Driftwood ClusterX series 2:│moon│ЅріzZ│nebula│drewnet│Slipstream │Redshift

sleewok

@Alight Good thing you were using the hack. I don't think the standard GH2 would pick up the definition in your arms ;).

ALight

@sleewok lol!

4CardsMan

Nebula Update - I ran into a problem with two clips out of five on a Sandisk 64/95 card. This clip would hang at 98% on import into FCPX, even after copying the card to a folder on my hard drive, so I converted it with Clipwrap. The QT file imported into FCPX, but the image broke up in several spots, as in the attachment. Playback in Quick Time shows the defect, which is where I got the screen grab. Interestingly, the original mts file will play in MPlayerX with no problem. In this clip, our dog, Buddha, was writhing in the grass after an encounter with a stink bug, so it appears that motion is a factor. However, another clip with very little motion shows breakup as well. Leica Summilux D, ISO 200, Standard profile, all -2.

driftwood

@CurtisMack Was the card pre-formatted out of interest?

haarec

@driftwood isn't it a Frame Limit issue? The problem described here http://www.personal-view.com/talks/discussion/comment/119118#Comment_119118 looks alike.

nemvan

Spizz trial 5 does not span neither on 24H nor on 24L setting using 30 MB/s Sandisk Extreme. (Tested with "deadly TV war of the ants".)

itimjim

Used Moon Trial 5 this weekend for a 48hr film short. Mixed feelings and it's making me thinking in doing something wrong.

First thing, random frames are breaking up around the bottom 1/5 of the image, where it just seems to stop recording and fills with a vertical smear to the bottom. Happens in about 1 in 12 shots. I'm shooting 24p using SanDisk 64GB/95mbps.

Then there's the noise. It's nice looking noise, but at ISO 200 a bit off putting, even in what are fairly well exposed areas. I was using middle row ISOs, mostly 200. ISO 1600 looked almost as clean as 200.

One thing had now got me thinking. I remember @shian mentioning something about it. I was manually setting WB rather than using the K scale. I think he mentioned a nose big seeing manual WB. Anyone know if that's true or not?

Well post some images later for examples. Cheers.

driftwood

@haarec @CurtisMack Either bad transcode error or playback/NLE errors interpreting the elementary stream. Proven time and time again. There's a thread about the slanting stuff. The players/transcoders are unable to support proper deblocking set by the GH2 encoder most of the time.

FYI Those same errors have been experienced on other ptool settings you'll note - not just Driftwood settings.

Always check the originals first in Movist on the MAC/OSX and Windows Media PLayer on the PC. VLC breaks up. Premiere Pro has problems. Ive yet to have these problems with my recordings and fcpx. Clipwrap (rewrap to Quicktime) then import to FCPX.

@itimjim Dropbox link me an original with the error.

omnidecay

@thepalalias like always, good advice! Its personally frustrating to me when there is a hidden factor not known to me. Part of me is like come on, these hacks are awesome and stop thinking too much. The other part of me says sure they rock but they should span. I really think it has more to do with the camera vs the cards but I am not knowledgeable enough to even begin to diagnose the issue. Thanks for the advice on FM2, I'm pretty sure thats the one I will go with.

@driftwood As soon as my battery is charged I will span test all 3 hacks in 24L vs chart'o'death.

@Zaven13 Yeah, I am going to have to. I rarely switch out of 24H so I sometimes forget about 24L lol. I honestly am not sure why Panasonic even included a 24L in this camera. Yeah, I can switch to a lower bitrate quickly but how often would anyone do that?

maddog15

Slipstream #1 Testting Update

Spanning Test: Left camera running pointed at a television so there would be visual changes, motion, color etc. Used Panny 14-140 Lens. Used AC power adapter.

Used the very affordable SanDisk Ultra SDHC 32GB 30MB/s card shooting 720 60p Low (H). Camera "beeped" and stopped recording after just over an hour. No crash. Camera just went back to "standby" mode. 5 MTS files spanned perfectly.
Used SanDisk Extreme Pro SDXC 64GB 95MB/s Card shooting 720 60p High (SH). Filmed for about 45 minutes non stop before manually stopping record. All files spanned. No Problems.

Really like this patch. Will either move to another Slipstream or DREWnet next. Really anxious to try that one after watching his review of it on the DREWnetwork. Kindly gave great props to Driftwood. (Search on Youtube if you've not seen his episodes. Really enjoy his stuff and have learned a lot.)

driftwood

@omnidecay Thanks for doing the test. PS Span test FM202 whilst youre at it on all modes with the death charts. Ta. Don't forget to try out Slipstreams too.

Its personally frustrating to me when there is a hidden factor not known to me. Part of me is like come on, these hacks are awesome and stop thinking too much. The other part of me says sure they rock but they should span. I really think it has more to do with the camera vs the cards but I am not knowledgeable enough to even begin to diagnose the issue. Thanks for the advice on FM2, I'm pretty sure thats the one I will go with.

They should do but its not necessarily just down to sdcard speed - its down to how the buffers are able to address the 4Gb limit when it reaches it and onto the next span file creation.

I'm not sure you are pretty sure thats the one to go with. Do more tests.

4CardsMan

@Driftwood - Re: Nebula update. Originally, the clip would hang at 98% when importing to FCPX, whether from the card, or after copying to a folder on the hard drive. After Clipwrap, it would import to FCPX, but showed the breakup. The mts , which I will supply on request, does play properly in MPlayerX. Nebula seems like a very workable patch, so I would like to definitively identify the problem.

BTW, I've now figured out how to get in-camera playback: do a short scratch clip first.

omnidecay

3 tests down and two to go...All tests done with SanDisk 8gig 30MB/s 24L:

Drewnet 8 - no span
Nebula 6 TrailB - no span
Spizz - SPANS!! get about 8MB/s

@driftwood is there to increase buffer size within the camera? Someone needs to build a camera that allows me to just hook my lens up to my computer lol.

driftwood

@CurtisMack I did tell people about recording a small file before switching off :-) My guess is something's gone awry. So you copied over all the Private folder first to your hard disk, then you clipwrapped rewrapped to Quicktime .mov or did you export to pro res? Its strange because Ive never had probs with Nebula.

@omnidecay Can you now span test the Slipstreams for me. Thank you. This is only on your 8gb card yeah? Remember each card can have different qualities. Sometimes cards are duff too - ie - not reporting the quoted datarate read/writes and some cards are better if they are of different size 16GB/32GB or 64Gb despite reporting the same data rate. Each card has its own personality so you are going to look for the best setting that likes your card? Am I right? As to the buffer size, its down to what is the datarate is at the time of the end of 4Gb limit (write EOF end of file) next file creation - continue write. If the image being recorded at that very point in time is bottling the buffer and maxing out then this could (or could not) be responsible for a span. Other users with different cards or sizes report success/failure. Nothing is written... 'Nothing is written'. Remember my words. ;-)

omnidecay

@driftwood

SanDisk 8gig 30MB/s 24L :

Slipstream1b - spans 4.75mb/s
Slipstream 2 - spans 6.13mb/s
Slipstream 3 - spans 10mb/s!
FM2 24L - spans 4MB/s
FM2 24H - did not span

I had a feeling that Slip3 wasnt going to span but I was sure happy when it did. So far thats the largest bitrate I've been able to span with. I'm stepping out for a moment but when I come back I'll do flowmotion2.

overexposure

Did a test today and have really sharp results with 24P High bitrate Slipstream #3: Release Date: 15/03/2013. Same high quality as HDMI out to atomos ninja2 Pro res HQ. The sharpness is just much lower on 25P 71Mbit HBR. Would be perfect to have 25P with the same settings as 24P

see attachements Screendumps of slipstream #3 video stills in 24P 150Mbit, 25P 71 Mbit ,25P Prores HQ HDMI Ninja 2

haarec

@driftwood I mentioned about a Frame Limit because the problem looks similar also to what happened to Ralph_B using Sanity 5.0 and caused an upgrade to v5.1 Apparently Cluster X does not have a Frame Limit set at all. When it comes to players, Windows Media Player is ok, how about Media Player Classic with K-Lite Codec Pack?

driftwood

@haarec Nothing to do with Frame Limit. Ralph used something very close to my Intra Frame limit value for his Sanity 5.1 correction if I remember to sort out an issue he had. Only Intra settings have this limit.

haarec

Yea, I thought Sanity 5 is a long GOP (18) setting.

dado023

what is "spans"?, is that when when it splits file after certain file size?

overexposure

is it true that 24P always looks better then 25P as the GH2 is a 24P camera? or is that history since the latest firmware?

greetings marc

haarec

@dado023 SD card file system can receive a 4GB max file. When you are recording lets say 20 minutes non stop, the file that camera produces would be bigger than 4GB, so the camera divides a recording to several 4GB parts, which is called that camera spans :-)

dado023

@haarec

thanks for the response....heh, new answer opens a new question, SD card file system is exFAT which should allow file size more than 4GB, actually it should be the same as NTFS regarding file size, so i wonder is that maybe rather GH2 limitation, since exFAT can accept big file sizes? Basicly i don't understand why are spans needed when exFAT overcomes the 4G file size limitation of FAT?

omnidecay

@dado023 like you said I believe it is a camera limitation.

itimjim

@driftwood I've just fired up the same clips that tear in the bottom 1/5 of the frame (only one frame) and they're fine :-/ ? Only difference between the two machines I used is that one uses NVidia (the one that tears) and the other ATI GFX card. I thought it might be the codec, but I transcoded to ProRes 422 and it's the same in that as the MTS. Anyone see anything weird like that?

I'm going round the chaps house later, so I'll try and grab a screen grab from the machine that shown the problem.

Howdy, Stranger!

Categories

Tags in Topic

Top Posters