Based on customer demand and feedback, starting January 1, 2017, we will implement a new pricing model for on-demand encoding of media with Azure Media Services. For the Standard Encoder and Premium Encoder, we will be calculating usage based on the total duration of the media files produced as output by the encoder. Further, for Media Reserved Units, we are lowering the rate, and making a change to calculate usage on a per-minute basis. In this post, I’ll be walking you through some examples to show how the new pricing models work.
New per-output-minute model for encoding
Starting on January 1, 2017, we will bill for encoding jobs based on the duration of the media files produced as output by the encoder. No code changes are needed in your application, as the new model will be automatically applied by our service. Please refer to the official pricing page after January 1, 2017, in order to verify the final price for each data center.
Calculation of output minutes
To calculate the total output minutes for an encoding task, we make use of the following multipliers:
Multipliers for calculating output minutes
Quality
Multiplier
Example
Audio only output
0.25x
20 minutes of audio count as 5 SD minutes
SD (less than 1280×720)
1x
20 minutes
HD (1280×720 – 1920×1080)
2x
20 minutes of HD output count as 40 SD minutes
UHD (up to 4096×2160)
4x
20 minutes of UHD output count as 80 SD minutes
Pricing example for Standard Encoder
The Standard Encoder is the best option if your goal is to transcode a wide variety of input video/audio files into an output format suitable for playing back on a variety of devices (smartphones, tablets, PCs, consoles, TVs). You can read more about the capabilities of this (Media Encoder Standard) media processor in our documentation.
To determine the overall cost for using this encoder, you need to know the duration of your input video and its resolution, and the encoding preset. Suppose you have a high-quality QuickTime video at 1920x1080p resolution, which is 20 minutes in duration. If you want to encode this 1080p input for delivery via adaptive streaming protocol to a variety of devices, then you would typically use a preset such as the “H264 Multiple Bitrate 1080p.” Using such a preset will incur billing for each output layer/bitrate, along with a multiplier depending on the video resolution. For this preset, the following multipliers apply:
Layer
Resolution
Multiplier
HD Video 1
1920×1080
2x
HD Video 2
1920×1080
2x
HD Video 3
1280×720
2x
SD Video 1
960×540
1x
SD Video 2
960×540
1x
SD Video 3
640×360
1x
SD Video 4
640×360
1x
SD Video 5
320×180
1x
Audio
N/A
0.25x
Total
11.25x
Based on the table above, your encoding of the QuickTime video will result in a total of 11.25 * 20 = 225 output minutes. After January 1, 2017, you can refer to the official pricing page for the current rates for one output minute, applicable to your data center, and multiply that by 225 to determine the final cost.
Pricing example for Premium Encoder
If your objective is to transcode between formats common to the broadcast or movie industry, or if your video workflow requires complex logic, then you would need the Premium Encoder. See our documentation for an in-depth comparison of the feature differences between the two encoders. Additionally, this blog offers a flow chart to help choose between the two encoders. Suppose you have a high quality ProRes/QuickTime video that is 20 minutes in duration, with 1920x1080p video resolution but two audio tracks – one in English and the other in Spanish. You can use the Premium Encoder to transcode this video into a single bitrate MXF file, with one video track and two audio tracks. In this case, the multiplier would be 2 (for HD video) plus 2*0.25 (for each audio) adding up to 2.5. Thus, your encoding of the QuickTime video into an MXF file would result in a total of 20*2.5 = 50 output minutes. After January 1, 2017, you can refer to the official pricing page for the current rates for one output minute, applicable to your data center, and multiply that by 50 to determine the final cost.
Per-hour pricing for Media Reserved Units
You need to add Media Reserved Units (MRUs) to your Media Service account if your workload requires one or more videos to be processed concurrently. You can increase the overall throughput from the service by (a) increasing number of MRUs to get more videos processed concurrently, and (b) by using faster MRUs (eg. S3). See the documentation for more information. Based on customer demand and feedback, we are changing the pricing model for MRUs to more closely track the actual usage of such units. As of January 1, 2017, you will be charged based on the time each MRU is active in your accounted, pro-rated on a per-minute basis.
Example of MRU Pricing
Here is an example, that shows that your charges will be based on actual minutes of usage of MRUs. Suppose you had zero MRU to begin with, and at 10:00 am on a particular day, you set your account to use 2 S1 MRUs. More videos arrive in the afternoon, so you change your account to use 4 S3 RUs at 1:15 pm. All the videos are processed by 4:00 pm, and then you turn off the MRUs in his account (set number of MRUs to zero). Your usage, for that day, is calculated as follows.
S1 Media Reserved Units: 2 units * 3.25 hours (10AM to 1:15PM) = 7.5 S1 hours
S3 Media Reserved Units: 4 units * 2.75 hours (1:15PM to 4PM) = 11 S3 hours
After January 1, 2017, you can refer to the official pricing page for the current rates for one S1 hour, and one S3 hour, and multiply them by 7.5 and 11 respectively, and add the results to get the total cost.
Contact us
Keep monitoring the Azure Media Services blog for updates on encoding capabilities.
Send your feedback and feature requests to our UserVoice page.
Quelle: Azure
Published by