Comparing Image Sequences Vs. Direct Video Rendering In Blender
What is the Difference Between Image Sequences and Video?
Image sequences and video files serve different purposes in Blender. An image sequence is a series of individual image files, numbered in sequence, that can be played back to simulate video. With image sequences, each frame is saved as a separate file on disk. Video files combine these image frames into a single compressed video file format such as MP4 or AVI.
Image sequences allow more flexibility for rendering and post-processing visual effects shots. Each frame can be re-rendered independently without needing to re-render the entire video. Video files are more efficient for playback and distribution, saving disk space and upload bandwidth compared to thousands of individual image files.
Purpose and Use Cases
The primary use case for rendering to an image sequence is visual effects and motion graphics work where shots will be composited with other rendered elements in post-production. Saving each frame independently makes them easier to re-render and modify later.
Rendering directly to a video file is preferred when the resulting animation will be played back directly without further compositing. This includes concept demos, animatics, presentations, video content, and final render output for clients.
Technical Details
From a technical perspective, image sequences store full 32-bit float color and alpha transparency information per pixel for every frame. This allows much more flexibility for color correction and compositing later on. Video codecs throw away visual information to save space, using 8-bit color depths and temporal compression between frames.
Due to their uncompressed nature, image sequences result in much larger file sizes on disk. A short 10 second animation at 30 FPS will generate 300 image files versus a single 10 second video file. However, video files become unwieldy above 4-8GB in size, making image sequences a safer approach for longer high-resolution animations.
Rendering Image Sequences
Setting Up the Render Output
Rendering to an image sequence is configured in the Output properties tab of the Properties Editor. Under the File Format dropdown, select a desired image format such as PNG, OpenEXR, or other options.
In the same Render Output panel, enable the checkbox next to “Save as Render Layers” and configure the Number of Digits used, such as 4. This will save each frame to sequentially numbered files like filename.0001.exr rather than overriding the same file each frame.
Controlling Frame Rate and Encoding
The frame rate of the image sequence animation is set under the Output Properties > Dimensions panel. Common rates are 23.98, 24, 25, 29.97, 30, 50, 60 FPS. The resolution can also be configured here.
On the File Format panel, certain image codecs have encoding/compression settings such as PNG and OpenEXR. Configuration options may include bit depth, color management, and lossy/lossless compression to balance visual quality and file size.
Example Render Settings
A sample image sequence configuration for visual effects might be:
- Output Format: OpenEXR
- Color Depth: 32-bit floating point
- Compression: DWAA Lossy near lossless
- Frame Rate: 24 FPS
- Resolution: 1080p or 4K
- Save Frames Individually Numbered
This preserves full visual quality with manageable file sizes. The individual numbered image files can easily be re-rendered as needed.
Rendering Video Files
Video Codecs and Formats
Blender supports rendering directly to video file formats including H.264, H.265, MPEG-4, AVCHD, and others. The specific list of available options depends on the platform and installed codecs.
H.264 and H.265 are common delivery codecs offering reasonable visual quality and file sizes. AVCHD is optimal for playback on consumer devices. MPEG-4 balances quality and file size. An advanced user might opt for lightly compressed or lossless codec like ProRes, DNxHD, or FFV1.
Encoding Settings and Optimization
Under the encoding settings for each video codec, the most critical setting is bitrate, measured in megabits per second (Mbps). This directly controls the final output quality and file size. A good target for 1080p is 5-10Mbps, increasing to 45+ Mbps for 4K UHD to maintain quality.
One can optimize the encoding profile settings such as GOP Size and motion estimation method to find the best size/quality trade-off. Using two-pass encoding analyzes then optimizes the video encoding but takes twice as long to render.
Example Encoding Settings
A versatile set of encoding settings for H.264 MP4 video output would be:
- Format: H.264 (MP4 container)
- Profile: High 4:2:0 8-bit
- Bitrate: 10Mbps 1080p, 45Mbps 4K UHD
- GOP Size: 60 frames
- Rate Control: Constant Bitrate (CBR)
This delivers great quality MP4 video for sharing online and with reasonable file sizes. Certain settings could be tweaked to optimize for certain target playback devices if needed.
Performance Considerations
Render Times and File Sizes
Due to video compression distributing data across frames, rendering straight to video files is typically 15-30% faster than image sequences for the same resolution and frame rate. However, video files must be re-rendered fully if any section needs to change meanwhile image sequences allow re-rendering only the needed frames.
The trade-off for faster encode times is that video files achieve much smaller file sizes by throwing away color and timing data. A 1080p animation consumes 187 GB as image sequences versus 756 MB in a H.264 MP4 video file without noticeable visual quality loss.
Playback and Editing
Reading thousands of image files requires more disk bandwidth and processing overhead compared to decoding compressed video frames. Playback performance is generally smoother at higher resolutions with video files.
Video files are also easier to edit in applications like Adobe Premiere without needing to import a long image sequence. Offline proxy workflows can aid editing image sequences by using lower resolution copies as placeholders.
When to Choose Image Sequences vs Video
The decision between rendering to image sequences or video output ultimately depends on the intended creative use case.
For final delivery of animations and renders at completion to clients or for video content, direct video output avoids excessive storage from image sequences and offers reliable playback on consumer devices.
For productions with visual effects shots that will integrate 3D rendered elements with live footage or other generated content, image sequences offer the most flexibility for compositing later. Animators and technical directors modify and re-render image sequence shots iteratively until finalized.
Understanding these factors empowers Blender artists to choose the best render methodology for each project’s needs.
