scope of some dependencies:

  • FrameExtractor has been removed from the main media3-inspector module, so please migrate your code to use the new media3-inspector-framemodule and update your imports toandroidx.media3.inspector.frame.FrameExtractor.

  • We have also moved theLottieOverlayeffect to a separate media3-effect-lottie module. As a reminder, this gives you a straightforward way to apply vector-based Lottie animations directly to video frames.

Please get in touch via the issue tracker if you run into any bugs, or if you have questions or feature requests. We look forward to hearing from you!

Monzo boosts performance metrics by up to 35% with a simple R8 update

Posted by Ben Weiss, Senior Developer Relations Engineer

Monzo is a UK digital bank with 15 million customers and growing. As the app scaled, the engineering team identified app startup time as a critical area for improvement but worried it would require significant changes to their codebase.

By fully enabling R8 optimizations, Monzo achieved a massive 35% reduction in their Application Not Responding (ANR) rate. This simple change proved that impactful optimizations don't always require complex engineering efforts.

Unlocking broad performance wins with R8 full mode

Monzo identified R8 full mode as an easy fix worth trying; and it worked, improving performance across the board:

  • Startup Reliability: Cold starts improved by 30%, Warm starts by 24%, and Hot starts by 14%.
  • Launch Speed: P50 launch times improved by 11% and P90 launch times by 12%.
  • Efficiency: Overall app size was reduced by 9%.
  • Stability: ANR reduction of 35%.

Enabling optimizations with a single change

Many Android apps use an outdated default configuration file which disables most functionality of the R8 optimizer. The main change Monzo made to unlock these performance improvements was to replace the proguard-android.txt default file with proguard-android-optimize.txt. This change removes the -dontoptimize instruction and allows R8 to properly do its job.

buildTypes {
  release {
    isMinifyEnabled = true
    isShrinkResources = true
    proguardFiles(
      getDefaultProguardFile("proguard-android-optimize.txt"),
    )
  }
}

After making this change, it's worth looking at your Keep configuration files. These files tell R8 which parts of your code to leave alone (usually because they're called dynamically or by external libraries). Tidying up unnecessary Keep rules means R8 can do more.

Improving scroll performance with Baseline Profiles

To further enhance the user experience, Monzo implemented Baseline Profiles, specifically targeting scroll and rendering performance on their main feed. This strategy ensured that the most common user journeys—opening the app and scrolling the feed—were fully optimized. The impact on rendering was substantial: P90 scroll performance became 71% faster, and P95 scroll performance improved by 87%. Now scrolling the app is smoother than before.

Monzo built this into their release process to maintain these improvements over time. "We trigger the baseline profile generation every week day (before running our nightly builds) and commit the latest changes once completed," Neumayer explains.

Keeping up with modern Android development

Monzo's experience shows what's possible when you stay up to date with Android build-tooling recommendations. While legacy apps often struggle with complex reflection usage, Monzo found the transition straightforward by documenting their Keep Rules properly. "We always add a comment explaining why Keep Rules are in place, so we know when it's safe to remove the rules," Neumayer notes.

Neumayer's advice for other teams? Regularly check your practices against current standards: "Take a look at the latest recommendations from Google around app performance and check if you're following all the latest advice."

To get started and learn more about R8, visit https://d.android.com/r8

An update on secondary calendar lifecycle changes and a new API

We recently announced a change to the ownership model of secondary calendars to improve data governance. As part of this, we emailed impacted customers to let them know that orphan secondary calendars would be deleted starting on April 27, 2026.

Since that announcement, we’ve received valuable feedback that to properly manage this new lifecycle, customers need better programmatic tools to handle secondary calendar data before it gets deleted.

To ensure you have the time and tools necessary to manage this transition smoothly, we are making two important updates:

  1. We are launching a new API endpoint by the end of June to transfer secondary calendars within your organization.
  2. We are postponing the secondary calendar lifecycle changes to October 5, 2026 for paid accounts.

Coming soon: new API endpoint to transfer secondary calendars

In the coming months, we’ll introduce a new endpoint in the Calendar API that will allow developers to programmatically transfer the ownership of secondary calendars. This endpoint will require the Calendar administrator privilege.

The API will mirror the existing capabilities in the Admin console—transferring secondary calendars within the same organization without requiring confirmation by the receiving user—and introduces the additional flexibility to transfer individual calendars.

The new API endpoint will be available for integration by June 2026. An announcement and technical documentation will be published when the API goes live.

Extended deadline for lifecycle changes

To give your teams ample time to adjust their workflows and integrate with the new API endpoint, we are officially pushing back the enforcement date for the secondary calendar lifecycle changes for paid accounts.

The new policy—where secondary calendars are permanently deleted upon the deletion of the owner's account—will now take effect on October 5, 2026, for paid accounts.

Until then, we will run a regular process for orphan calendars that auto-assigns ownership to a user who has “Make changes and manage sharing” access. This process will stop on October 5, 2026. Instead, make sure to ask the owner to transfer relevant secondary calendars to a colleague before they leave - or make sure an administrator executes the transfer using the Admin console or the new API endpoint.

Note that the changes to the secondary calendar lifecycle will still take effect on April 27, 2026, for users with personal Google accounts.

Additional details

Secondary calendars owned by an organization must be owned by a user within that same organization, and ownership transfers are restricted to users in the same domain. However, you can continue to share calendars with users outside your organization—including with high-level permissions such as "Make changes and manage sharing"—provided your organization policies allow it.

Example of a secondary calendar owned by the dwelling.com organization, with [email protected] as owner

Example of a personal secondary calendar (not owned by an organization), with [email protected] as owner 

Getting started

  • Admins: Stay tuned for more details on the new API endpoint when it launches.
  • End users: There is no end user setting for this feature. End users can already transfer secondary calendars to other users within their organization. Visit the Help Center to learn more.

Rollout pace

  • Rapid Release and Scheduled Release domains
    • New API endpoint: Gradual rollout (up to 15 days for feature visibility) starting in June (to be announced on Workspace Updates blog when available)
    • Secondary calendar lifecycle change: Gradual rollout (up to 15 days for feature visibility) starting on October 5, 2026
  • Users with personal Google accounts
    • Secondary calendar lifecycle change: Gradual rollout (up to 15 days for feature visibility) starting on Apr 27, 2026

Impact

  • All Google Workspace customers and users with personal Google accounts are impacted by these changes 

Resources