ABC iview is a video on demand service from the Australian Broadcasting Corporation. In 2015, we built a new service in-house to handle transcoding for all of iview's content. Metro is built on AWS, and uses FFmpeg, node.js and go components to transcode content quickly and cost-efficiently. Presented at GOTO Chicago on 2nd May 2017

1. Video Transcoding at the
ABC with microservices
Daphne Chong
@daphnechong

6. media
transcoder

7. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future roadmap

8. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future roadmap

10. What changes
are we making?

13. High speed
Medium speed
Low speed

14. Standard Encoding
H.264 MP4

15. Transcoding
Example

16. Step 1: Prepare watermark

17. Step 1: Prepare watermark

18. Step 1: Prepare watermark

19. Step 2: Input video (GXF)

20. Step 3: Asymmetrical cropping

21. Step 4: Scale and set aspect ratio

22. Step 5: Apply watermark

23. Step 6: Audio Normalization

24. Original file audio track

25. After audio normalization

26. Step 7: Combine changes

27. Step 8: Split into segments

28. That’s a lot of
changes...

29. Transcoding
Profiles

30. GXF
Profile:
- Renditions (6)
- Crop
- Normalize audio
- Adjust ratio
- Add watermark
- GXF to MP4

31. GXF

32. GXF
MOV
Profile:
- Renditions (4)
- Adjust ratio
- ProRes MOV to MP4

33. GXF
MOV

35. ./ffmpeg -i input.gxf output.mp4

36. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

37. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

38. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

39. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

40. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

41. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

42. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

43. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

44. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

45. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

46. ./ffmpeg -i INPUT -c:v libx264 -filter_complex
"[0:v]crop=720:576:0:32,yadif,scale=1024x576,setdar=16/9[ma
in], [1:v]scale=54:-1[ovrl],
[main][ovrl]overlay=main_w-overlay_w-47:main_h-overlay_h-44
" -crf 22 -maxrate 1404k -bufsize 1500k -pix_fmt yuv420p
-profile:v high -level 4.0 -refs 3 -r 25 -g 50 -c:a
libfdk_aac -b:a 96k -filter_complex "[0:a:0][0:a:1]
amerge=inputs=2,
dynaudnorm=f=500:g=31:p=0.82:m=30.0:r=0.0:s=0.0 [aout]" -ar
44100 -ac 2 -map 0:0 -map "[aout]" -movflags +faststart
-threads 0 -y OUTPUT
In reality, it looks like this

47. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future roadmap

48. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future roadmap

49. Workflow

65. 8 minutes 27 seconds

66. Original
length
Time to produce rendition
60 min
7 min
42 min
Lowest
Quality
Highest
Quality

67. AWS & Languages

68. Languages:
Node.js
Go

71. RDS - state storage
EC2 - instances
SQS - messaging
S3 - file storage
SNS - notifications
AWS Infrastructure:

77. Advantages
Deployments
Independent scaling
Really easy to expand
Clear responsibilities

78. Challenges
Deployments
General overhead

79. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future optimizations

80. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future optimizations

82. When should
you scale?

83. Scaling Considerations:
Content priority
Ad-hoc requests
Varied content length

84. AWS Considerations:
Hourly billing
Cost vs Speed
Fleet reaction time
Scaling down

85. MVP:
Don’t make things wait
too long

86. MVP:
Supply enough
capacity at all times

87. GXF
Profile:
- Renditions (6)
- Crop
- Normalize audio
- Adjust ratio
- Add watermark
- GXF to MP4

93. var autoscaling = new AWS.AutoScaling(...);
var params = {
AutoScalingGroupName: groupName,
DesiredCapacity: capacity
Some white text
};
autoscaling.setDesiredCapacity(params, ...);

94. var autoscaling = new AWS.AutoScaling(...);
var params = {
AutoScalingGroupName: groupName,
DesiredCapacity: capacity
Some white text
};
autoscaling.setDesiredCapacity(params, ...);

95. Which instance size?

96. Time vs Cost
for 1 hour of content
Time Cost
?

97. Time vs Cost
for 1 hour of content
Time Cost
?

98. Winner:
c4.xlarge

99. When to scale down?

100. Up: step scaling
Down: stop everything

106. An interesting puzzle :)

107. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future optimizations

108. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future optimizations

109. What about
commercial transcode
providers?

110. Full-Service
Transcoding

111. $(a lot)

112. DIY on-demand

114. $45k

115. Metro’s costs?

116. $11.2k

117. Some daily examples

118. (10am - 10am)

119. $42.86

120. EC2 Instances: $34.31
EC2 (other): $ 4.24
RDS: $ 2.66
S3: $ 1.65
SQS: $ 0.00
Total (USD): $42.86
Transcoding costs: 24 hrs

121. Next 24 hours

122. $25.93

124. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future optimizations

125. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future optimizations

126. Throughput
optimizations

129. Transcoding in
chunks/segments

131. Introspection &
Validation

132. GXF
Profile:
- Crop
- Normalize audio
- Adjust ratio
- Add watermark
- GXF to MP4

133. ffprobe

137. 1. What is Transcoding?
2. Architecture
3. Autoscaling transcoders
4. Costs
5. Future optimizations

138. Summary

143. 8 minutes 27 seconds

144. $11.2k

146. @daphnechong
Thanks!