Responsive images

In this article, we'll learn about the concept of responsive images — images that work well on devices with widely differing screen sizes, resolutions, and other such features — and look at what tools HTML provides to help implement them. This helps to improve performance across different devices. Responsive images are just one part of responsive design , a future CSS topic for you to learn.

Prerequisites: You should already know the basics of HTML and how to add static images to a web page .
Objective: Learn how to use features like srcset <picture> element to implement responsive image solutions on websites.

Why responsive images?

Let's examine a typical scenario. A typical website may contain a header image and some content images below the header. The header image will likely span the whole of the width of the header, and the content image will fit somewhere inside the content column. Here's a simple example:

Our example site as viewed on a wide screen - here the first image works OK, as it is big enough to see the detail in the center.

This works well on a wide screen device, such as a laptop or desktop (you can see the example live and find the 源代码 on Github.) We won't discuss the CSS much in this lesson, except to say that:

  • The body content has been set to a maximum width of 1200 pixels — in viewports above that width, the body remains at 1200px and centers itself in the available space. In viewports below that width, the body will stay at 100% of the width of the viewport.
  • The header image has been set so that its center always stays in the center of the header, no matter what width the heading is set at. If the site is being viewed on a narrower screen, the important detail in the center of the image (the people) can still be seen, and the excess is lost off either side. It is 200px high.
  • The content images have been set so that if the body element becomes smaller than the image, the images start to shrink so that they always stay inside the body, rather than overflowing it.

However, issues arise when you start to view the site on a narrow screen device. The header below looks OK, but it's starting to take up a lot of the screen height for a mobile device. And at this size, it is difficult to see the people within the first content image.

Our example site as viewed on a narrow screen; the first image has shrunk to the point where it is hard to make out the detail on it.

An improvement would be to display a cropped version of the image which displays the important details of the image when the site is viewed on a narrow screen. A second cropped image could be displayed for a medium-width screen device, like a tablet. The general problem whereby you want to serve different cropped images in that way, for various layouts, is commonly known as the art direction problem .

In addition, there is no need to embed such large images on the page if it is being viewed on a mobile screen. And conversely, a small raster image starts to look grainy when displayed larger than its original size (a raster image is a set number of pixels wide and a set number of pixels tall, as we saw when we looked at vector graphics ). This is called the resolution switching problem .

Conversely, it is unnecessary to display a large image on a screen significantly smaller than the size it was meant for. Doing so can waste bandwidth; in particular, mobile users don't want to waste bandwidth by downloading a large image intended for desktop users, when a small image would do for their device. Ideally, multiple resolutions would be made available to the user's web browser. The browser could then determine the optimal resolution to load based on the screen size of the user's device.

To make things more complicated, some devices have high resolution screens that need larger images than you might expect to display nicely. This is essentially the same problem, but in a slightly different context.

You might think that vector images would solve these problems, and they do to a certain degree — they are small in file size and scale well, and you should use them wherever possible. However, they aren't suitable for all image types. Vector images are great for simple graphics, patterns, interface elements, etc., but it starts to get very complex to create a vector-based image with the kind of detail that you'd find in say, a photo. Raster image formats such as JPEGs are more suited to the kind of images we see in the above example.

This kind of problem didn't exist when the web first existed, in the early to mid 90s — back then the only devices in existence to browse the Web were desktops and laptops, so browser engineers and spec writers didn't even think to implement solutions. Responsive image technologies were implemented recently to solve the problems indicated above by letting you offer the browser several image files, either all showing the same thing but containing different numbers of pixels ( resolution switching ), or different images suitable for different space allocations ( art direction ).

注意: The new features discussed in this article — srcset / sizes / <picture> — are all supported in modern desktop and mobile browsers (including Microsoft's Edge browser, although not Internet Explorer.)

How do you create responsive images?

In this section, we'll look at the two problems illustrated above and show how to solve them using HTML's responsive image features. You should note that we will be focusing on <img> elements for this section, as seen in the content area of the example above — the image in the site header is only for decoration, and therefore implemented using CSS background images. CSS arguably has better tools for responsive design than HTML, and we'll talk about those in a future CSS module.

Resolution switching: Different sizes

So, what is the problem that we want to solve with resolution switching? We want to display identical image content, just larger or smaller depending on the device — this is the situation we have with the second content image in our example. The standard <img> element traditionally only lets you point the browser to a single source file:

<img src="elva-fairy-800w.jpg" alt="Elva dressed as a fairy">

						

We can however use two new attributes — srcset and sizes — to provide several additional source images along with hints to help the browser pick the right one. You can see an example of this in our responsive.html example on Github (see also the source code ):

<img srcset="elva-fairy-480w.jpg 480w,
             elva-fairy-800w.jpg 800w"
     sizes="(max-width: 600px) 480px,
            800px"
     src="elva-fairy-800w.jpg"
     alt="Elva dressed as a fairy">

						

srcset and sizes attributes look complicated, but they're not too hard to understand if you format them as shown above, with a different part of the attribute value on each line. Each value contains a comma-separated list, and each part of those lists is made up of three sub-parts. Let's run through the contents of each now:

srcset defines the set of images we will allow the browser to choose between, and what size each image is. Each set of image information is separated from the previous one by a comma. For each one, we write:

  1. An image filename ( elva-fairy-480w.jpg )
  2. A space
  3. The image's intrinsic width in pixels ( 480w ) — note that this uses the w unit, not px as you might expect. An image's intrinsic size is its real size, which can be found by inspecting the image file on your computer (for example, on a Mac you can select the image in Finder and press Cmd + I to bring up the info screen).

sizes defines a set of media conditions (e.g. screen widths) and indicates what image size would be best to choose, when certain media conditions are true — these are the hints we talked about earlier. In this case, before each comma we write:

  1. A media condition ( (max-width:600px) ) — you'll learn more about these in the CSS topic , but for now let's just say that a media condition describes a possible state that the screen can be in. In this case, we are saying "when the viewport width is 600 pixels or less".
  2. A space
  3. width of the slot the image will fill when the media condition is true ( 480px )

注意: For the slot width, you may provide an absolute length ( px , em ) or a length relative to the viewport ( vw ), but not percentages. You may have noticed that the last slot width has no media condition (this is the default that is chosen when none of the media conditions are true). The browser ignores everything after the first matching condition, so be careful how you order the media conditions.

So, with these attributes in place, the browser will:

  1. Look at its device width.
  2. Work out which media condition in the sizes list is the first one to be true.
  3. Look at the slot size given to that media query.
  4. Load the image referenced in the srcset list that has the same size as the slot or, if there isn't one, the first image that is bigger than the chosen slot size.

And that's it! At this point, if a supporting browser with a viewport width of 480px loads the page, the (max-width: 600px) media condition will be true, and so the browser chooses the 480px slot. The  elva-fairy-480w.jpg will be loaded, as its inherent width ( 480w ) is closest to the slot size. The 800px picture is 128KB on disk, whereas the 480px version is only 63KB — a saving of 65KB. Now, imagine if this was a page that had many pictures on it. Using this technique could save mobile users a lot of bandwidth.

注意: When testing this with a desktop browser, if the browser fails to load the narrower images when you've got its window set to the narrowest width, have a look at what the viewport is (you can approximate it by going into the browser's JavaScript console and typing in document.querySelector('html').clientWidth ). Different browsers have minimum sizes that they'll let you reduce the window width to, and they might be wider than you'd think. When testing it with a mobile browser, you can use tools like Firefox's about:debugging page to inspect the page loaded on the mobile using the desktop developer tools.

To see which images were loaded, you can use Firefox DevTools's Network Monitor tab.

Older browsers that don't support these features will just ignore them. Instead, those browsers will go ahead and load the image referenced in the src attribute as normal.

注意: <head> of the example linked above, you'll find the line  <meta name="viewport" content="width=device-width"> : this forces mobile browsers to adopt their real viewport width for loading web pages (some mobile browsers lie about their viewport width, and instead load pages at a larger viewport width then shrink the loaded page down, which is not very helpful for responsive images or design).

Resolution switching: Same size, different resolutions

If you're supporting multiple display resolutions, but everyone sees your image at the same real-world size on the screen, you can allow the browser to choose an appropriate resolution image by using srcset with x-descriptors and without sizes — a somewhat easier syntax! You can find an example of what this looks like in srcset-resolutions.html (see also the source code ):

<img srcset="elva-fairy-320w.jpg,
             elva-fairy-480w.jpg 1.5x,
             elva-fairy-640w.jpg 2x"
     src="elva-fairy-640w.jpg"
     alt="Elva dressed as a fairy">

						

A picture of a little girl dressed up as a fairy, with an old camera film effect applied to the image In this example, the following CSS is applied to the image so that it will have a width of 320 pixels on the screen (also called CSS pixels):

img {
  width: 320px;
}

						

在此情况下, sizes is not needed — the browser works out what resolution the display is that it is being shown on, and serves the most appropriate image referenced in the srcset . So if the device accessing the page has a standard/low resolution display, with one device pixel representing each CSS pixel, the elva-fairy-320w.jpg image will be loaded (the 1x is implied, so you don't need to include it.) If the device has a high resolution of two device pixels per CSS pixel or more, the elva-fairy-640w.jpg image will be loaded. The 640px image is 93KB, whereas the 320px image is only 39KB.

Art direction

To recap, the art direction problem involves wanting to change the image displayed to suit different image display sizes. For example, a web page includes a large landscape shot with a person in the middle when viewed on a desktop browser. When viewed on a mobile browser, that same image is shrunk down, making the person in the image very small and hard to see. It would probably be better to show a smaller, portrait image on mobile, which zooms in on the person. The <picture> element allows us to implement just this kind of solution.

Returning to our original not-responsive.html example, we have an image that badly needs art direction:

<img src="elva-800w.jpg" alt="Chris standing up holding his daughter Elva">

						

Let's fix this, with <picture> ! Like <video> and <audio> <picture> element is a wrapper containing several <source> elements that provide different sources for the browser to choose from, followed by the all-important <img> element. The code in responsive.html looks like so:

<picture>
  <source media="(max-width: 799px)" srcset="elva-480w-close-portrait.jpg">
  <source media="(min-width: 800px)" srcset="elva-800w.jpg">
  <img src="elva-800w.jpg" alt="Chris standing up holding his daughter Elva">
</picture>

						
  • <source> elements include a media attribute that contains a media condition — as with the first srcset example, these conditions are tests that decide which image is shown — the first one that returns true will be displayed. In this case, if the viewport width is 799px wide or less, the first <source> element's image will be displayed. If the viewport width is 800px or more, it'll be the second one.
  • srcset attributes contain the path to the image to display. Just as we saw with <img> above, <source> can take a srcset attribute with multiple images referenced, as well as a sizes attribute. So, you could offer multiple images via a <picture> element, but then also offer multiple resolutions of each one. Realistically, you probably won't want to do this kind of thing very often.
  • In all cases, you must provide an <img> element, with src and alt , right before </picture> , otherwise no images will appear. This provides a default case that will apply when none of the media conditions return true (you could actually remove the second <source> element in this example), and a fallback for browsers that don't support the <picture> 元素。

This code allows us to display a suitable image on both wide screen and narrow screen displays, as shown below:

Our example site as viewed on a wide screen - here the first image works OK, as it is big enough to see the detail in the center. Our example site as viewed on a narrow screen with the picture element used to switch the first image to a portrait close up of the detail, making it a lot more useful on a narrow screen

注意: 应该使用 media attribute only in art direction scenarios; when you do use media , don't also offer media conditions within the sizes 属性。

Why can't we just do this using CSS or JavaScript?

When the browser starts to load a page, it starts to download (preload) any images before the main parser has started to load and interpret the page's CSS and JavaScript. That mechanism is useful in general for reducing page load times, but it is not helpful for responsive images — hence the need to implement solutions like srcset . For example, you couldn't load the <img> element, then detect the viewport width with JavaScript, and then dynamically change the source image to a smaller one if desired. By then, the original image would already have been loaded, and you would load the small image as well, which is even worse in responsive image terms.

Use modern image formats boldly

New image formats like WebP and AVIF can maintain a low file size and high quality at the same time. These formats now have relatively broad browser support but little "historical depth".

<picture> lets us continue catering to older browsers. You can supply MIME types inside type attributes so the browser can immediately reject unsupported file types:

<picture>
  <source type="image/svg+xml" srcset="pyramid.svg">
  <source type="image/webp" srcset="pyramid.webp">
  <img src="pyramid.png" alt="regular pyramid built from four equilateral triangles">
</picture>

						
  • Do not 使用 media attribute, unless you also need art direction.
  • <source> element, you can only refer to images of the type declared in type .
  • Use comma-separated lists with srcset and sizes , as needed.

Active learning: Implementing your own responsive images

For this active learning, we're expecting you to be brave and go it alone ... mostly. We want you to implement your own suitable art directed narrow screen/wide screen shot using <picture> , and a resolution switching example that uses srcset .

  1. Write some simple HTML to contain your code (use not-responsive.html as a starting point, if you like).
  2. Find a nice wide screen landscape image with some kind of detail contained in it somewhere. Create a web-sized version of it using a graphics editor, then crop it to show a smaller part that zooms in on the detail, and create a second image (about 480px wide is good for this).
  3. 使用 <picture> element to implement an art direction picture switcher!
  4. Create multiple image files of different sizes, each showing the same picture.
  5. 使用 srcset / size to create a resolution switcher example, either to serve the same size image at different resolutions, or different image sizes at different viewport widths.

Test your skills!

You've reached the end of this article, but can you remember the most important information? You can find a detailed assessment that tests these skills at the end of the module; see Mozilla splash page .

摘要

That's a wrap for responsive images — we hope you enjoyed playing with these new techniques. As a recap, there are two distinct problems we've been discussing here:

  • Art direction : The problem whereby you want to serve cropped images for different layouts — for example a landscape image showing a full scene for a desktop layout, and a portrait image showing the main subject zoomed in for a mobile layout. You can solve this problem using the <picture> 元素。
  • Resolution switching : The problem whereby you want to serve smaller image files to narrow-screen devices, as they don't need huge images like desktop displays do — and to serve different resolution images to high density/low density screens. You can solve this problem using vector graphics (SVG images) and the srcset with sizes 属性。

This also draws to a close the entire 多媒体和嵌入 module! The only thing to do now before moving on is to try our multimedia assessment, and see how you get on. Have fun!

另请参阅

In this module

发现此页面有问题吗?

最后修改: , 由 MDN 贡献者

  1. Complete beginners start here!
  2. Web 快速入门
    1. Getting started with the Web overview
    2. 安装基本软件
    3. What will your website look like?
    4. 处理文件
    5. HTML 基础
    6. CSS 基础
    7. JavaScript 基础
    8. 发布您的网站
    9. How the Web works
  3. HTML — Structuring the Web
  4. HTML 介绍
    1. Introduction to HTML overview
    2. Getting started with HTML
    3. What's in the head? Metadata in HTML
    4. HTML text fundamentals
    5. Creating hyperlinks
    6. Advanced text formatting
    7. Document and website structure
    8. Debugging HTML
    9. Assessment: Marking up a letter
    10. Assessment: Structuring a page of content
  5. 多媒体和嵌入
    1. Multimedia and embedding overview
    2. Images in HTML
    3. Video and audio content
    4. From object to iframe — other embedding technologies
    5. Adding vector graphics to the Web
    6. Responsive images
    7. Assessment: Mozilla splash page
  6. HTML 表格
    1. HTML tables overview
    2. HTML table basics
    3. HTML Table advanced features and accessibility
    4. Assessment: Structuring planet data
  7. CSS — Styling the Web
  8. CSS 第一步
    1. CSS first steps overview
    2. What is CSS?
    3. Getting started with CSS
    4. How CSS is structured
    5. How CSS works
    6. Using your new knowledge
  9. CSS 构建块
    1. CSS building blocks overview
    2. Cascade and inheritance
    3. CSS 选择器
    4. The box model
    5. Backgrounds and borders
    6. Handling different text directions
    7. Overflowing content
    8. Values and units
    9. Sizing items in CSS
    10. Images, media, and form elements
    11. Styling tables
    12. Debugging CSS
    13. Organizing your CSS
  10. 样式化文本
    1. Styling text overview
    2. Fundamental text and font styling
    3. Styling lists
    4. Styling links
    5. Web fonts
    6. Assessment: Typesetting a community school homepage
  11. CSS 布局
    1. CSS layout overview
    2. Introduction to CSS layout
    3. Normal Flow
    4. Flexbox
    5. Grids
    6. Floats
    7. 位置
    8. Multiple-column Layout
    9. Responsive design
    10. Beginner's guide to media queries
    11. Legacy Layout Methods
    12. Supporting Older Browsers
    13. Fundamental Layout Comprehension
  12. JavaScript — Dynamic client-side scripting
  13. JavaScript 第一步
    1. JavaScript first steps overview
    2. What is JavaScript?
    3. A first splash into JavaScript
    4. What went wrong? Troubleshooting JavaScript
    5. Storing the information you need — Variables
    6. Basic math in JavaScript — Numbers and operators
    7. Handling text — Strings in JavaScript
    8. Useful string methods
    9. 数组
    10. Assessment: Silly story generator
  14. JavaScript 构建块
    1. JavaScript building blocks overview
    2. Making decisions in your code — Conditionals
    3. Looping code
    4. Functions — Reusable blocks of code
    5. Build your own function
    6. Function return values
    7. 事件介绍
    8. Assessment: Image gallery
  15. 引入 JavaScript 对象
    1. Introducing JavaScript objects overview
    2. Object basics
    3. 对象原型
    4. Object-oriented programming concepts
    5. Classes in JavaScript
    6. Working with JSON data
    7. Object building practice
    8. Assessment: Adding features to our bouncing balls demo
  16. 异步 JavaScript
    1. Asynchronous JavaScript overview
    2. General asynchronous programming concepts
    3. Introducing asynchronous JavaScript
    4. Cooperative asynchronous Java​Script: Timeouts and intervals
    5. Graceful asynchronous programming with Promises
    6. Making asynchronous programming easier with async and await
    7. Choosing the right approach
  17. 客户端侧 Web API
    1. 客户端侧 Web API
    2. Introduction to web APIs
    3. Manipulating documents
    4. Fetching data from the server
    5. Third party APIs
    6. Drawing graphics
    7. Video and audio APIs
    8. Client-side storage
  18. Web forms — Working with user data
  19. Core forms learning pathway
    1. Web forms overview
    2. Your first form
    3. How to structure a web form
    4. Basic native form controls
    5. The HTML5 input types
    6. Other form controls
    7. Styling web forms
    8. Advanced form styling
    9. UI pseudo-classes
    10. Client-side form validation
    11. Sending form data
  20. Advanced forms articles
    1. How to build custom form controls
    2. Sending forms through JavaScript
    3. CSS property compatibility table for form controls
  21. Accessibility — Make the web usable by everyone
  22. Accessibility guides
    1. Accessibility overview
    2. What is accessibility?
    3. HTML: A good basis for accessibility
    4. CSS and JavaScript accessibility best practices
    5. WAI-ARIA basics
    6. Accessible multimedia
    7. Mobile accessibility
  23. Accessibility assessment
    1. Assessment: Accessibility troubleshooting
  24. Tools and testing
  25. Client-side web development tools
    1. Client-side web development tools index
    2. Client-side tooling overview
    3. Command line crash course
    4. Package management basics
    5. Introducing a complete toolchain
    6. Deploying our app
  26. Introduction to client-side frameworks
    1. Client-side frameworks overview
    2. Framework main features
  27. React
    1. Getting started with React
    2. Beginning our React todo list
    3. Componentizing our React app
    4. React interactivity: Events and state
    5. React interactivity: Editing, filtering, conditional rendering
    6. Accessibility in React
    7. React resources
  28. Ember
    1. Getting started with Ember
    2. Ember app structure and componentization
    3. Ember interactivity: Events, classes and state
    4. Ember Interactivity: Footer functionality, conditional rendering
    5. Routing in Ember
    6. Ember resources and troubleshooting
  29. Vue
    1. Getting started with Vue
    2. Creating our first Vue component
    3. Rendering a list of Vue components
    4. Adding a new todo form: Vue events, methods, and models
    5. Styling Vue components with CSS
    6. Using Vue computed properties
    7. Vue conditional rendering: editing existing todos
    8. Focus management with Vue refs
    9. Vue resources
  30. Svelte
    1. Getting started with Svelte
    2. Starting our Svelte Todo list app
    3. Dynamic behavior in Svelte: working with variables and props
    4. Componentizing our Svelte app
    5. Advanced Svelte: Reactivity, lifecycle, accessibility
    6. Working with Svelte stores
    7. TypeScript support in Svelte
    8. Deployment and next steps
  31. Angular
    1. Getting started with Angular
    2. Beginning our Angular todo list app
    3. Styling our Angular app
    4. Creating an item component
    5. Filtering our to-do items
    6. Building Angular applications and further resources
  32. Git and GitHub
    1. Git and GitHub overview
    2. Hello World
    3. Git Handbook
    4. Forking Projects
    5. About pull requests
    6. Mastering Issues
  33. Cross browser testing
    1. Cross browser testing overview
    2. Introduction to cross browser testing
    3. Strategies for carrying out testing
    4. Handling common HTML and CSS problems
    5. Handling common JavaScript problems
    6. Handling common accessibility problems
    7. Implementing feature detection
    8. Introduction to automated testing
    9. Setting up your own test automation environment
  34. Server-side website programming
  35. 第一步
    1. First steps overview
    2. Introduction to the server-side
    3. Client-Server overview
    4. Server-side web frameworks
    5. Website security
  36. Django Web 框架 (Python)
    1. Django web framework (Python) overview
    2. 介绍
    3. 设置开发环境
    4. Tutorial: The Local Library website
    5. Tutorial Part 2: Creating a skeleton website
    6. Tutorial Part 3: Using models
    7. Tutorial Part 4: Django admin site
    8. Tutorial Part 5: Creating our home page
    9. Tutorial Part 6: Generic list and detail views
    10. Tutorial Part 7: Sessions framework
    11. Tutorial Part 8: User authentication and permissions
    12. Tutorial Part 9: Working with forms
    13. Tutorial Part 10: Testing a Django web application
    14. Tutorial Part 11: Deploying Django to production
    15. Web application security
    16. Assessment: DIY mini blog
  37. Express Web Framework (node.js/JavaScript)
    1. Express Web Framework (Node.js/JavaScript) overview
    2. Express/Node introduction
    3. Setting up a Node (Express) development environment
    4. Express tutorial: The Local Library website
    5. Express Tutorial Part 2: Creating a skeleton website
    6. Express Tutorial Part 3: Using a database (with Mongoose)
    7. Express Tutorial Part 4: Routes and controllers
    8. Express Tutorial Part 5: Displaying library data
    9. Express Tutorial Part 6: Working with forms
    10. Express Tutorial Part 7: Deploying to production
  38. Further resources
  39. Common questions
    1. HTML questions
    2. CSS questions
    3. JavaScript questions
    4. Web mechanics
    5. Tools and setup
    6. Design and accessibility