IPB vs ALL-I settings

[Wilmark]

Has anyone done any extensive comparison about choosing between these two settings other than the fact that one produces much larger files. The manual says it is easier to edit in post.

[Axilrod]

Has anyone done any extensive comparison about choosing between these two settings other than the fact that one produces much larger files. The manual says it is easier to edit in post.

All-I may have a bit more latitude in terms of post, but I thought the IPB responded to changes just fine and I honestly couldn't tell the difference between the two.  I've heard if there is going to be a lot of motion involved that ALL-I is better, but for relatively static subjects I think IPB is just fine (and much smaller!)

[psolberg]

IMO all-i looks worse often and not better often enough so I wouldn't use it for anything as the file size doesn't really yield better quality.

[HurtinMinorKey]

I think the basic deal is that you get more latitude in post, but only if you are editing frame by frame. If you are applying edits to the whole take, then IPB is just as good for editing. Furthermore, IPB has more detail in each frame because it isn't reliant on each frame providing all the information for the entire frame.  Many people have reported block artifacts with ALL-I, because it requires significant compression in each frame.

Think about it, truly uncompressed 1080, ALL-I should be (2M(resolution)*3bytes(24bit color)*(24fps)=144 MB/s=1.1Gb/s

The 5dmkiii max data rate is 1/10th of that, so it requires massive compression in each frame. IPB circumvents this issue somewhat by putting more detail in a key frame.

[Cptn Rigo]

I think the basic deal is that you get more latitude in post, but only if you are editing frame by frame. If you are applying edits to the whole take, then IPB is just as good for editing. Furthermore, IPB has more detail in each frame because it isn't reliant on each frame providing all the information for the entire frame.  Many people have reported block artifacts with ALL-I, because it requires significant compression in each frame.

Think about it, truly uncompressed 1080, ALL-I should be (2M(resolution)*3bytes(24bit color)*(24fps)=144 MB/s=1.1Gb/s

The 5dmkiii max data rate is 1/10th of that, so it requires massive compression in each frame. IPB circumvents this issue somewhat by putting more detail in a key frame.

144 MB/s=1.1Gb/s mhhh.... nope 

[Drizzt321]

I think the basic deal is that you get more latitude in post, but only if you are editing frame by frame. If you are applying edits to the whole take, then IPB is just as good for editing. Furthermore, IPB has more detail in each frame because it isn't reliant on each frame providing all the information for the entire frame.  Many people have reported block artifacts with ALL-I, because it requires significant compression in each frame.

Think about it, truly uncompressed 1080, ALL-I should be (2M(resolution)*3bytes(24bit color)*(24fps)=144 MB/s=1.1Gb/s

The 5dmkiii max data rate is 1/10th of that, so it requires massive compression in each frame. IPB circumvents this issue somewhat by putting more detail in a key frame.

144 MB/s=1.1Gb/s mhhh.... nope 

Actually, I believe that is correct, provided you only go to 1 decimal places. Google Calculator results.

[Graham King]

144MB x 8bit/byte = 1.152Gb

[HurtinMinorKey]

144 MB/s=1.1Gb/s mhhh.... nope 

144 MB/s=1.15Gb/s=.144GB/s

Everybody happy now?

If you think about it, the 5d mkiii can handle 6 fps @ 25MB per shot, so its capable of  doing over short bursts(at least in terms of processing).  But they can definitely do better than the bit-rate they currently offer.

[psolberg]

144 MB/s=1.1Gb/s mhhh.... nope 

144 MB/s=1.15Gb/s=.144GB/s

Everybody happy now?

If you think about it, the 5d mkiii can handle 6 fps @ 25MB per shot, so its capable of  doing over short bursts(at least in terms of processing).  But they can definitely do better than the bit-rate they currently offer.

the pathways for stills record raw data. video needs to be processed and tons of data discarted. this bitrate is likely baked in the circuits and can't be really upped or they would have started higher. if you want higher rates you need a clean uncompressed HDMI out. hopefly magic lantern will find a way since that is the only real solution to bitrate issues.

[HurtinMinorKey]

the pathways for stills record raw data. video needs to be processed and tons of data discarded. this bit rate is likely baked in the circuits and can't be really upped or they would have started higher. if you want higher rates you need a clean uncompressed HDMI out. hopefully magic lantern will find a way since that is the only real solution to bit rate issues.

I don't know. Just think about what the 1D-C can pump out, with Raw 4k.  the DIGI 5 processor is a big leap over 4.

[marvinhello]

the pathways for stills record raw data. video needs to be processed and tons of data discarded. this bit rate is likely baked in the circuits and can't be really upped or they would have started higher. if you want higher rates you need a clean uncompressed HDMI out. hopefully magic lantern will find a way since that is the only real solution to bit rate issues.

I don't know. Just think about what the 1D-C can pump out, with Raw 4k.  the DIGI 5 processor is a big leap over 4.

1D C doesn't have Raw 4K, it uses Motion JPEG compression, at 8bit 4:2:2, bitrate 500Mbps. Still highly compressed.

[Axilrod]

Rega
1D C doesn't have Raw 4K, it uses Motion JPEG compression, at 8bit 4:2:2, bitrate 500Mbps. Still highly compressed.

This is very true, but it still looks damn good.

[HurtinMinorKey]

So the 5D3 should be able to rock it at 250Mbs. Thanks for the correction on the lack of Raw on 1DC.

[marvinhello]

So the 5D3 should be able to rock it at 250Mbs. Thanks for the correction on the lack of Raw on 1DC.

5D3 uses H264 compression, it's a bit different from Motion JPEG, I prefer Motion JPEG