codec and bitrate setting for slomo

[Cees van Kempen]

This question is a follow up on my post about bitrate quality in Piotr's noise ratio discussion, and so well answered by Piotr and Dan. Because my next question has nothing to do with the initial noise discussion I now make this new post.

For slomo I shoot 720p/60 in my nano, which I ingest as 25p through Cinema tools in FCP. This can only be done with I-Frame Only and for capacity reasons (see the other post) I do not want to go higher than 140Mbps. I believe I read somewhere that a clip shot in LongGop can be transfered in post to an I-Frame format and than be transfered to a slower frame rate in FCP. My questions:
- is this so, and if yes, how?
- would you then recommend 720p/60 in 100Mbps LongGop, to be postprocessed to an I-frame format and slowed down, or would you recommend to shoot in 140Mbps I-Frame Only straight in the nano (I know higher bitrate is better, but I want to rul out that option).

[Dan Keaton]

Originally Posted by Cees van Kempen

For slomo I shoot 720p/60 in my nano, which I ingest as 25p through Cinema tools in FCP.

For slow-mo, 720p60 is better than 1080p30 as there are more frames of video per second.

This can only be done with I-Frame Only and for capacity reasons (see the other post) I do not want to go higher than 140Mbps.

I believe I read somewhere that a clip shot in LongGop can be transfered in post to an I-Frame format and than be transfered to a slower frame rate in FCP.

I do not understand why "I-Frame Only" enters into the situation here.

One can certainly create slow-mo from Long-GOP and one can certainly transcode to a I-Frame Only codec such as ProRes if that is necessary to use Cinema Tools.

My questions:
- is this so, and if yes, how?

- would you then recommend 720p/60 in 100Mbps LongGop,
to be postprocessed to an I-frame format and slowed down,

or would you recommend to shoot in 140Mbps I-Frame Only straight in the nano
(I know higher bitrate is better, but I want to rul out that option).

This is your choice, but 100 Mbps Long-GOP or even 50 Mbps Long-GOP should work fine for you.

I hope this helps.

[Piotr Wozniacki]

Originally Posted by Dan Keaton

This is your choice, but 100 Mbps Long-GOP or even 50 Mbps Long-GOP should work fine for you.

I'm in the 50 Hz Region too Cees, so as far as slomo goes, I usually do 720@25/50p. I'd advise 180Mbps - you would end up with 90 Mbps @25 fps, which is good enough for serious post.

I believe setting the nanoFlash to 50 Mbps will not be good enough in this case.

Piotr

[Cees van Kempen]

Piotr, your 180Mbps statement is clear, but in this specific project I want to keep bitrate lower. Dan, the I-Frame comes in because of the Cinema Tools. it does not take LongGop. Will have to find my way to transcode it to Prores (told you I am new).
So the main question is: Is it right to assume that 100Mbps LongGop transcoded to prores and the slowed down is better than 140Mbps I-Frame Only in nano and then slowed down?

I do not use the overcrank option in the nano (which I believe is always I-Frame Only anyway?) and prefer 720p/60, because that gives a slomotion of 25/60 instead of 25/50. Being in a pal country does not have to withhold you from shooting 60p for slomotion, for as long as you bring it down to 25p and not 30p.

[Dan Keaton]

[QUOTE=Cees van Kempen;1572877]
So the main question is: Is it right to assume that 100Mbps LongGop transcoded to prores and the slowed down is better than 140Mbps I-Frame Only in nano and then slowed down?
[QUOTE]

No, the 140 Mbps will be better than 100 Mbps.

Based on objective, but limited tests, 140 Mbps I-Frame Only is slightly better than 140 Mbps Long-GOP, but this is not the important part.

The important part is that 140 is better than 100.

But one should not think that 140 is 1.4 times as good as 100.

[Piotr Wozniacki]

Originally Posted by Dan Keaton

No, the 140 Mbps will be better than 100 Mbps.

Based on objective, but limited tests, 140 Mbps I-Frame Only is slightly better than 140 Mbps Long-GOP, but this is not the important part.

I beg to disagree - 140 I-Frame Only is not even comparable with 100 L-GoP.
The slowest acceptable I-Fo is 180; 220 being OK and 280 - Master Quality.

[Dan Keaton]

Dear Friends,

I will carefully parse my statement:

"Based on objective, but limited tests, 140 Mbps I-Frame Only is slightly better than 140 Mbps Long-GOP, but this is not the important part."


"Based on objective": Tests based on the Video Clarity System
"but limited tests": Our tests, using the Video Clarity System was with their "Football Game" video, and not all of the other reference videos that they provide with their excellent system.

"140 Mbps I-Frame Only is slightly better than 140 Mbps Long-GOP": The objective results of this test were very close.

These tests are created by Video Clarity using professional cameras, creating low-noise videoes, using enough light to obtain a great, clean image from the camera.

Your mileage will vary, especially if one records images that are not low in noise, such as those that can occur with natural light or artifical light that is at a level that is not optimum for the camera in use.

In these cases, lower bit-rates can mask the noise in the image.

Personally, I prefer 50 or 100 Mbps Long-GOP, but this is a personal preference, one not based on the objective, but limiited tests that we performed with the nanoFlash using the objective Video Clarity system.

[Cees van Kempen]

Hope you do not mind that this is difficult stuff for the less experienced. I 'technically' understand all the statements you both make, but it seems there is no final conclusion. Which make me realize that both options might be more or less okay, cause the quality out of the nano will meet my requirements anyhow. The one thing I can not substract from your answers: if choosing LongGop 100 and transcoding to Prores, would the transcoding lead to significant loss of quality, so the LongGop looses it's advantage above acquiring in I-Fo in the first place?

[Dan Keaton]

Dear Cees,

Of course, when one renders a video file to another format, there is some loss of quality.

But, it does not matter if the source is Long-GOP or I-Frame Only, everything else being equal.

When one renders to another format,

first the existing file is decoded using the original codec,

which produces a uncompressed frame, or group frames,

which is then rendered into the destination codec.

If you have the nanoFlash, a camera, Final Cut Pro, and Cinema Tools, you could run your own tests and see if you can detect a difference.

Please note the "Everything else being equal", is tough to achieve.

Long-GOP has its inherent advantages,

and I-Frame Only has its inherent advantages.

[Piotr Wozniacki]

Cees,

If I opted for as high bitrate for recording slomo as possible, is because the effective data rate "per frame" of the resultant material will only be half of the bitrate set on the nanoFlash.

When I do 25/50 slomo on my EX1, it increases the 50 fps recording bitrate above the HQ's 35 Mbps so that the above mentioned effective data rate per frame recorded is still the HQ's 35 Mbps.

Unfortunately, it seems the nanoFlash doesn't do that; when I do a 25/50 overcranking at 180 Mbps, the nanoFlash says "Recording 90 Mbps".

If I'm wrong in this, Dan, please correct me :)

But if I'm right, then setting the nano for 100 Mbps, your overcranked stuff will only be effectively recorded with datarate of 50 Mbps (assuming a 25/50 fps slomo). Definitely too low for I-Fo!

Piotr

[Daniel Doherty]

How can 140 Mb/s I-frame be better than 140 Mb/s Long Gop when the Long Gop is 2 to 3 times more efficient?

[Dan Keaton]

Dear Daniel,

The efficiency of a codec varies with the bit-rate.

At lower bit-rates, a Long-GOP codec can be 2 to 2.5 times as efficent as an Intra-Frame codec.

At higher bit-rates they can become closer.

But, there is no one answer.

Also, things change depending on the type of video.

No one test can provide all of the answers for all types of shooting.

[Daniel Doherty]

Dan, thanks for helping to clarify.

[Steve Phillipps]

Cees, reading your post it sounds as if you are recording 60P onto the Nano rather than using the Nano's Overcranking function - is that right? Otherwise there'd be no need to go to Cinema Tools, the material would come out as slomo natively from the Nano. That way you could record Long GOP as you'd not need to conform in CT.
Steve

[Piotr Wozniacki]

Originally Posted by Steve Phillipps

Cees, reading your post it sounds as if you are recording 60P onto the Nano rather than using the Nano's Overcranking function - is that right? Otherwise there'd be no need to go to Cinema Tools, the material would come out as slomo natively from the Nano. That way you could record Long GOP as you'd not need to conform in CT.
Steve

AFAIK, overcranking is always recorded as I-Frame Only on the nanoFlash - regardless of its original setting.

Dan?