PTool v3.62d topic

cbrandin

@Zorba

My settings file only affects 1080p, you'll need to set "Video bitrate for FSH/SH" to 44M as well. I haven't tried raising the buffers for 1080i, so I'm not sure how you would do that. Actually, they probably don't matter too much, so I wouldn't mess with those.

Chris

vladnik

@cbrandin
what is minimum quantizer value when we use auto qauantizer 4 ?
how much low it can go?

towi

for 50i I wouldn't raise the bitrate over 32Mbits (default doubled)... lots of artefacts can happen in motion otherwise. I never understood the values suggested in PTool.
The new AQ patches however DO improve IQ... 50i looks remarkably better than the stock GH2.

Zorba

@chris, thank you!

I will also try towi's tip: 32mbs + AQ 3 or 4.
Some question for AQ: must i have IQ enabled for AQ?

ignatius

Please delete.

cbrandin

@vladnik

It depends on how much you have raised the bitrate. If you have doubled the bitrate New QP = Old QP-6. If you are using 88M, for example, New QP becomes Old QP-12. If you set AQ lower than 4 the decrease in QP is proportionally less. If you set AQ to 0 it doesn't change it at all. The factory default QP is 20. The actual QP can change from frame to frame depending on the nature of detail and type of frame. Typically I frames use a QP range of QP-2 to QP+2, P frames use QP, and B frames use QP+2.

Chris

vladnik

@cbrandin
so if i use 88m AQ4 minimal QP can go to 8 , factory QP 20-12.
what will hapen if i add minimal value 1 in AVCHD ST.
can qp go that low?

cbrandin

If you go too low with QP the codec runs out of bandwidth before finishing a frame and the lower part looks like crap. Try it, it may be possible to go a little lower. You have to use the manual QP settings in the tester section, though.

Chris

vladnik

that means we can use auto qauantizer and manual QP settings together?

vladimir

chrimsbroome

Guys, lads, girls and men..............dont you think we should have a post that is about what settings are working and what are not. Like a checklist because is it just me or is there lots going on.

derek

Спасибо за ваш хак Виталий

cbrandin

I think the manual settings override AQ.

cbrandin

I think Panasonic gets around some of the filtering CPU overhead by using pre-calculated tables. Those tables don't change below a QP of about 24. My guess is that they didn't bother with lower settings because they assume the QP will never go below 20 (the factory low limit). I think I know where they are, but the parameter tables have other stuff mixed in. I'll have to decipher what they are doing.

Chris

Stray

Wouldn't changing your FB sizes based on your 24H setting really knock the ratio between I frames to B and P frames badly out of whack for the 24L setting ? In that it would be a good idea to set your 24L bitrate not that far out from your 24H bitrate ?

kron

PTool v3.62d cannot produce stable settings for me, the uneven cadence always appear at some point in the clip. I never went over 44Mb, played with buffers, AQ, GOP, or used defaults, or variations of such, but always the cadence issue. I reverted to stock firmware for now.

p.s.: I did all my test with the voigtlander 25mm and on manual mode only.

Stray

I'm getting the same Kron, I've tried settings at both 44 and 66MB with double and triple buffer sizes respectively. There seems to be no predictability to it either except that it will happen sometime when you've shot over a minute. I've mostly been shooting long static shots. Oddly shooting outdoors with an ND filter seems to make it less likely to happen, I'm going to just try and turn the AQ off before going back to the stock firmware.

Stray

Sorry, forgot to say, I've mostly been shooting with a GOP of 12. Also, there actually doesn't seem to be much wrong with the footage itself when I play it back, but I guess I'd notice it in post if I had time to play with it.

Vitaliy_Kiselev

If you have problems with cadence you can try 1080p24 T4 patch (in testers section) and select other table as replacement (1080p24 T1, T2 or T3).

mistan

Vitaliy,
Are there any drawbacks to doing that? Is there a compromise of details or motion? Will it lower any kind of quality?

Vitaliy_Kiselev

@mistan
Just try and see if it helps.

mistan

Ok will do, thanks.

Stray

Yes it definitely helps. I've not done proper rigourous tests, I've just shot the same plant outdoors, so there is slight motion. As the cadence problem is so dramatically visible in streamparser no matter what you shoot, I figure its okay, though I will test it some more. It does seem that I frame size does drop off as it records, but the cadence itself is a lot smoother. I can't say for sure, but movement does also cause a slight drop in I frame size, but it does recover.

Replacing T4 2with T3 doesn't improve things much, T2 and T1 however does seem to remove the cadence problem, although I and P frame sizes do drop after about 20-30 seconds of recording. B Frame sizes increase however, and it remains a solid consistent pattern from 20 seconds onwards on a completely static shot. I took another indoor shot around 2 minutes long to confirm this suspicion which are the last two files attached here.

Edit : GOP 12, Frame buffers x2, 44MB on all tests.

vladnik

@cbrandin
GH2 filtering image at 24 qp min before being decoded.
it looks like a significant loss of quality.
Perhaps this is a reason we do not see a drastic difference between 44 and 88 mbit .
if we could reduce the value or to exclude filtering it would allow the codec to work with the original image values.

towi

@Stray
"Replacing T4 2with T3 doesn't improve things much, T2 and T1 however does seem to remove the cadence problem"

??? ...

so you've altered this setting -> Patches for testers - Quantizer tables - 1080p24 T4 ??

And you've altered 1080p24 T4 to T2 (T1 respectively)... is this correct?

Any difference from T1 to T2 ... what do these settings do at all?

Thanks in advance.

Vitaliy_Kiselev

>what do these settings do at all?

They alter quantizer tables used.
T4 is fallback tables used if encoder thinks that remaining bitrate is very low.

Howdy, Stranger!

Categories

Tags in Topic

Top Posters