6 gop vs 15 gop????

[PappasArts]

I was having a discussion and the subject came up about 6 gop vs 15 gop.

Which is better for image quality. Just curious to know the benefits of either one!

Thanks!

 

[Barry_Green]

As with all questions, "it depends."

The longer the GOP, the more efficient the compression can be.

But the more susceptible to bit-starved ugliness. If the compression runs out of bits for its group, the whole group gets downgraded and blocky.

JVC's implementation appears to deliver more consistent performance with more resilience to macroblocking than Sony's does. But JVC also allocates a lot more bits per pixel, depending on what frame rate you're shooting at: Sony's trying to encode about 60% more pixels than JVC is, and in 24P mode JVC's putting even fewer pixels to the tape, so the proportionate bits-per-pixel are even higher. In 24P mode JVC's encoding about 22 million pixels per second, whereas Sony is trying to encode 46.6 million pixels per second. So the Sony is doing more than twice the pixels, but has only 30% more bandwidth to do that job with. The only reason it can look competitive is because of the greater efficiency of the longer GOP.

But when it goes to crap, it goes to crap for a longer period of time and is impacted harder. I'll have to take a JVC down to the Sirens of T.I. show and put it through the same torture test I put the Sony through, that should really give us an example of seeing how they perform relative to each other in the worst possible case.

But in general, I find the JVC to be more robust and more artifact-resilient than the Sony.

Haven't had an opportunity to put Canon's through its paces yet. Canon talked a lot about the efficiency of their MPEG encoder so they seem pretty proud of it.