Just curious what size I-frames everyone is getting according to StreamParser. I know that this patch is variable bit-rate so depending on the scene you will get smaller/bigger I-frame, but just curious because I noticed I am getting rather low frame sizes. Well compared to what I remembered getting with previous hacks I've used. My shots weren't anything scientific, just shooting around my friends house, tried to get the most detailed scenes I could. According to streamparser, I am getting I-frame sizes between 350,000 bytes and 400,000 on average, with sometimes going as high as 500,000. Are people getting high rates than that? FYI, I was using a Transcend class 10 at the time. Didn't have my equipment with me to use my 30mb/s card, just what was left in the camera when I left the house. Also, haven't had time to watch the footage, just ran it through streamparser on some downtime at work, my computer here can't run the files smoothly.

@driftwood, @cykla - I understand that bitrate drops with VBR and it's not a loss of picture quality I am observing here. And it's also not the effect of 24 fps vs 60 fps with smooth motion of fast moving subjects in front of the camera.

It looks like an actual recording hiccup (the drop/freeze effect that @kholi was also noticing) when the camera pans fast. It's not doing it with Quantum 100 v3, it's not doing it with FlowMotion 1.1, only with Quantum X. And that hiccup is also present when the footage is transcoded to ProRes 422. I do have a very fast and current Mac (with maxed out RAM) that doesn't choke on anything.

I'll see if I can put together a sample a little later, so that we're all on the same page.

Thanks for everything so far.

@driftwood, I've also noticed this dropframes or hickups, after conversion also, both in clipwrap and 5dtorgb, and directly in the NLE, one of the files, from QuantumXv1, doesn't finish conversion in 5dtorgb, and with clipwrap, after conversion, it repeats the first bugged frame whenever it happened till the end of the clip. In the NLE I can see it going frame by frame in the viewer. I was using a Sandisk Extreme 32Gb 30Mbs, and the shots where static as per the frames on the previous page. Thanks in advance Nick.

@eyefi and on Quantum X v2?

QuantumX v2, happened exactly the same the only difference was that 5dtorgb was able to get to the end of the process, but results were the same, various frames or dropped or out of order. I can upload the mts if you want.

@driftwood I sent you a dropbox link in a PM. It's for the QuantumX v2 25pHBR, as posted on the previous page.

Had quick play with SpanmybitchupV2 in HBR, instant card lock up at 12,800 (Sandisk 30mb/s) if that is of any use, thanks.

I had to batt drop in HBR, as well.

@proaudio4 - Thanks, dude. I shot what I could with the 15 minutes that I had!

My good experience. Using Rocket V2 I changed 24L setting to higher bitrate 66M. After some tests I filmed at very nice quality 5 memory cards Sandisk 32GB 30M yesterday. No crashes, spanning, extele and camera playback works fine. It takes about 1 hour on 32GB card. Thank you all!

@THX1965 , Completely agree about the look. It's amazing. The GH2 is changing the way I film docs and finding a stable 25p offering with spanning is almost a dream come true. Actually I can live without spanning :) but it would be nice.

@act - thanks for your feedback. When I say "fast pans", I don't mean panning across a landscape in a fast way. I mean a quick pan left or right - as a narrative tool. You start with a static shot and quickly pan (10-20 frames) to your next static shot - let's call it "adjustment of perspective". That's when Quantum X v2 chokes in my case - producing hiccups, freezframes as the camera seems to try to catch up while writing to the 30mb/s Sandisk card.

@Kount_Michula nice

Here is my first crack at Quantum v2 Rocket: