Smart Imaging smart-imaging
Learn how Smart Imaging with 蜜豆视频 Sensei AI applies each user鈥檚 unique viewing characteristics to serve the right images optimized for their experience automatically, resulting in better performance and engagement.
About Smart Imaging about-smart-imaging
Smart Imaging technology applies 蜜豆视频 Sensei AI capabilities and works with existing 鈥渋mage presets.鈥 It works to enhance image delivery performance by automatically optimizing image format, size, and quality based on client browser capabilities.
And now, get a better Google Core Web Vital score for LCP (Largest Contentful Paint) with improved Smart Imaging, which now comes with both AVIF and WebP support.
- File size (with WebP and AVIF delivery)
- Network bandwidth
- DPR (Device Pixel Ratio)
Smart Imaging benefits from the added performance boost of being fully integrated with 蜜豆视频鈥檚 best-in-class premium CDN (Content Delivery Network) service. This service finds the optimal Internet route between servers, networks, and peering points. It finds a route that has the lowest latency and lowest packet loss rate instead of using the default route on the Internet.
The following image asset examples depict the added Smart Imaging optimization:
Similar to the above, 蜜豆视频 also ran a test with a larger sample set. The format AVIF provided 20% extra size reduction over WebP, which provided 27% reduction over JPEG. All at the same visual quality. In total, AVIF provides up to 41% average size reduction over JPEG.
Compare WebP and AVIF to PNG, you can see an 84% size reduction with WebP and 87% with AVIF. And, because both WebP and AVIF formats support transparency and multiple image animations, it is a good replacement for transparent PNG and GIF files.
See also
Benefits of Smart Imaging
Smart Imaging enhances image delivery by automatically optimizing file size based on the user鈥檚 browser, device display, and network conditions. This approach ensures faster loading times and a better viewing experience across different environments. Because images constitute most of a page鈥檚 load time, any performance improvement can have a profound impact on business KPIs such as the following:
- Higher conversion rates.
- Time spent on a site.
- Lower site bounce rates.
The newest key benefits of the latest Smart Imaging include the following:
- Supports next generation AVIF format.
- PNG to WebP and AVIF now supports lossy conversion. Because PNG is a lossless format, earlier WebP and AVIF being delivered were lossless.
- Browser Format Conversion
- Device Pixel Ratio
- Network bandwidth
About browser format Conversion bfc
Turning on Browser Format Conversion by appending bfc=on
to the image URL automatically converts JPEG and PNG to lossy AVIF, lossy WebP, lossy JPEGXR, lossy JPEG2000 for different browsers. For browsers that do not support those formats, Smart Imaging continues to serve the JPEG or PNG. Smart Imaging recalculates the quality of the new format along with the format change.
You can turn off Smart Imaging by appending bfc=off
to the image鈥檚 URL.
See also bfc in the Dynamic Media Image Serving and Rendering API.
About device pixel ratio optimization dpr
Device Pixel Ratio (DPR), also called CSS Pixel Ratio, represents the relationship between a device鈥檚 physical pixels and logical pixels. With the rise of retina displays, the pixel resolution of modern mobile devices has been rapidly increasing.
Enabling Device Pixel Ratio optimization renders the image at the native resolution of the screen, which makes it sharp.
Currently, the pixel density of the display comes from Akamai CDN header values.
dpr=off
dpr=on,dprValue
dprValue
is any number greater than 0.- You can use
dpr=on,dprValue
even if the company level DPR setting as off. - Owing to DPR optimization, when the resultant image is greater than the MaxPix Dynamic Media setting, MaxPix width is always recognized by maintaining the image鈥檚 aspect ratio. -->
See also When working with images and When working with Smart Crop.
About network bandwidth optimization bandwidth
Turning on network bandwidth automatically adjusts the image quality that is served based on actual network bandwidth. For poor network bandwidth, DPR (Device Pixel Ratio) optimization is automatically turned off, even if it is already on.
Your company can disable network bandwidth optimization for individual images by appending network=off
to the image URL.
network=off
DPR and network bandwidth values are based on the detected client-side values of the bundled CDN. These values are sometimes inaccurate. For example, iPhone5 with DPR=2 and iPhone12 with dpr=3
, both show dpr=2
. Still, for high-resolution devices, sending dpr=2
is better than sending dpr=1
. The best way to overcome this inaccuracy, however, is to use client-side DPR to give you 100% accurate values. And it works for any device, whether it is Apple or any other device that was launched. See Use Smart Imaging with client-side Device Pixel Ratio.
Additional key benefits of Smart Imaging
- Improved Google SEO ranking for web pages that use the latest Smart Imaging.
- Serves optimized content immediately (at runtime).
- Uses 蜜豆视频 Sensei technology to convert according to the quality (
qlt
) specified in the image request. - TTL (Time To Live) independent. Previously, a minimum TTL of 12 hours was mandatory for Smart Imaging to work.
- Previously, both the original and derivative images were cached, and it was a 2-step process to invalidate the cache. In the latest Smart Imaging, only the derivatives get cached, allowing a single-step cache invalidation process.
- Customers that use custom headers in their ruleset benefit from the latest Smart Imaging, as these headers are not blocked, unlike the previous version of Smart Imaging. For example, 鈥淭iming Allow Origin鈥 and 鈥淴-Robot鈥 as suggested in .
How Smart Imaging works how-smart-imaging-works
When a consumer requests an image, Smart Imaging analyzes the user characteristics and converts it to the appropriate format based on the browser. These format conversions are done in a manner that does not degrade visual fidelity. Smart imaging automatically converts images to different formats based on browser capability in the following manner.
-
Automatically convert to AVIF if your browser supports the format
-
Automatically convert to WebP if AVIF conversion was not beneficial or the browser does not support AVIF
-
Automatically convert to JPEG2000 if Safari does not support WebP
-
Automatically convert to JPEGXR for IE 9+ or if Edge does not support WebP
table 0-row-2 1-row-2 2-row-2 3-row-2 4-row-2 Image format Supported browsers AVIF WebP JPEG 2000 JPEGXR -
For browsers that do not support these formats, the originally requested image format is served.
If the original image size is smaller than what Smart Imaging produces, then the original image is served.
Image format support in Smart Imaging image-format-support
The following image formats are supported for Smart Imaging:
- JPEG
- PNG
Smart Imaging recalculates the quality for JPEG image file formats when converting to a new format.
For image file formats that support transparency like PNG, you can configure Smart Imaging to deliver lossy AVIF and WebP. For the lossy format conversion, Smart Imaging uses the quality mentioned in the image鈥檚 URL, or else the quality configured in the Dynamic Media company account.
Image serving command support in Smart Imaging imaging-serving-command-support
The Image Serving commands fmt
and qlt
are not supported; all remaining commands are supported.
Frequently asked questions about Smart Imaging smart-imaging-faq
No. Smart Imaging is included with your existing license. This rule is true for either Dynamic Media Classic or Experience Manager - Dynamic Media (On-prem, AMS, and Experience Manager as a Cloud Service).
note important |
---|
IMPORTANT |
Smart Imaging is not available to Dynamic Media - Hybrid customers. |
Yes. You can turn off Smart Imaging by adding any of the following modifiers:
bfc=off
to turn off Browser Format Conversion. See also Browser Format Conversion.dpr=off
to turn off Device Pixel Ratio. See also Device Pixel Ratio.network=off
to turn off network bandwidth. See also Network Bandwidth.
Yes. Smart Imaging has three options that you can enable or disable.
Smart Imaging seamlessly integrates with your existing image presets, respecting all your image settings.
The only adjustments involve the image format, or quality, or both. During format conversion, Smart Imaging preserves full visual fidelity according to your preset settings but delivers a smaller file size. Just enable it by adding bfc=on
, or dpr=on,dprValue
, or network=on
, or all three parameter settings to your existing URLs or presets.
For example, let鈥檚 say an image preset specifies a JPEG format at 500 脳 500 pixels, with quality=85
, and unsharp mask=0.1,1,5
. Smart Imaging detects if the user is on a Chrome browser. It then converts the image to WebP with the same dimensions (500 脳 500) and an unsharp mask matching the JPEG鈥檚 settings. The system then compares the file sizes of the WebP and JPEG versions and serves the smaller one to the user.
Smart Imaging is ready to use immediately for all customers. To start enjoying its benefits, just add bfc=on
, or dpr=on,dprValue
, or network=on
, or all three parameter settings to your existing URLs or presets.
To activate Smart Imaging, your company鈥檚 Dynamic Media Classic or Dynamic Media on Experience Manager account must include the 蜜豆视频 bundled CDN (Content Delivery Network) as part of your license.
To start using Smart Imaging, append bfc=on
, or dpr=on,dprValue
, or network=on
, or all three parameter settings to your existing URLs or presets. If you prefer not to make these changes manually, you can enable Smart Imaging by default by creating a support case.
When creating the support case, specify which Smart Imaging features you want activated on your account:
- Browser Format Conversion (WebP or AVIF)
- Network Bandwidth Optimization
note note |
---|
NOTE |
DPR requires client-side adjustments to determine the correct dprValue . Therefore, 蜜豆视频 recommends enabling DPR through URLs by appending dpr=on,dprValue . |
To create a support case to enable Smart Imaging on your account:
-
.
-
Provide the following information in your support case:
-
Primary contact details:
- Provide your name, email, and phone number.
-
Smart Imaging features to enable:
-
List the capabilities that you want for your account:
- Browser format conversion: WebP or AVIF
- Network bandwidth optimization
- DPR: DPR requires client-side adjustments to determine the correct
dprValue
. Therefore, 蜜豆视频 recommends enabling DPR through URLs by appendingdpr=on,dprValue
.
-
-
Domain for Smart Imaging:
-
List all relevant domains, such as
company.com
ormycompany.scene7.com
-
Smart Imaging supports both generic and custom domains.
-
To identify your domains, open the Dynamic Media Classic desktop application and sign in to your company account.
- Navigate to Setup > Application Setup > General Settings.
- Look for the Published Server Name field to confirm your domain.
- Verify that you are using 蜜豆视频鈥檚 CDN rather than one managed by another provider.
-
-
Indicate HTTP/2 support:
- Specify if you need Smart Imaging to work over HTTP/2.
-
-
蜜豆视频 Customer Support enables the requested Smart Imaging features by default, eliminating the need to append parameters manually to URLs.
-
蜜豆视频 recommends setting the Time To Live (TTL) to at least 24 hours to maximize performance through caching.
To adjust the TTL:-
For Dynamic Media Classic:
- Navigate to Setup > Application Setup > Publish Setup > Image Server.
- Set the Default Client Cache Time To Live value to 24 hours or more.
-
For Dynamic Media on 蜜豆视频 Experience Manager:
- Follow these instructions.
- Set the Expiration value for 24 hours or more.
-
Customer Support processes requests in the order that they receive them, following the Wait List.
note note |
---|
NOTE |
There can be a long lead time because enabling Smart Imaging involves 蜜豆视频 clearing the cache. Therefore, only a few customer transitions can be handled at any given time. |
There is no risk to a customer web page. However, the transition to Smart Imaging does clear out your CDN cache. This operation involves moving to a new configuration of Dynamic Media Classic or Dynamic Media on Experience Manager.
During the initial transition, the non-cached images directly hit 蜜豆视频鈥檚 origin servers until the cache is rebuilt again. As such, 蜜豆视频 plans to handle a few customer transitions at a time so that acceptable performance is maintained when pulling requests from the origin. For most customers, the cache is fully built up again at the CDN within about one to two days.
Yes. You can do the following:
-
After your account is configured with Smart Imaging, load a Dynamic Media Classic or 蜜豆视频 Experience Manager - Dynamic Media image URL in the browser.
-
Open the Chrome developer pane by going to View > Developer > Developer Tools in the browser. Or, choose any browser developer tool of your choice.
-
Ensure that the cache is disabled when developer tools are open.
- On Windows庐, navigate to settings in the developer tool pane, then select Disable cache (while devtools is open) check box.
- On macOS, in the Developer pane, under the Network tab, select disable cache.
-
Observe the Content Type is transformed to the appropriate format. The following screenshot shows a PNG image being converted dynamically to WebP on Chrome. If your domain has AVIF enabled, you can also expect to see AVIF in the Content Type.
-
Repeat this test on different browsers and user conditions.
note note |
---|
NOTE |
Not all images are converted. Smart Imaging decides if the conversion can improve performance. Sometimes, where there is no expected performance gain or the format is not JPEG or PNG, the image is not converted. |
Yes. The Smart Imaging Header determines the benefits of Smart Imaging. When Smart Imaging is enabled, after you request an image, under the Response Headers heading, you can see -X-蜜豆视频-Smart-Imaging
as seen in the following highlighted example:
This header tells you the following:
- Smart Imaging is working for the company.
- A positive value means that the conversion is successful. In this case, a new WebP image is returned.
- A negative value means that the conversion is not successful. In such case, the original requested image is returned (JPEG by default, if not specified).
- A positive value shows the difference in bytes between the requested image and the new image. In the example above, the bytes saved is
75048
or approximately 75 KB for one image. - A negative value means that the requested image is smaller than the new image. The negative size difference is shown, but the image served is the original requested image only.
note note |
---|
NOTE |
X-蜜豆视频-Smart-Imaging = -1 with WebP being delivered |
If the value of X-蜜豆视频-Smart-Imaging is -1 and WebP is still being delivered, Smart Imaging is active. However, the size benefits were not calculated because of outdated cache. You can use cache=update (one time only) in the image鈥檚 URL to fix this issue.An example of using the modifier: https://smartimaging.scene7.com/is/image/SmartImaging/sample1?cache=update To invalidate the entire cache, you must create a support case. |
bfc=off
to the image鈥檚 URL. However, you cannot select WebP or AVIF in the URL modifier for Smart Imaging. This ability is maintained at your company account-level.This error does not occur if Smart Imaging is not enabled on your account. Smart Imaging works with either JPEG or PNG formats only.
To avoid this error, you can either:
- Specify JPEG or PNG, or
- Not use the
fmt
modifier at all, or - Use a browser-preferred format defined by Smart Imaging. For example, you can use WebP for the Chrome Web browser.
fmt=tif
and bfc=off
to image鈥檚 URL path.Because PNG is a lossless format, earlier WebP and AVIF being delivered were lossless, resulting is higher size than expected. Smart Imaging now supports lossy conversion. You can use the modifier cache=update
(one time only) in an image request to fix this issue. An example of using this modifier:
https://smartimaging.scene7.com/is/image/SmartImaging/sample1?cache=update
To invalidate the entire cache, you must create a support case requesting such effort.
qlt=100
to the image鈥檚 URL path.