@zzap64 yes with those settings. Its all because of the limits. Any higher and youre talking buffer underruns etc... You have to do extrenuous testing when you start changing bitrates- use streamparser or try d/l trial versions of elecard software - because what might seem stable and all nice recordings to you, on the charts could be huge dropouts, sudden drops in bitrate and oif course buffer underruns/overruns plus cadence issues. Test using small increments/decrements and one setting at a time. 1080i60 and 720p60 are the prblem zones as far as the hack goes. A long gop is probably better suited to these two. The way the GH2 handles 720/60 and 1080i60 is hard to nail down as you'llk find if you read past posts @Ralph_B discovered smearing at odd frames too. Be wary in all your testing. I think 720p60 is the last bastion for the hack - noone has found satisfactory improvements to date including my own. Anyone who finds an incrdible perfect setting will certainly get my salutations.
@Huckster Either. AQuarius just brings in 1080i60 on INTRA again and it has a slight drop to 170M in bitrate to try & improve card writes. Not really much between them as they are both on AQ3 at 1080p24
That's what I was hoping to hear... that the drop from 176M to 170M is negligible and doesn't really make a difference in picture quality.
Any thoughts on what's coming next? Do you think we've maxxed out the picture quality on the GH2 for 1080p24 INTRA or do you think there still may be a bit of extra quality to eke out somehow?
@driftwood I really liked smoothasfuck because I would get higher bitrates around 120 on detailed scenes and as low as 50 on more average ones. They would even span so this is close to an all purpose setting. What should I modify to keep the same general dynamic but somewhat larger maximum frames? Btw the quality on the latest Aquarius is just astounding. Thanks.
I have been doing some experimenting and came up with something I think will work quite well for me. I took Driftwood's Aquarius V2 and, after much experimenting, settle on changing 24l to 66mbits. The images look very close to the 170 of 24h and the grain still has that beautiful quality. It spans no problem (as my 88mbit test did) on 45mbit UHS cards. The 88 spanned also but had a strange spiking behavior in streamparser. 66 has a perfectly even stream except for the first few frames. It has played well and edited in everything I have tested.
Question, could there be a version of ptool where 24h would have a different gop than 24l or does the architecture fundamentally make that impossible?
Hey all -- is the 44M patch the highest one that will span? Sorry if this should be in another forum, but wondering if it been decided whether the 45MB/s SD cards work as well as the 30mb/s? I need to get a 32GB card for the higher bit rate, low gop patches. Thanks!
*** NEW *** Driftwood 'SeAQuake' - INTRA & Progressive Seg Frame 1080p24 GOP1 INTRA & 1080i PsF both at AQ3 & 720p at AQ2. 1080p24 and 720p50 all INTRA; 1080i50/60 INTRA PsF, 720p60 (GOP3). Includes all new Encoder settings to improve macroblocking and QP. Bitrate is also primed to the minimum to help work on decent cards and above.
1080i60 converts well to 29.97 as far as I can see - tested using P2 mxf ;-)
If you have a poor card, don't come crying.
Otherwise, enjoy, it's tasty!
Driftwood 'SeAQuake' - its INTRA all round apart from 720p60 (GOP3). All settings now with enchanced Encoder QP - setc.zip
@driftwood filmed a video installation with the last intra, super image! Even at 720. Would have loved to send it into the collaborative project.... Will post a few scenes...
PS: Lens 35mm Samyang 1.4 PSS: Image not original 720p - sorry- will repost when I have file at hand :-(
@Driftwood Thank you very much for your new patch release. For me, the great news is the ability to get 25 P via PsF. My doubt is about the workflow to get this. Any idea?
@driftwood Every test has been rock solid. How should I expect the buffer under run to manifest. VK mentioned that 3.63 allows very low 24l bitrates. Could that be why this is working for me?
@Driftwood Looking closer at your chart, I see it quickly drop from the h rate to the l rate, which is where I want it to be. Why should I care about those early frames when they do not appear to cause any practical interference? What should I be worried about? Thanks