This is outdated. Please have a look here if you want to up/downgrade your camera.
This is a test of the new firmware @nabilfathi15 and I developed together.
You can download and install the firmware here: Olympus OM-D E-M1 custom video firmware
The firmware allows shoot better video. More dynamic range, better ISO and more...
But PLEASE test the firmware for yourself before you judge it.
We are working all day to improve this and other features.
Please respect the work, thanks for your support
Technical Specs of the video:
Shutter 1/4000 - ISO 3200 - No Grading - 40/150mm Lens
Camera : Olympus OM-D E-M1
Head Image
How's the difference on lower ISO? The "No full read" part looks absolutely crap in this test - how does it compare using ISO 200/400?
ReplyDeleteBoth are same by quality, but lower is shot further and miss details because of that.
DeleteNo, what is a crop sensor, the other is a full sensor read, that the difference.
DeleteIf you don't understand how a sensor work, don't comment.
Is it possible to change firmware number to 2.0a or 2.1 to distinguish which firmware is flashed when in camera menu?
ReplyDeleteAll the luck for your work guys.
ReplyDeleteBut why is the lower video so bad quality, when compared to what the E-M1 can shoot with official firmware? I would understand if these were 100% crops from 4k video but if these are just 100% crops from 1080, I don't understand.
Hi,
ReplyDeleteThank you for working on that, I really crave for slow-mo feature to document animal behavior.
But... I honestly don't understand that you put this sample as showcase: like people above I think the "lower video" is so bad it cannot be used to judge anything in comparison, whereas the (original) E-M1 is capable of so much better. And the video maker is a professional. Moreover it looks like both videos were taken at different distances, it makes any comparison even more difficult, don't you think ?
You ask people to try the firmware before judging it, but IMHO by providing such an example you are going to make them reluctant instead, it's a pity.
I strongly suggest that you make a significant test, using a specific scene with both videos taken in the same conditions, at the best quality you can reach in both cases.
Good luck, thanks for the effort anyway !
The best test would be 2 E-M1 side by side. But who has 2 E-M1 just layin' around? :)
DeleteThis comment has been removed by a blog administrator.
ReplyDeleteLooks like the full sensor read is 4K - this is the sort of difference one would expect to see between 4k and 1080. Very hard to tell from this video.
ReplyDeleteI have flashed to my own EM1 and have to say it is very hard to see any differences under any circumstances, BUT not saying I dont believe, just possibly the upper one was actually done with the real 4K firmware.
To aid in getting people onboard, you should think about setting up a test scene. Shoot the test scene with the stock 2.0 firmware, then shoot with your customized revisions.
ReplyDeleteThat way you can have a baseline, then compare your work against it.
This will:
1. Aid in your development
2. Give potential users something substantial to view and help gain traction for your efforts and maybe some support $$$.
Test scene recommendations, use stuff you can stand to loose. Get a pice of cardboard, put newspaper on it, glue your scene to it, that way you can move it out of the way if you need too.
Use the same tripod, lens, focus point on each shot.
Set up ideas:
1. desk lamp for light.
2. Page from a newspaper, to test for detail, sharpness, and moire.
3. Any type of ball that will cast a shadow onto the table, to dest dynamic range.
4. Something with color, a color chart print out, box of crayons, easter eggs, whatever you have.
5. Clear drinking glass, with or without water, to test for chromatic abbreviations when the lamp is lighting from top or behind.
6. Anything cool to look at or interesting that you think can show viable differences in the image results.
Post your results, do a youtube video comparison. show each in it's full, then a over and above combination. 10 or 20 sec a take should be fine.
Also you can post ooc .mov's so users can see the originals. Use mpegStreamClip to trim your files with out re-transcoding. Post to cheap cloud storage or your server, or send to me I'll host them for you.
I think you still don't understand.
DeleteLike i said the last time.
Full Sensor Read is the Custom Firmware
No Full Sensor Read is the Original Firmware
To clear this out, higher sensor read give you better pixel
but don't change the internal encoder.
Both version of the video use the same lens, the same distance (not the same position) but please take a look clearly to the video, check the corner, etc...
I'm out of my budget, i'm working with my own OMD E-M1 day and night, i take the risk, i make the video and i do the test (i usually never use tripod or my own material for the video)
I don't have tripod but only a monopod, i will not buy any tripod for now, just to make a video that enough to show the difference.
You are not going to teach me how to do my test.
I clearly stated on the video
Full Sensor Read 4:4:4, Internal Encoder 1080P 4:2:0
No Full Sensor Read 4:4:4, Internal Encoder 1080p 4:2:0
One look more blurry than the other because on doesn't use the Full Sensor Read.
That all.
Let's stop here the speculate or critic the way how i do my test.
If you don't want to use the custom firmware, don't use it, if you are not happy, well i can't do anything.
Here again, i work my own E-M1 day and night, that i was not supposed to work on something like this but because i want to help, well i do my best, so stop asking more than i can do.
Thank you
I can't replicate any the told benefits. Why so?
DeleteThat all the problem.
DeleteYou need to read correctly what the benefits is about.
And yes you can replicate the told benefits on extreme condition with lot of information.
You will not see the difference by recording your cat or just a piece of glasses, etc...
Go outside, go record a lot of information, watch the sky, watch the light etc...
I think people who can reproduce are just people who don't want to see it or check clearly.
The beneficit is there but we could have more beneifict by rewriting the encoder part.
The quality difference when changing the way it reads from the sensor (whole frame or line scanning) should not depend on the amount of information in the picture. That has to do with the compression and the quality of the codec.
DeleteSo if you can't see a difference even in a static scene with high frequency information (close to the resolution of the camera) then there is really no benefit.
I am not sure how the above video is created and why there is so much difference, but one thing is certain that it is not just because of the released hacked firmware since nobody has managed to see any difference.
Let make it simple for you and the others. Because apparently you do not realize how much a Full Sensor read can help in term of details.
DeleteTake two shot, one at 3840x2160 compressed to 1920x1080 and the other take one at 1920x1080 compressed in 1920x1080.
Which one would give you more details ? exactly same size, same compression.
There again, you don't know what Full Sensor read mean.
Nobody has manager to see the difference because we talk about details.
Most people in there just want recording in 4K, this as nothing to do with recording in 4K
You will not see as much details than in 4K but you will take more details than it would be with a not full sensor read.
Is like comparing F1.8 to F.10 for example and say that have no difference.
It is depending of the information in the picture, i can show you two shot, one in 4K and one with a good 1080p camera on a specific scene, you would not see the difference.
Again, let's make it clear, i choose to film the grass (colors, reflect, light, etc...)
If people do the same stuff and correctly, they would see the difference.
Until now i see no one show that they did really test the firmware or make a real test.
Most of them trash talk, that all
I do realize how much a Full Sensor read can help in term of details, and that's why I doubt that the hacked firmware that is published actually does it.
DeletePersonally I don't care for 4K. Whole sensor readout with much improved bitrate is enough.
I appreciate all the work that you put in this but unfortunately the only person that I have seen to trash talk is you.
Please stop being so defensive and don't stop working on the firmware. Keep in mind that EVERYBODY would like to see this project work out. So if people don't see a difference and you do, there is probably something wrong with the published firmware and it has to be fixed.
Please don't take my comments as an insult.
ReplyDeleteI was just offering some helpful suggestions to help you "showcase" the work you're doing on improving the quality of the video.
Having a standard image as a baseline for comparison can be beneficial in a number of ways.
I don't take as an insult.
DeleteI just point out that we already know,
But at the moment i have no choice, i use almost 100% of my empty time
to work on the custom firmware, i don't have time to think about only testing, i also need to work on coding, fix issues, check the correct string.
People think it is so easy to just add their features like that, when is not, it take time and even more when you are the only working on this.
Anyway thanks for your support
If all goes well I should have my E-M1 in a week or so, and I'll help out with testing.
DeleteI appreciate your efforts.