[mythtv-users] Concert footage and video artifacts

[Michael T. Dean]

On 11/18/2016 12:33 PM, Ian Evans wrote:
> On Nov 17, 2016 4:24 PM, "Jay Foster" wrote:
>> On 11/17/2016 12:40 PM, Ian Evans wrote:
>>> This isn't a pressing issue, but more a "just curious, learn something
>>> everyday" type of thing.
>>>
>>> Sometimes when watching a music awards show or concert footage, the
>>> picture can go crazy if there's a quick change of lighting or fast strobing
>>> effects. Suddenly the picture goes from the crystal clear bronze glow of
>>> the Zildjian cymbals and chrome on the drum set to a blocky visual worthy
>>> of an 8-pixel video game emulator.
>>>
>>> Is that how it's coming from the source? Is it a temporary deinterlacing
>>> issue caused by elves? Will the right video card prevent it?
>>>
>>> Curious minds want to know. :-)
>>>
>> That is compression artifacts.  Usually due to insufficient bandwidth to
>> encode the video.  This is typically introduced by the local broadcaster,
>> most of which choose quantity (i.e., more subchannels) over quality (i.e.,
>> a higher bit rate).
> Oddly I usually see it on the Grammys, on an OTA channel with zero
> subchannels.

Some of the "celebration" scenes in awards shows, etc.--especially those 
with tons of flashing/moving lights and glitter "snow" being shot in the 
air or similar--require tremendously high bitrates to encode such that 
it looks correct (I wouldn't be surprised if transmitting only 
intra-coded pictures (I-frames--basically fully-specified images) and 
not even using the "bit-saving" predicted (P-frames/delta-frames) or 
bi-predictive (B-frames) pictures during some such scenes would take 
fewer bits).  While encoders try to account for change-heavy scenes 
using variable-bitrate encoding (giving higher bitrate to more 
complex/faster-changing scenes, as required), the maximum bitrate limits 
applied are generally below what's required for these "chaos" scenes.  
The maximum bitrate limits may be applied to account for a variety of 
things beyond just making room for subchannels--including keeping the 
bitrate low enough that decoders "in the wild" can generally keep up 
with the decoding (where not every TV in use today has a 2016 
decoder--such as my ~2005 HLR-6768W, which struggles and drops frames 
even in low bitrate situations, but no worries since MythTV is the 
decoder I'm using, and it does fine thanks to a good frontend).  So, in 
general, most broadcasters decide that a short section of garbage video 
for chaotic scenes is good enough.  I don't think I've ever seen any 
such chaos scene that looked good, regardless of 
source/broadcaster/decoder/display/...

The funny thing is that these scenes looked just fine in our analog 
transmission formats (which, really, just sent intra-coded 
pictures--full frames).  Digital TV broadcast uses lossy compression 
and, therefore, has issues with some types of scenes/video.  Dark scenes 
are also a huge challenge for our digital TV encoding formats--which is 
funny because dark became popular (ref Christopher Nolan/Batman Begins 
:) about the time the US (and, really, much of the world) was making the 
change to digital.  And, no, I'm not saying we should go back to analog 
transmission--just saying that we did have to make some sacrifices to 
fit 60M pixels/sec of images into our 6 (or 8) MHz channel allotments.

FWIW, the MPEG (Moving Picture Experts Group, not any format) is 
actually noticing these limitations, too, and are working to improve 
their video-coding standards to account for such weaknesses in the 
existing standard. Some of these improvements should make it to our 
broadcast standards in the next 50+ years or so.  :)

Mike