In approaching this topic I considered many of the driving factors from the apparent misunderstanding of video buffering and compression for suitable low latent playback. However in organizing my never before documented axioms on this, I quickly realized I should touch base on optimization and rightsizing for all forms of media.
E / P = X2 … formulaic on video encoding strategy where:
“ Encoding “
(( Resolution px + Audio kbps ) + Bitrate X TimeLength / Transcoding Speed)
__________________________ = “ Viewing Experience “
(OVER) (DIVIDED BY)
“ Playback “
( Bandwidth / (Buffering Performance + Local Playback Scaler ))
Mastermind video editing software considerations to optimize online broadcast BEFORE source upload :
- Maintain audio bit consistency across a series or “channel library” of videos.
- Budget ample production time for slower transcoding to empower advanced file compression.
- Locally encode your post-productino videos at 15 Mbps bitrate optimized for cloud uploading.
- Reconsider to ever allow online based down-sampling from your original uploaded source files.
- Drop all media framerates to 30fps before uploading to avoid re-encoding degradation.
Gross Generalizations to attempt an approach for Viewer Normalization :
- Estimated 60% viewers will watch video in 720p, great for fullscreen web playback
- Other majority of audience will experience 1080p, appropriate for connected TVs.
- Majority of Marketshare Devices rescale at a divisible of 16 for compressions; thus championing 720p.
- To avoid digital artifacting reduce your streaming solutions rescaling as much as possible.
Device Playback Index
- iPod < 320p : 1.5 Mbps
- iPadI < 720p
- iPadII < IV 1080p
- iPhone4 = Baseline H.264 profile
- Most standard definition camcorders resoluton 640 X 480
- Xbox360 is 1080p (1920p wide x 1080 high)
- Windows 7 Phone is 480 x 800 H.264 / HTML5
YouTube (optimized) suggested specs :
- Codec : MPEG-4 video format via DivX or XviD
- Audio : MP3 format
- Resolution :
- Most Instructional videos 480p ( 640 pixels wide by 480 pixels high )
- Extended MOOC lectures when visual aid is less important than oral 320p “HVGA”
- ( 320p would be 320 pixels wide x 240 pixels high : one-half “VGA” 640 x 480 video )
- “HD720p” When visual education is critical in an instructional video use 720p ( 1280 wide x 720 high )
- Frame-rate : 30fps or drop frames down to as low as 15fps for long audio lectures w/ less critical video
- Timelength : 15 mins or validate account for extended longer timelengths
- File-Size : upto 128 GBs as a validated account
Other general considerations and tips
- Avoid future streaming service compression changes; encode Library Repository sourcefiles at 1080p / 50 Mbps as complex content may warrant higher bitrate stored.
- For most common needs stay with MAIN level of H.264 profile for compatability though HIGH offers better compression.
- Before recording consider pillar/letterboxing needs and use of 16:9 ratio ( divisible by 8 not leaving to devices to rescale as suggested at 16 division to expedite calculations)
- Old “analog” TV formats 640×480 is 4:3 standard 480i NTSC broadcast of unsuare pixels on a computer monitor. Conventional Digital Broadcast is 16:9 and 8 square pixels is native.
- Remember any monitor “resolution” is independent of displayed viewport media; as GUI normally displays specified embedded playback aspect ratio, whatever your computer screen resolution.
- When deciding resolution remember to retain aspect ratio of original capture.
Degradation warnings :
- YouTube re-encoding is problematic with green backgrounds and all YouTube videos stream limits upto 30 fps quality so any greater is bloated wasteful processing.
- YouTube enables HDTV broadcast widescreen 1280X720 pixels
- Flash Player, does NOT offer the best H.264 decoder.
- Dont use the Web 352p suggestion as it is unconventional.
As with any website there is always something to be done… The daily beefing-UP of meta rich content while minimizing page load and strategize a normalizing web cache.
I explored the revamped HealthCare.gov and for me I STILL see many areas where a budgeted $300 MILLION website could improve!
The government funded optimization team immediately began in what ACC defines as “crunch mode” normalized top anticipated web-assets and reallocated hardest hitting performance then onto web-cache libraries
Image asset optimization is simple and as such c
Leverage dark data to gain previously uncovered insights that you can take action on to improve your customer retention and maximize your service’s marginal utility ceiling.
Im my opinion there is only great theory on this not much is really capable to take advantage of as each dual thread is in concept intended . in the end CPU brings it to one and there still lies your original bottleneck.
Web caching is one of the most misunderstood technologies on the Internet.
Distributed Reflectance and other hardware resources, local and in the cloud like AWS CloudFront.
and in hardware the more CPU-head Disc “hammers” you have performing actual writes the more sound this systen is than if you payed out of the wazzoo to get a faster.
Tell me as you are reading this now, how old is your last web status update!