Native 24P/25p settings/problems discussion

Just go out and point it at a large patch of grass and record for a while.

I don't have the camera with me or I would, but it should be failry easy to test.

FOr the record, my settings are D settings. I thought they were C, but they are D:

jobless said:
New tests:
Video buffer 40.000.000
Video bitrate adjustment - c)40000000 d)50000000 e)60000000
Overall bitrate adjustment - c)46000000 d)56000000 e)66000000
Limiting bitrate adjustment - c)50000000 d)60000000 e)70000000
c) Gives you Overall bit rate : 34.2 Mbps
Maximum Overall bit rate : 46.0 Mbps
http://www.nikolicnemanja.com/Csettings.MTS
d) Gives you Overall bit rate : 37.4 Mbps
Maximum Overall bit rate : 56.0 Mbps
http://www.nikolicnemanja.com/Dsettings.MTS
e) Gives you Overall bit rate : 4 755 Kbps
Maximum Overall bit rate : 66.0 Mbps
http://www.nikolicnemanja.com/Esettings.MTS

Unpatched Version:
http://www.nikolicnemanja.com/Unpatched.MTS

With E setting bitrate drops! I double checked all parameters and everything was ok...
Also I noted that on C and D first frame is with MUD, but others are not.. And, there is difference in sharpens between C,D and unpatched .. :)
Click to expand...

I am using Ptool3 still, and YES I am using VIDEO BUFFER. I don't care what anyone says, when I moved video buffer beyond 40 it gave me weird images, and if I turned it off the recording errors started.

I left it on, I have no issues. I'm not changing it. lol
 
Svart said:
...having issues when they least expect it, like during a... childbirth or something... You know, something you can't simply stop everything to pop the battery out and put it back in!
Click to expand...

"Hey sorry guys... my cam froze up... honey we're going to need another take... Nurse! Shove that baby back in!"
 
I can get my GH1 to seize-up about 100% of the time if I do this:
http://www.dvxuser.com/V6/showpost.php?p=2033148&postcount=245

All I have to do is use my Panasonic kit 14-140mm zoom lens or Lumix 20mm pancake lens _and_ use PTool v3.36 AVCHD settings 24pN (no pulldown) _and_ either PTool's A, B or C bit-rates _and_ shoot a locked-down 1080p shot of a "static scene of high detail death" (in which mostly nothing moves) which contains a _lot_ of fine detail such as concrete, finely textured carpet, etc. Bingo; happens every time I try it. This is with a Panasonic Class 10 card.

However, just as Vitaliy Kiselev recommended, if I instead use my old manual Nikon 55mm SLR lens I can't freeze-up my GH1 shooting 1080p AVCHD with these same settings & scenes. Or, if I turn 24pN off and record 1080p24 with pulldown, and at high bit-rates, I can use all my lenses, even for static scenes of high detail death.

Meanwhile, I was able to shoot 1080p for several hours with no freezes using PTool's 24pN & C settings & all my lenses on this 1972 'Cuda shoot in which every take is either a moving dolly shot and/or there's no static scenes with fine textured detail:
http://www.dvxuser.com/V6/showthread.php?t=215424

I'm not complaining. In fact, I sent Vitaliy my second $25 donation today. :-)

YMMV.
 
Last edited:
Peter J. DeCrescenzo said:
I can get my GH1 to seize-up about 100% of the time if I do this:
http://www.dvxuser.com/V6/showpost.php?p=2033148&postcount=245

All I have to do is use my Panasonic kit 14-140mm zoom lens or Lumix 20mm pancake lens _and_ use PTool AVCHD settings 24pN _and_ either A, B or C bit-rates _and_ shoot a locked-down shot of a "static scene of death" (in which mostly nothing moves) which contains a _lot_ of fine detail such as concrete, finely textured carpet, etc. Bingo; happens every time I try it. This is with a Panasonic Class 10 card.

However, just as Vitaliy Kiselev recommended, if I instead use my old manual Nikon 55mm SLR lens I can't freeze-up my GH1 shooting AVCHD with these same settings & scenes. Or, if I turn 24pN off and record 1080p24 with pulldown, I can use all my lenses, even for static scenes of death.

Meanwhile, I was able to shoot for several hours with no freezes using PTool's 24pN & C settings & all my lenses on this 1972 'Cuda shoot in which every take is either a moving dolly shot and/or there's no fine textured detail:
http://www.dvxuser.com/V6/showthread.php?t=215424

YMMV.
Click to expand...

I don't have a kit lens. All my lenses are Nikon primes.
 
So it sounds like the lens communicating with camera can be a problem. I don't have the kit lens or the pancake, but are people getting same errors with old manual lenses?

Also, Svart, why were you testing at such high bitrates? Did you already test with the lower C settings of 50,52,60 with errors? I just think the higher you push then of course you're going to get errors and camera freezes.
 
Some found that the lens "pulsing" slightly while maintaining focus actually helped this issue. This issue is strange because it's not bitrate issues in the normal sense because when you are panning the camera, the bitrates look *normal*. When the error happens during static shooting an I frame will suddenly become 10x bigger than it should be and immediately error the camera while other I frames will be OK. Some folks have captured P frames doing this too. Some have found double I frames where P frames should have been, etc.

Some folks with the Sandisk cards have caught these I frames, in the pictures I posted in an earlier post, simply because the Sandisk card can take a much higher bitrate before pooping out.

I generally use the "C" settings. I tested high bitrates without the Native 24p to see how much the cards could sustain with a normal bitstream.

With the Native 24p I could get the stock 16-18-20 to fail with the right scene.
 
Last edited:
Svart said:
I generally use the "C" settings. I tested high bitrates without the Native 24p to see how much the cards could sustain with a normal bitstream.

With the Native 24p I could get the stock 16-18-20 to fail with the right scene.
Click to expand...


Did you just enable 24P patch? Or were other hacks enabled as well?
 
Kholi said:
... I am using Ptool3 still, and YES I am using VIDEO BUFFER. ...
Click to expand...

Hi Kholi: Does PTool 3.36 (or newer) have a "video buffer" setting for AVCHD? I don't see this option in my copy of 3.36. Are you using an older version of PTool? Just wondering. Cheers.

EDIT: Nevermind, I see what you meant now. You _are_ using an older version. Cool.
 
Peter J. DeCrescenzo said:
Hi Kholi: Does PTool 3.36 (or newer) have a "video buffer" setting for AVCHD? I don't see this option in my copy of 3.36. Are you using an older version of PTool? Just wondering. Cheers.

EDIT: Nevermind, I see what you meant now. You _are_ using an older version. Cool.
Click to expand...

Yeah I am using Ptool 3.0.0.0 and I've got the Video Buffer enabled.

Honestly, I read a LOT about people saying Vid Buffer doesn't do anything. You'd be hard pressed to tell me otherwise because when I had the other Transcend media card and VId buffer was at 50 or 60 it destroyed my image: it looked like water colors smeared all over water color paper.

I backed it off to 40, reset camera, took out the battery, loaded everything back up (a very hard reset for any electronic device after firmware update tends to be good, as they all operate like computers) and since then I haven't had a single issue.

I could definitely try to record some super high detail static stuff when I get a chance to get my hands on another camera, and I'll even do it with the widest lens I can find.

"D" settings, Ptool3, Native 24p is working for me with no lockups.
 
I have just done the "Patch of Grass" test.

Gh1 Pal
Results:

Panasonic Gold Class 10 4GB (19.80 Read Speed and 10.88 Write Speed)

40 | 46 | 50 1080p Froze after 18 seconds 1080i NO FREEZING after 2 minutes

38 | 40 | 42 1080p Froze after 30 seconds 1080i NO FREEZING after 2 minutes

I switched the camera off after 2 minutes, after this test I have decided to leave mine at 40 | 46 | 50 1080i

Cheers

Paul
 
I have also been using 38/40/42 1080i and 720p on a PAL camera (no mpeg hack and only FD lenses)

Have shot over 200 clips of various scenes, static and motion, on 3 different cards ...

Transend 16gb class 6
ADATA 16gb class 6
Eagletec 16gb class 6

No errors, bitrates vary from 13mbps to 34mbps, average seems to be 27mbps

No mud, more color, images look fantastic compared to original which always had pulsing in high detail scenes when viewed on large LCD TV, now gone.

I'm sticking with these as i don't see any reason to go higher and i convert everything to cineform anyway, so removing interlacing is just a tick in a box.

Not having IS the only difficulty i can see.

Wow the EVF is huge compared to the tiny one in the HMC 152.

720p is almost jelloless

Internal meter and histogram is helpful, zebras would be nice though.

GH1 body $899 (retail)
FD 50mm f1.4 $45 (ebay)
FD 20 - 35mm F1:3.5 L $65 (ebay)
FD 100mm F1:2.8 $35 (ebay)

Total around $1000 for this quality image..... awesome. Just need FaderND.

Thank you Tester and Co.
 
Last edited:
While I agree with Svart that there’s a bug in the firmware (or hack) that has issues with 24pN footage of high-detail static shots, some people ARE successfully using the firmware hack even with the bug. Issac reported that he could shoot the star chart with his camera/settings. So it can be done.

Until this is fixed at a firmware level, it seems pragmatic to sort out who can currently use it and why. Doing so will help more people use it as-is, which -- if it works -- is a good thing. There seems to be evidence that, stability aside, the 24pN can look better if all other variables are equal.

If there was absolutely no way of getting the 24pN patch to work, Isaac wouldn't have gotten it to work on a static star chart. So there has to be some variable (or set of variables) that are creating problems/success.

So let’s look at variables:

SDHC CARD: Isaac is using a card that's acknowledged as one of the fastest around; it works for him. Others using the Sandisk Extreme cards (also super fast) have reported success with 24pN. Svart is using patriot class 10 and Adata class 6 cards; 24pN doesn’t work for him. While this could be a misleading variable, evidence suggests the card you use matters. And there could, of course, be other variables that interact with this one. (Look at card speeds here.)

MANUAL LENSES: Isaac suggested that the culprit could be using lenses that send metadata. That, too, was what I was thinking, and Peter's post about success/failure with manual lenses seemed to also support that theory. But then Svart said he was only using Nikon primes. So that suggests that the fact of using manual lenses alone is not the issue.

LENS SHARPNESS: Peter joked to Svart that his Nikon lenses could be too sharp. Lens sharpness could be an issue -- Nikons are mighty sharp. But, then again, the Contax Zeiss lenses that Isaac is using successfully are pretty friggin' sharp too. So I’m skeptical that having "too sharp lenses", at least for now, is the issue.

OTHER PTOOL SETTINGS: Isaac has reported that he's not checking anything except the 24pN and "C" settings. No "3rd party battery", no MJPEG, etc. Svart, have you tested while unchecking settings like battery, MJPEG, etc.? Isaac asked you this and I’ve not seen an answer. This could be an issue.

OTHER SETTINGS:
I'm just brainstorming here… not subscribing to this theory -- but could it be something totally unsuspected, like…

- Film Mode settings -- sharpness, noise reduction, etc.
- other camera menu settings?
- even something else totally off the wall like what you have your camera's AF-S / AF-C / MF dial set to?
- whether or not you're using the neckstrap. Just kidding. :grin:

Unlikely that some of these things would interfere or be involved, but since we don't fully understand everything about the way the camera is programmed and built anything is still on the table as far as I'm concerned. Until people report these factors, we have no way of knowing.

Isaac: Assuming the 24pN is still working for you, what other settings do you have dialed in on the camera?

I know those things seem highly unlikely, but it doesn't hurt to consider. The fact is, if the patch is working for at least one person, then operator variables are the cause, at least in part, of failure.
 
I don't think anybody is suggesting that you can never get Native 24p to work. I think that it has been proven, however, that you can't make it reliable. One person even got it to produce a write error when it was the only patch set.

I have lots of data that proves that it isn't working right under any circumstances. Now, whether your camera setup can overcome the issues most of the time - that's another issue.

Ultimately what happens is that the "rogue frames" effectively temporarily double the data rate. Get two or three GOP's with rogue frames in a row and you'll get an error no matter what else you set other parameters to. It's just a matter of time.

If you can tolerate the occasional inevitable error - go for it.

Running the exact same test as somebody else and hoping for a different outcome is the definition of .......... Anybody? Anybody?

Chris
 
^^well insanity. but in this case, i'd say eternal "hope" LOL!

P. Harrill: great post. in order to properly test and document successes and failures, every possible variable must be put on the table so we can figure out what works, what doesn't, what's broken, etc. Then you get multiple people to validate the claimed results by attempting to replicate said results while operating under the exact same conditions (about the only variable that would be different would be the glass but that really shouldn't matter if this hack is ever going to be deemed "golden").

Something that I haven't heard anyone say yet is maybe the physical hardware is simply incapable of delivering error free 24P...
 
Last edited:
The firmware certainly has bugs. I can't imagine why hardware would have a problem with it. After all, 720 60p produces more data than 1080 24p.

Chris
 
Well in the Native 24p problem thread it was clearly established that the Sandisk cards *CAN* mask the problems with the Native 24p. That means that YES they will *work* but the flaw in the Native 24p patch is still there and will cause problems for people with any other card than the Sandisk cards.

I'm not going to run out and buy a Sandisk card for 3X the cost of a generic class 10 simply because it allows me to use a patch that is buggy. It would be better just to fix the bug.

You guys trying to justify using the Sandisk cards, you have to remember that we are dealing with bitrates that should EASILY dealt with by a class 10 card and generally the class 6 cards as well. Having to purchase a special card just to utilize a patch pretty much nullifies the robustness of that patch.
 
This is what happens when you get people testing left and right. That is why big companies get a small gathering of EXPERTS in the field and do their testing for any new product. You can't listen to dozens of results and assertions from dozens of different people that may or may not have the expertise. We need to wait for more work from Vitaliy. He is doing research himself and learning more about what HE needs to do to get this thing to work perfect. People should make their findings known but without making any conclusions about anything before he does.
 
You must log in or register to reply here.
Back
Top