The Best Settings for GH1 MJPEG & AVCHD Hack

Professional Films? Best Settings for GH1 MJPEG & AVCHD Hack

GH13 was made possible by holy Tester13. The input numbers go in his Ptool firmware update tool, that is provided by holy Tester13 and his mighty crew of programmers, code crackers, design fellowers, design testers and cummonity supporters.

What the enthousiast hackers haven't told you, is that the improved Mbit causes serious crashes in the middle of recordings! This commonly occurs with high detailed rich motion scenes at high mbit/s. All cinematographic creative footage is lost if you are not carefull on the hacked settings.

settings

Aster AVCHD Settings | Aster MJPEG Settings


Receive our final optimized settings?

To receive an update of our final conclusions, subscribe below

Email address:

worldwide subscribers

 



Creative Cinematographic Filming with MJPEG settings

- T = 4 4,4,4,4 is the lowest possible setting it seems.
There are rumors, e.g. PappasArts has proven, that the YCC 4:2:2 mode seems not working perfectly yet.
True cinematographic proportions and native sensor halving 1088x612 are now in test.


Creative Cinematographic Filming with AVCHD264 settings

- AVCHD 1920x1080 is sharper than MJPEG 1920x1080. Less color banding, more colour space, smaller files.
- PAL users: convert to NTSC, to achieve 24p miraculously. Plus bit more bandwidth headroom at your disposal.
- Our own test for 1920x1080 stable settings: 52 / 52 / 52 / 60 / 70 and GOP=4 on NTSC formatted class10 Transcend 16GB card Zeiss and Canon 50mm primes at aperture f/2.8 - f/16 to get maximum detail, finding reliability needs really tight aperture = lots of detail and spikes, filming urban life of Amsterdam, Holland. It filmed at average of 27 Mbit/s hitting 40 mbit/s max, this is all ACVHD. Practical and professionally reliable long recordings for those pros out there.
- Our test run for an astonishing 01 hours 47 minutes recording continuously, generating massive MTS footage spanning three 4192 MB files and some more over the fourth one: almost full, had it not been for the battery that died... Green nerds wanting to test this at home, set your lcd OFF after you have composed your boring test shot. Battery used was an original one.
- Due to changing Preset Bit rate patch now even higher bit rates.
- HOWEVER: it seems these massive Mbits do not directly improve smoothness: the video hicks, skips some frames once a second or so, as if the choke limitation of 70mbit doesnt allow to dump all frames... Comparing to smoother GOP 6 or 12, this could mean its wasting bandwidth!?
- GOP = 2 gives the smoothest motion, yet requires heavy frame dumping, almost like MJPEG to so say.
- UPDATE: set GOP to 2 for massively high image quality as well. Our socks flew right off our feet!
Best settings so far are at GOP=2 for video smoothness! 24P native, change the Preset Bit rate under Patches for Testers, default was 234. Also I set the Sensor Mode4 under Patches for Testers to nr7, default was 4, sothat my viewfinder would show 23,97 fps smoothly as well. Stay tuned for more daily updates!

 

Regular Resolution: Filming City Life at Very Tight Apertures:


FPSRESVideo
Bitrate
Overall
Bitrate
Limiting
Bitrate
GOP
Preset
Bitrate
Max Time Filmed
Continuously
Avg!
Mbit
Resolution
& Contrast
Motion Quality
(e.g. for sports)
Hicking
@1fps
These tests below are done one Ultra Resolution Chart (extremely detailed, smoothly moving scene). Hope hacked GH1 survives this torture! If it doesnt lock up here, then your most complex sea/forests will function reliably as well. Attention: this is a NON real life experiment with packed details in motion!
25FHD16182015234Till battery dead17 Default = BADNormalNone
24FHD5862701!!580CityLife: > 1min okay59AstoundingVery badhicks
24FHD5963711!!590CityLife: > 1min okay61AstoundingVery badhicks
24FHD5963724570CityLife: > 1min okay56 Bit distortedSmoothok
24FHD5963722590CityLife: 8 Sec & crashXXXX
24FHD6061802500CityLife: Failed XXXX
24FHD6066716600CityLife: Failed XXXX
24FHD6064701!!600CityLife: Failed XXXX
24FHD60657512550CityLife: Failed XXXX
24FHD6466684500CityLife: Failed XXXX
24FHD60708012600CityLife: Failed XXXX
24FHD55627112530CityLife: +1min okay24GoodNormal 
24FHD5562712530CityLife: +1min okay43Astounding Bad Hicks
24FHD5663724540CityLife: +1min okay31GoodSmooth 
24FHD56637212540CityLife: +1min okay21GoodSmooth 
24FHD5862704580CityLife: +1min okay31GoodSmooth 
24FHD5862724580CityLife: +1min okay32GoodSmooth 
24FHD5565713520CityLife: +1min okay33GoodSmooth 
24FHD5568724550CityLife: +1min okay34GoodNormal 
24FHD4068701!!400CityLife: +1min okay41AstoundingNormalExtreme
24FHD5068702500CityLife: 13 Sec & crash45AstoundingVery Smooth?
24FHD5268742520CityLife: +1min okay42AstoundingVery Smooth?
24FHD5368802530CityLife: +1min okay45AstoundingVery Smooth?
24FHD5568902550CityLife: bit unreliable47AstoundingVery SmoothSlight
24FHD55688212550CityLife: +1min okay24GoodSmooth 
24FHD5570904550CityLife: +1min okay30GoodSmooth 
24FHD5770904570CityLife: +1min okay32GoodSmooth 
24FHD5667962560CityLife: 11 min & crash42AstoundingVery SmoothMinimal
24FHD5667702560CityLife: 12 min & crash40AstoundingVery SmoothMinimal
24FHD5464682540CityLife: 4 min & crash51Astounding!SmoothMinimal
24FHD5760962570CityLife: 2 min CARD ERR52Astounding!SmoothLot
24FHD5360962530CityLife: >26 Min50Astounding!!!Smoothest!!!Minimal

 

Ultra Resolution Motion Chart @ GOP=2 = Smoothest Motion!

 

FPSRESVideo
Bitrate
Overall
Bitrate
Limiting
Bitrate
GOP
Preset
Bitrate
Max Time Filmed
Continuously
Avg!
Mbit
Resolution
& Contrast
Motion Quality
(e.g. for sports)
Hicking
@1fps
All tests below are done one Ultra Resolution Chart (extremely detailed, smoothly moving scene) f/5.6. Hope hacked GH1 survives this torture! If it doesnt lock up here, then your most complex sea/forests will function reliably as well. Attention: this is a NON-real-life experiment with packed details in motion!
24FHD5360642530UltraRes: 3 sec & crash57Astounding Lot
24FHD5360992530UltraRes: 2 sec & crash62Astounding bad
24FHD5360612530UltraRes: 2 sec & crash57Astounding bad
24FHD5063642500UltraRes: 2 sec & crash66Astounding bad
24FHD4060642400UltraRes: 11 sec & crash42  very bad
24FHD5658602560UltraRes: 1 sec & crash62Astounding bad
24FHD5051702500UltraRes: 7 sec & crash52Astounding bad
24FHD4050962400UltraRes: 2 sec & crashX  X
24FHD3040702300UltraRes: 1 sec & crashX  X
24FHD5667962560UltraRes: 6 sec & crashX  X
24FHD6061962560Immediately crashedX  X
24FHD3062702400UltraRes: 7 sec & crashX  X
24FHD4042602300UltraRes: 8 sec & crash42   
Fucking Hell GOP = 2 is so unreliable! Bye Bye Super Smooth Motion. Lets continue using regular GOP settings, at our ultra resolution motion test:
24FHD5060964500UltraRes: 3 sec & crashX   
24FHD5052604500UltraRes: 11 sec & crashX   
24FHD4050604300UltraRes: 11 sec & crashX   
24FHD4050706400UltraRes: > 3min OKAY30AlrightAlrightNone
24FHD5051706400UltraRes: > 3min OKAY31   
24FHD5060964unsetUltraRes: 7 sec & crashX   
24FHD50609612400UltraRes: 7 sec & crashX   
24FHD50526012500UltraRes: 1m 10s crash    
24FHD40424412300UltraRes: 1m 30s crash28AlrightAlrightNone
24FHD50525424400UltraRes: 10 sec & crash36   
24FHD50525448400UltraRes: 10 sec & crashX   
24FHD50525472400immediate crashX   
Conclusion: GOP must be between 1 - 48. Interesting! Perhaps it was not the GOP 2 that was causing the choke, but just these absurdly high detailed motion scenes at tight apertures. Good news: we CAN use lower GOP and achieve very high Mbits AVCHD, as long as, bad news, we stay out of extremely detailed scenes, since this latter causes the crashes / write speed error / card error. By the way, who is going to turn a tight aperture on a fast prime lens anyways? Lets redo the tests but with wider apertures mimicking more real world bokehs at f1.4 - f/2.8 on our wide primes, on tele you can!
24FHD3040502300UltraRes: 43 sec & crash30GoodLovely!Damn Hicks !!
24FHDunsetunsetunset4unsetUltraRes: > 2min OKAY16Bad Hicks !!
25FHDunsetunsetunset2unsetUltraRes: 12 sec & crashX GoodGood No hicks
24FHDunsetunsetunset2unsetUltraRes: 1m 9s crash16BadExtremely Good!Hicks
24FHD3236402500UltraRes: 47 sec & crash33 Extremely Good!Hicks
24FHD32364024500UltraRes: 42 sec & crash16OkayGoodNice No Hicks
Above line are exact doubles of the default: 16>32 / 18>36 / 20>40 15>24 and 234>500 turned out okay in terms of quality but unreliable at this test.
24FHD32364024700UltraRes: > 2min OKAY27GoodGreatNice No Hicks!
24FHD40424424700UltraRes: 22 sec & crash32GoodGreatNice No Hicks!
24FHD40424424550UltraRes: >2min OKAY31GoodGreatNice No Hicks!
24FHD32424424550UltraRes: 31 sec & crash??? 
Lets try HD 720p, hoping for massively higher Mbits (=quality) making camera more reliable in extreme detailed motion scenes (trees, water ripple, etc)
24HD40424424700UltraRes: >2min OKAY20BadAlright!Nice No Hicks
24HD50607024720UltraRes: >2min OKAY22BadAlright!Nice No Hicks
Such bad quality at 720p! forget AVCHD 720p folks: FHD is the way to go. Now lets get those unreliable settings correct for extreme scenes:
24FHD50607024720UltraRes: 31 sec & crash32GoodAlrightHikcing
24FHD16505224700UltraRes: >2min OKAY15 !!Muddy  
24FHD40426024700UltraRes: 5 sec & crashX   
24FHD40414224700UltraRes: 9 sec & crashX   
24FHD36687024700UltraRes: 5 sec & crashX   
24FHD36374024700UltraRes: 4 sec & crashX   
24FHD36374024420UltraRes: 5 sec & crashX   
24FHD363740unset500UltraRes: >3min OKAY33GreatAlrightNone!
24FHD505260unset700UltraRes: >3min OKAY34GreatAlrightNone!
24FHD525456unset700UltraRes: >3min OKAY47Astounding!!AlrightNone!
24FHD525354unset670UltraRes: >2min OKAY46Astounding!!AlrightNone!
24FHD505152unset700UltraRes: 32 sec crash46Astounding!AlrightNone!
24FHD485058unset720UltraRes: 58 sec crashXXXX
24FHD485052unset570UltraRes: 1m 14s crash36   
24FHD474960unset700UltraRes: >6min OKAY37   
24FHD485258unset670UltraRes: >27min OKAY    
24FHD465258unset590UltraRes: 58 sec crashX   
24FHD47484924500UltraRes: 21 sec crash39   
24FHD474849unset500UltraRes: 10 sec crash45Astounding!!AlrightNone!
24FHD4747,147,2unset500UltraRes: 10 sec crashX   
24FHD485258unset960UltraRes: 12 sec crash41Astounding!!AlrightNone!
24FHD484848unset500UltraRes: 32 sec crash    
24FHD484848unset500CityLife: > 15min !! OK    
24FHD464850unset500UltraRes: >2min44Astounding!!AlrightNone!
24FHD444650unset520UltraRes: >10m RELIABLE40Very GoodAlrightNone!

 

Changing Native Resolution @24 fps

 

FPSRESVideo
Bitrate
Overall
Bitrate
Limiting
Bitrate
GOP
Preset
Bitrate
Max Time Filmed
Continuously
Avg!
Mbit
Resolution
& Contrast
Motion Quality
(e.g. for sports)
Hicking
@1fps
Experiment with changing Sensor Mode 4 Adjustments: W=sensor w | H = sensor h | RW = raw w| RH = raw h (default settings in red row below)
25FHD16182015234Till battery dead17w=4400 | h=2478 | rw=1119 | rh=984None
24FHDunsetunsetunset24unsetNot a single Effect!?/w=2200 | h=1239 | rw=960 | rh=844 
24FHDunsetunsetunset24unsetNot a single Effect!?/above & sensor h cut = 500 w cut =0 

 

Conclusions & Recommendations for Experts Filming AVCHD on 16GB SDHC class 10 card:

 

FPSRESVideo
Bitrate
Overall
Bitrate
Limiting
Bitrate
GOP
Preset
Bitrate
Max Time Filmed
Continuously
Avg!
Mbit
Resolution
& Contrast
Motion Quality
(e.g. for sports)
Hicking
@1fps
Aster advices for family/holiday/long play/conferences/students/presentation videos
24FHD32~5234~6060~704~24500~55001h47min (batt)± 27GoodReliable @ Low/Normal Res ScenesTiny
Aster advices for professionals/advertising agencies/filmmakers/documentary/70mm Cinema Fans: now testing...
24FHD42~5044~5246~58unset testing... testing...Reliable @ High Res Scenes 


Have a favourite settings you want to include? We are looking for GH1 settings that are:

- Reliable high quality modes that are edged towards an extreme

- Cinematographic modes, e.g. 1920x800 or 1080x400
- Sensors native downsizes from 4352 x 2448 ,16:9, to 1088x600 native sensor downsizings etc
- Famous settings that are widely used on the internet, reliable for pros
Please provide your video size width x height, MJPEG settings, eg Q/T, or AVCHD settings, videobitrates/GOP, and nickname and post it to the official forum, see link at the top of page and subscribe for updates.

Receive our final optimized settings?

To receive an update of our final conclusions, subscribe below

Email address:

worldwide subscribers