Native 24P/25p settings/problems discussion
- Thread starter NavinRJohnson
- Start date
rambooc1
Well-known member
Yes, i think you also need to "know" what triggers errors and avoid those situations. I see myself making use of sliders or mini dollys for the static tripod shots so that there is always slight movement in the frame, looks better anyway.
I can't see 99% of real world scenes causing any problems, if you know what to avoid and if it is critical not repeatable shot, use L mode at 32000000 or MPEG, it never fails.
Will definitely buy a faster card though for peace of mind, those I frames are still too big.
R
PappasArts
Veteran
Did "Video Bitrate Adjustment Simplified" in the last version of Ptool get named to "Video Bitrate FHD/SH" in the new PTOOL?
Pappas
Pappas
rambooc1
Well-known member
,
Yes, it was confusing. Now everyone should start using the latest PTool and only take notice of usage notes referring to it and not the very early versions.
R
Last edited:
PappasArts
Veteran
So the "Video Bitrarte L" & " Video Bitrate H" are not necessarily needed to be ticked?
Some of my best footage came from 24/60i at ( 64-72-80 15gop 24/60i ) set up in the old ptool- and now in the new Ptool I want to retry these- however now, apply the new patches to see if it makes the high bitrate settings give less write errors on high detail shots. Those setting looked quite scary good on a 18foot screen last week- even though it was played back raw files from camera to projector.
This is what I will try again.
Native 24p/25p= XXXXX- NOT CHECKED
Video Bitrate FHD/SH=64000000
Video Bitrate H= XXXXX- NOT CHECKED
Video Bitrate L= XXXXX- NOT CHECKED
Overall Bitrate=72000000
Limiting Bitrate=80000000
Preset bitrate=640
Preset bitrate 2=640
Pappas
rambooc1
Well-known member
Gezzus, you will burn a hole in the 18 ft projection screen with those settings ..haha
I'm wondering of the I frames stay the same height and the P frames get bigger with those settings, that would be a good outcome assuming you are going to get upper data rates of 40 to 50mbps in high detail scenes.
R
PappasArts
Veteran
( 64-72-80 15gop 24/60i ) was pretty insane good looking even though it's in the 60i wrapper.
That said 50-56-60 24PN might looked just the same as that high bit rate even though it's lower- the fact that 24PN has less frames to deal with and throws more bits at the 24P and its to I/P frames.
Where as the 60 field video has way more frames to contend with- redundant ones- however more. Who knows though. On paper it sounds right- real world it might not matter that much.
Pappas
In theory (well as far as I understand it anyway), I frames sizes should mostly be higher than P frames sizes, especially so on static shots. P frames encode changes in scene, which in static shots are minimal, so I dare to call the screenshot of I/P frame size distribution you posted "normal."
rambooc1
Well-known member
Yes and the less changes the smaller the P frame and bigger the I frame gets. I was hoping (still are) that maybe 1 and 2 preset patches or other patch could increase P and reduce I in static shots, not sure this has been the case, but with native enabled it's definable more stable.
R
There must be some funny business going on in 24PsF mode. As far as I understand, at low level AVCHD encoder reads a flag that tells it whether two half fields or just one full one are used to reconstruct a "frame." So 24 frames split into 48 half frames. That still leaves us with 12 fields need to fill the 60 fields per second. My guess is that every 4th frame is duplicated. I don't think it necessarily means "physically duplicated", as in "taking up bandwidth". It would be interesting to find out what is really going on.
rambooc1
Well-known member
Well that is a pretty common wrapper for PAL and NTSC cameras, so it should be fairly easy to find out with a Google search. Obviously with 25p wrapped in 50i it is much simpler.
Wiki has this http://en.wikipedia.org/wiki/Progressive_segmented_frame
Wiki has this http://en.wikipedia.org/wiki/Progressive_segmented_frame
For pal results of 50-56-60 see below
For pal results of 50-56-60 see below
The low detail of 25pN was OK, but grass sustained locked the camera-froze, battery out, file lost. At low level, spikey P frames still and was into the 80mbits. Sandisk extreme. Also, am thinking pN is showing much more RS at long focal lengths, any reason why? I then went on to test the wrapper version, didnt freeze and good at low level, will progress these settings further incl. your new higher setting Pappas in wrapper. To get pN stable we need to back off imo to much lower rates, possibly 40, 45 50 may be a good compromise.
I need to get my settings logged though because after this new batch of test's have forgotten which version I was originally on and they were super stable and in the wrapper I had no RS in 1080, ahhhhhhhhhhhhhhhhh!
For pal results of 50-56-60 see below
The low detail of 25pN was OK, but grass sustained locked the camera-froze, battery out, file lost. At low level, spikey P frames still and was into the 80mbits. Sandisk extreme. Also, am thinking pN is showing much more RS at long focal lengths, any reason why? I then went on to test the wrapper version, didnt freeze and good at low level, will progress these settings further incl. your new higher setting Pappas in wrapper. To get pN stable we need to back off imo to much lower rates, possibly 40, 45 50 may be a good compromise.
I need to get my settings logged though because after this new batch of test's have forgotten which version I was originally on and they were super stable and in the wrapper I had no RS in 1080, ahhhhhhhhhhhhhhhhh!
rambooc1
Well-known member
One reason i like the C settings with Native unticked, is the data rates stay higher and the height relationship between I and P frames is much better, approximately 2 to 1. With Native ticked the same footage is lower in data rate and the the I and P frame difference is huge.
Going thru both frame by frame, the unticked footage has less MUD and is cleaner overall.
Here is the unticked streameye result.
Going thru both frame by frame, the unticked footage has less MUD and is cleaner overall.
Here is the unticked streameye result.
PappasArts
Veteran
Something that I can swear I see in footage that is in the 60i wrapper too- especially when you crank the bitrates high in wrapper, there seems to be even more benefit. After all the camera was designed to do 24psf/wrapper in the first place.
Scanning rate difference- however I don't think that anything was adjusted there. I don't know... Look forward to seeing how those settings work out for you in the wrapper.
Now I haven't used the new patches for ( 64-72-80 15gop 24/60i ) - so next I'm gonna see if the new patches make this high bit rate in wrapper even more stable.
**
Native 24p/25p= XXXXX- NOT CHECKED
Video Bitrate FHD/SH=64000000
Video Bitrate H= XXXXX- NOT CHECKED
Video Bitrate L= XXXXX- NOT CHECKED
Overall Bitrate=72000000
Limiting Bitrate=80000000
Preset bitrate=640
Preset bitrate 2=640
Pappas
Last edited:
rambooc1
Well-known member
I'm just backing up a bit to C settings, 50i and no preset patches ticked.
I's just more normal looking P and I behavior from the stream in streameye like original Gh1 but at double the data rate. The actual image confirms it.
Good luck with your new elevated rates Pappas.
It's 3-30am in the morning in CA, don't you ever sleep.?
R
Last edited:
PappasArts
Veteran
I saw Chris's tests in the research forum- interesting results. The H L function seemed to be an issue.
Now what about the bitrate presets- do we use both or one? Has that been established?
Important to keep the "Video Bitrate FHD/SH" & "Preset bitrate" the same- or at least that's what were told is best.
Yeah the hi-bit rate settings are topping out there- however I do wonder if the write errors I do get in 60i are truly the card not being fast enough- or it's running into something else. Curious to see if the "Preset bitrate" and setting it to match the "Video Bitrate FHD/SH" setting will make that more stable in 60i mode.
Sleep- What is that?
Pappas
rambooc1
Well-known member
Yes i'm now just going to add preset patches back into the C settings one at a time and at different rates. All in 50i of course and see what the effect is. Not expecting much as confirmed by Chris's tests. Might disable L and H as well.
Given up on native until Vitaliy has advanced more with it's stability.
R
don romolo
Member
Forgive me, I may be a bit naive and my observation may be explained away fairly easily. Feel free to delete or disregard my post due to my noobness.
I just hacked my gh1 and loaded Ipowell's known stable 35Mbps mjpeg settings two days ago. I was playing with the shutter and got a write errors on shutter speeds below 1/30. (I posted in another thread but noone has replied). I don't know if this is documented and people rarely post their shutter speeds in tests! I was using a sandisk extreme III shooting at 4.0, 1/5, iso 400 with Iplowell's known 100% stable 720p 30p settings. tripod shot in the living room. nothing too high detail. class 10 card. I've consistently repeated this with write errors.
So I was wondering if this is in any way related to the problem with 24pN? Might not be very well thought out but my theory is this:
Shutter speeds for video are not mechanic shutter speeds like they are for stills. They are (I think) like refresh rates. 1/60 being like 60Hz.
Frame rates are also like refresh rates. 60p being like 60Hz. (In a video camera -- my background -- 'shutter off' is essentially shutter at 1/60)
30p, 60p, 60i, 50p, 50i and even 25p in some instances seem to be stable. 24pN has virtually no reports of stability.
I will just throw out that there is no shutter of 1/24. There is no shutter that would make the math simpler for 24pN (i.e. 120). If there was a 1/24 shutter speed it would essentially be like the shutter is off and there wouldn't be any more math or processing to be done?
I could be completely off base. Let me know what you think. feel free to yell at me and call me stupid.
I just hacked my gh1 and loaded Ipowell's known stable 35Mbps mjpeg settings two days ago. I was playing with the shutter and got a write errors on shutter speeds below 1/30. (I posted in another thread but noone has replied). I don't know if this is documented and people rarely post their shutter speeds in tests! I was using a sandisk extreme III shooting at 4.0, 1/5, iso 400 with Iplowell's known 100% stable 720p 30p settings. tripod shot in the living room. nothing too high detail. class 10 card. I've consistently repeated this with write errors.
So I was wondering if this is in any way related to the problem with 24pN? Might not be very well thought out but my theory is this:
Shutter speeds for video are not mechanic shutter speeds like they are for stills. They are (I think) like refresh rates. 1/60 being like 60Hz.
Frame rates are also like refresh rates. 60p being like 60Hz. (In a video camera -- my background -- 'shutter off' is essentially shutter at 1/60)
30p, 60p, 60i, 50p, 50i and even 25p in some instances seem to be stable. 24pN has virtually no reports of stability.
I will just throw out that there is no shutter of 1/24. There is no shutter that would make the math simpler for 24pN (i.e. 120). If there was a 1/24 shutter speed it would essentially be like the shutter is off and there wouldn't be any more math or processing to be done?
I could be completely off base. Let me know what you think. feel free to yell at me and call me stupid.
Filmkid2003
Well-known member
Can I do it in person? Just kidding. :happy: