AVS Forum banner

1 - 12 of 12 Posts

·
Registered
Joined
·
2,043 Posts
Discussion Starter · #1 ·
weird question.


i watch all the Yankee's baseball games and usually here in NY they are on the YES Network. but sometimes they have the games on Fox or ESPN. tonight's game is on ESPN. i noticed when they do the slow motion replays that there is almost no blur. on Yes when they do the replays there is usually a lot of blur. is there a reason that there is less blur on slow motion replays on ESPN? are the ESPN games possibly shot at a higher frame rate?
 

·
Registered
Joined
·
2,043 Posts
Discussion Starter · #3 ·

Quote:
Originally Posted by owine /forum/post/17051311


Certain cameras shoot at a higher frame rate which leads to this.

yea, that was my guess. i guess the YES Network could use some better cameras.
 

·
Registered
Joined
·
1,090 Posts
Most regional networks employ a single high framerate camera on their broadcast. YES coins it the SuperShot. Others call it X-Mo. National broadcasts like Fox and ESPN can justify higher costs which can lead to multiple high frame rate cameras.
 

·
Registered
Joined
·
2,043 Posts
Discussion Starter · #5 ·

Quote:
Originally Posted by owine /forum/post/17051344


Most regional networks employ a single high framerate camera on their broadcast. YES coins it the SuperShot. Others call it X-Mo. National broadcasts like Fox and ESPN can justify higher costs which can lead to multiple high frame rate cameras.

nice, i'm getting schooled.
 

·
Registered
Joined
·
7,956 Posts
If you replay conventional 50/60 field/frame video as slow motion you just end up repeating frames/fields and the result appears juddery and blurry. (If you shutter the cameras you can reduce the motion blur - but this then looks unpleasant live)


There are a number of different styles of high frame-rate cameras.


Sony and Philips/Thomson/GrassValley have 3xframe/field rate SD and HD cameras that look just like conventional production cameras, but that feed 180/150 fields/frames down their camera cables rather than 50/60 fields/frames that conventional cameras use. They are recorded onto special EVS (usually) disc recorders that allow replay at up to 1/3rd speed with no repeated frames - giving very smooth slow motion.


These cameras usually have live 50/60 field/frame outputs that allow them to be used live - but this usually is just achieved by outputting every third frame, which can mean the live output looks to be very shuttered and juddery.


For this reason the super slow-mo cameras are usually deployed in addition to the live cameras, and seldom cut to live (as the judder sticks out) and instead just used for ISOed replays.


There are more specialised cameras that run at much higher frame rates. Some of these buffer the frames in camera, and "play out" from the camera head rather than a remote disc recorder, which reduces the requirement to have high bandwith full-rate camera cabling. I think these systems offer 300fps plus and are being increasingly used on sports like Golf and Tennis, including in-match stuff.


One of the early high frame rate devices was the Tornado from Arri/Nac - which was very much a replay device. It had to go through quite extensive processing in a Quantel box to get it to air - but runs at very high frame rates allowing some incredible "effects" slow motion - such as analysis of a tennis serve or a golf swing. However this isn't useful for "in match" stuff - more for post-match analysis.


Can't comment on US sport coverage - but in the UK they usually get a couple of the 150/180 fps models for football, cricket etc., with the higher frame rate ones used for higher-end stuff like Wimbledon and the Open.
 

·
Registered
Joined
·
4,893 Posts
I wonder why they don't average the three frames instead of just outputting every third. The integration time of the imagers is nearly the entire frame, so this would essentially emulate the longer shutter. I would think that sort of processing is rather inexpensive now. Artificial motion blur works by multiple stamping an image with spatial offsets in the direction of the motion and it's quite effective.
 

·
Registered
Joined
·
7,956 Posts

Quote:
Originally Posted by TVOD /forum/post/17060090


I wonder why they don't average the three frames instead of just outputting every third. The integration time of the imagers is nearly the entire frame, so this would essentially emulate the longer shutter. I would think that sort of processing is rather inexpensive now. Artificial motion blur works by multiple stamping an image with spatial offsets in the direction of the motion and it's quite effective.

I've wondered the same thing as well...


In the early days I believe they linked the CCUs to the disc recorders with 3xSDI outputs, so every 3rd frame went to a different, standard, SDI (pre-HD days) output, and the recorder coped with replaying them in the correct order (and you didn't need anything other than clever software to make it work) - and the live output was just one of these "every 3rd frame" feeds. Not sure if they still use this solution with HD-SDI (or SDI) systems - or if they now use high-speed single cable links? (It's slightly analogous to the way the BBC used to record HDTV digitally in the late 80s/early 90s before HD DVTRs were available - where they instead used 4xD1s and a converter that took in an 1152 line HD source and created 4x576 line SD D1 outputs - so effectively the image was broken down into 2x2 pixel blocks and each of the 4 pixels in this block was sent to a different SD output)


It would seem obvious that just averaging three frames to create a single "live frame" at 1/3rd frame rate would be a better solution. I guess it is a bit more of a hassle with interlaced systems (as if you run at 3xfield rate you'd have to de-interlace and re-interlace at 3xframe rate - as otherwise wouldn't you have 2xodd and 1xeven or 2xeven and 1xodd fields as sources to derive odd and even real-time fields?)
 

·
Registered
Joined
·
4,893 Posts
I also had thought about the interlace issue. With row pairing the first and third field would be correct, and the middle one would be the opposite. That field could be line interpolated to vertically center it to the first and third fields. While this would soften this field vertically a bit, it is only contributing a third of the total image and would likely not have a great impact on vertical detail. I think it would look much better than a shutter time one third the field duration.
 

·
Registered
Joined
·
10,679 Posts
One time I've seen sports productions use stuff from the high speed camera played at regular speed (with the strobing effect) is for the "bumpers" before they go to a commercial. They'll edit a string of one second clips together of the highlights from the game so far and the strobing effect which is normally irritating makes the clips look more exciting.
 
1 - 12 of 12 Posts
Top