Tragic Lantern for EOS M

Started by coutts, April 17, 2013, 01:43:28 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

1%

Didn't work for me. Dot problem might have a solution from 650D

qsara

Quote from: 1% on June 15, 2013, 05:57:27 PM
Didn't work for me. Dot problem might have a solution from 650D

in my last tests I'm not getting any AF dots on frames, are there other kind a dots that I'm not aware of?

1%

This would kill dots without enabling af_raw and ruining sharp edges.

mixer2

do i understand it correct, that the solution from 650d might work for eos m to get best from both raw types? no dots and no edge detection? that would be great, hope there is soon a version to try.
btw with handheld night i still get the same 32+32+29 = 93, as with multishot nr. but few MB more or less don't make a difference anyway. and handheld night is also just available in photo mode. as long as you get continuous recording on maximum card speed, the benefit is some stability (but don't think few mb make a big difference) and some more frames when recording at a rate that exceeds your card (or controller) speed.
the bug with photo mode raw rec activation, when switching from video to photo mode is nearly fixed in the current version. but there are some combinations, that still cause problems, i'll make a video to demonstrate later. hope that helps to reproduce the bug.

1%

There is one bug I caught this morning that fake movie mode wasn't turning off. Maybe related to that.

qsara

in addition to that bug, anyone experienced a problem with EOS M power-off? because when I turn off the camera, it still stays hot until I remove the battery.

1%

Weird.. I have been powering off and on and no drain or heat.

qsara

mine maybe a faulty camera..

1%

I wouldn't say that yet, maybe some setting or bug.

xpstudio

@1% I'm testing [EOSM] BugfixNewSystem.zip

In photo mode: load modules -> nr -> enable raw (1600x544, 24 frames), camera turn off after 2 seconds of recording

Works fine with [EOSM] FixUnhack.zip and frame rate 24
EOS M 22mm | 18-35mm

1%

I tested on 29MB/s card and it stops/runs out of buffer then unhacks. BTW, need to tweak timer A a little I think.. 510 is too low, screen goes black.

xpstudio

Is there any option to automatically restart recording after 4GB? I tried "Movie restart' but it doesn't work.
EOS M 22mm | 18-35mm

qsara

Quote from: xpstudio on June 15, 2013, 10:46:22 PM
Is there any option to automatically restart recording after 4GB? I tried "Movie restart' but it doesn't work.

format your card as exfat, use cardtools to make it bootable, it can record over 4GB

mdwhaley

I'm still using FixUnhack. The max at 23.976fps I can get is 2.35:1 1728x448 continuous. 1728x512 2:1 said it would work and worked for awhile but eventually stopped itself after maybe 2 minutes. This is on a Sandisk 45 MB/sec card.

A problem I am having now though is the files I got are 4.29 GB and raw2dng says they aren't valid. RAWMagic doesn't like them either. They add .R00 files when they go over 4GB. Am I missing something here

qsara

have you tried eyeframeconverter or batchelor ?

1%

Its supposed to split after 4gb on fat32 cards.

mdwhaley


mdwhaley

Edge artifacts aside, there is a shocking amount of dynamic range



qsara

Quote from: mdwhaley on June 16, 2013, 05:33:38 AM
Edge artifacts aside, there is a shocking amount of dynamic range



well, yeah, it is raw.

mixer2

i can confirm, that the edges seem to be much smoother. (Edit: edges are not smoother, was just the angle) but i still have small edge detection dots, which are much less disturbing than the unclean edges. looks nice already! have to do some videos and test around with postprocessing to check how good the dots can be removed in post and how disturbing they are. but i think this may be the first version were eos m raw video quality may be better than h.264 in all points.

qsara

Quote from: mixer2 on June 16, 2013, 10:30:32 AM
i can confirm, that the edges seem to be much smoother. but i still have small edge detection dots, which are much less disturbing than the unclean edges. looks nice already! have to do some videos and test around with postprocessing to check how good the dots can be removed in post and how disturbing they are. but i think this may be the first version were eos m raw video quality may be better than h.264 in all points.


it was always better than h.264! what are you really comparing here? h264 is  50mbit/s RAW is 300mbit/s! just think about the amount of data!

qsara

Quote from: mdwhaley on June 16, 2013, 04:23:51 AM
I'm on MAC

I'm on PC, sorry. you can virtualize windows if you have mac PRO.

mixer2

the amount of data isn't a qualified factor for judging quality at all. the amount of data says nothing about the amount of information. and especially it doesn't say anything about the amount of useful information.

qsara

Quote from: mixer2 on June 16, 2013, 11:06:43 AM
the amount of data isn't a qualified factor for judging quality at all. the amount of data says nothing about the amount of information. and especially it doesn't say anything about the amount of useful information.

oh really, extract frames on h264 video and try editing them as RAW :) I'm waiting.

and to answer your further posts, please watch this then try to talk again about h.264

http://vimeo.com/52269416

mixer2

i never said. that you can't get better quality with raw, i just said, that recording more data doesn't directly lead to better image quality. and the image quality of the raw video possible on eos m just wasn't better than h.264 so far, because of the distorted edges and the af-dots.

hurm, i'm still not completely sure about the edges in the current version. made a test with the table edge yesterday and got cleaner edges as in the test before. but today tested with a flower the edges wasn't really cleaner than with the "[EOSM] FixUnhack.zip"-Version.
have to do some more tests to see if the quality improved or just anything in the test went wrong.

i did the test with the table edge again and the stair step effect does seem to depend more on the angle than on the used version.
versions i used from top to bottom: [EOSM] BugfixNewSystem.zip, [EOSM] FixUnhack.zip, [EOSM] UsableRaw.zip


all 3 seem to have an identical amount of stair step effect. But it's not as strong as the bottom image of my last test and not as weak as on the top image: