Is there any Difference between 720p60 or 720p30 data wise?

[Barry_Green]

There is a flag in the timecode, I think it's in the user bits, that flags active frames versus duplicate frames. OnLocation is smart enough to throw away duplicate frames and only store the active frames, but only in one of its recording modes (Type II avi instead of Type 1, or something... don't remember exactly which one).

[DJDecay]

Originally Posted by Barry_Green

There is a flag in the timecode, I think it's in the user bits, that flags active frames versus duplicate frames. OnLocation is smart enough to throw away duplicate frames and only store the active frames, but only in one of its recording modes (Type II avi instead of Type 1, or something... don't remember exactly which one).

Ooh, that's very cool, the TC track has an active/dup frame flag. Awsome.
Currently I've been matching frames at 1F byte with 07 and 03 values - Which is if I reparse it - and set Quicktime durating to say 1020 frame blocks at 59.94 its equiv. to 510 frame blocks at 29.97 in seconds of playback and playspeed (realtime or 1x) - but the active frame thing is a way better mechanism than finding pairs so I can tell if the stream is indeed 60P (which needs a larger index table).

Nice side-effect though is instant slow-motion if the parser makes mistakes and wrong fps or duration is assigned in frames.:-). I'm sure FCP would puke on its import, but its possible to just assign a 33fps or 10fps to Quicktime file.

Thanks alot for the tip I really appreciate the help.

[Cranky]

I am sorry for "duh".

Is 720p30 recorded within 60p with actual duplication of frames or with "repeat frame" flags?

DJDecay, why you are excluding "special" frame rates like 720/30PN ? Does OnLocation NOT support this format?

[Barry_Green]

The frames are actually duplicated. An intraframe codec means that every frame exists, in its entirety, individually, hence no frame duplication flags.

30pN doesn't travel down the firewire, so OnLocation won't see an HVX/HPX in 30pN mode.

OnLocation has the ability to simulate 30pN mode though, either it's Type 1 or Type 2 AVI that lets OnLocation discard duplicate frames and only store the active frames.

[Cranky]

Originally Posted by Barry_Green

The frames are actually duplicated. An intraframe codec means that every frame exists, in its entirety, individually, hence no frame duplication flags. 30pN doesn't travel down the firewire, so OnLocation won't see an HVX/HPX in 30pN mode.

So, recording 720/30pN onto a card and then dumping it on a computer will yield a higher-quality video than streaming 720p30 via FireWire?

I am not sure that "an intraframe codec means ... no frame duplication flags." It means no "delta" frames, but why not repeating a whole frame? I googled a bit, did not find anything definitive.

[Barry_Green]

Originally Posted by Cranky

So, recording 720/30pN onto a card and then dumping it on a computer will yield a higher-quality video than streaming 720p30 via FireWire?

It will result in an identical-quality video stream, but in half the space.

I am not sure that "an intraframe codec means ... no frame duplication flags." It means no "delta" frames, but why not repeating a whole frame? I googled a bit, did not find anything definitive.

Intraframe means that every frame is compressed individually and stands alone. By using frame-repeat flags, it's no longer intraframe, because inherently the duplicate frame is dependent on the prior frame.

[Cranky]

Now I look like an idiot, this always happens when I say "duh". I thought that one could explicitly choose bitrate AND frame rate for DVCPROHD, but instead one chooses frame rate, and bitrate is chosen automatically: 100 Mbit/s for 60fps, 50 Mbit/s for 30 fps native, 40 Mbit/s for 24 fps native. What about 25/50fps? Are these 50/100 Mbit/s or 40/80 Mbit/s? What about overcranking? Is overcranked video recorded with higher bitrate so that when played back with nominal speed (say, 24fps) it had nominal bitrate (say, 40 Mbit/s)?

Also, wikipedia says that DVCPROHD uses 4 parallel codecs, is this true? I mean, is the whole frame divided into four super-blocks and each block is encoded separately?

[Barry_Green]

The actual bitrate used is dependent on the frame rate, because each and every frame is allocated the exact same amount of bandwidth.

DVCPRO50 uses two codecs ganged together, DVCPRO-HD uses 4. However, DVCPRO50 and DVCPRO-HD also get 4:2:2 color, so I'm not 100% sure how they use those 4 DV codecs in parallel, because DV in and of itself is 4:1:1.

[Cranky]

Originally Posted by Barry_Green

The actual bitrate used is dependent on the frame rate, because each and every frame is allocated the exact same amount of bandwidth.

I got this, but is it 100/50 Mbit/s or 80/40 Mbit/s for 50/25 fps, considering that 60 fps is 100 Mbit/s. Ok, I just took a look at the HPX500 info, and it seems that 50/60fps and 25/30fps use the same bitrate, 100 and 50 Mbit/s respectively. Well, duh (may I "duh" myself?), considering that DVCPROHD is a variant of DV, which uses the same bitrate for 50/60 Hz versions.

The last digression in this topic: is it possible to record 24/25/30pN onto tape? I don't mean miniDV tape but large tape.

[Barry_Green]

Tape is always 100mbs. The whole reason the HVX/HPX line have a 720p mode in addition to a 720pn mode, and 1080/24p with 2:3 or 2:3:3:2 pulldown, is because of the way tape (and streaming) work. They have to have that fixed data rate.