filmov
tv
Rec 709 vs OM-Log400 - E-M1X Video Contrast Test

Показать описание
Here's a small handful of comparison shots I managed to take the other day between the standard Rec 709 video profile vs the new OM-Log400 log gamma profile in the Olympus E-M1X.
This is a warts and all comparison in which I've purposely shot where there's extreme contasts i.e. strong sun, and deep shadows just to see how the new OM-Log400 behaved. No sharpening or noise reduction has been applied to the output (essentially it is unmastered) because I wanted to see what I was dealing with. Only a basic grade has been applied to the log profiles - and it is to my taste. No LUT was used in the grades. Clips were compiled and graded through Davinci Resolve. Footage was captured externally to an Atomos Shogun Inferno. However, as is always the case, it looks like youtube is doing its own internal compressing thing.
You may see some odd things in a couple of the shots. In the first shot there is an area of softness around the upper left of the frame - unfortunately, the ND64 I was using got scratched really bad as I was setting up and so it affected the image there. More noticeable when I was shooting at a smaller aperture, but tried to stay around my lenses' widest apertures f/1.8 and f/2.8 mainly.
Exposure was relatively simple. Exposed to 100 IRE and then take a 1/3 step back via aperture - so basically an uneducated ETTR as I didn't know at what IRE clipping occurs yet. Shutter speed was set at 1/48th @ true 24fps. OM-Log400's base sensitivity is ISO 400 which is good in that it's very clean and you don't have to go so crazy with NDs. I noticed that I only ended up using an ND8 and ND64 - even in relatively bright situations, ND500 was a bit strong.
A couple of things that I noticed once I got the footage back to my computer at home. OM-Log400 seems to clip highlights at a value greater than 100IRE. Also, it seems to also retain values below zero (at least that's what my waveforms are telling me at the time of exposure as well as through my scopes on Davinci Resolve). I was able to pull back highlights from a lot higher on my waveform than I expected to so perhaps I could ETTR just a touch further.
I haven't done any scientific tests - others will I'm sure. I found that I was exposing around 1.3 - 2 stops over the exposure I was getting for Rec 709, and in post, I was able to pull down the file quite easily using the gain wheel. All that said, there is a very good improvement in dynamic range by having access to a log gamma profile.
As far as log gammas are concerned, it's not the flattest I've ever come across. In saying that, it's probably the most responsible 8-bit log gamma I've tinkered around with. Let me explain...
I've graded other camera's log gammas in 8-bit in the past and whilst some technically do have more dynamic range, the file's malleability is particularly compromised due to the combination of a wider dynamic range vs a limited number of values that could be assigned to address it i.e. 0-255. I found that I was able to stretch the Olympus file a bit more without the file breaking down so quickly compared to other 8-bit log gammas I've used in the past.
Definitely possible to get some really good footage from this camera. Now that I have a bit of a better idea of what I'm looking at exposure-wise, I'll probably post something more interesting going forward. This was a bit more of a stress test and you can definitely see that the log footage does capture more data than the default standard Rec 709 profile. So stay tuned in the near future.
This is a warts and all comparison in which I've purposely shot where there's extreme contasts i.e. strong sun, and deep shadows just to see how the new OM-Log400 behaved. No sharpening or noise reduction has been applied to the output (essentially it is unmastered) because I wanted to see what I was dealing with. Only a basic grade has been applied to the log profiles - and it is to my taste. No LUT was used in the grades. Clips were compiled and graded through Davinci Resolve. Footage was captured externally to an Atomos Shogun Inferno. However, as is always the case, it looks like youtube is doing its own internal compressing thing.
You may see some odd things in a couple of the shots. In the first shot there is an area of softness around the upper left of the frame - unfortunately, the ND64 I was using got scratched really bad as I was setting up and so it affected the image there. More noticeable when I was shooting at a smaller aperture, but tried to stay around my lenses' widest apertures f/1.8 and f/2.8 mainly.
Exposure was relatively simple. Exposed to 100 IRE and then take a 1/3 step back via aperture - so basically an uneducated ETTR as I didn't know at what IRE clipping occurs yet. Shutter speed was set at 1/48th @ true 24fps. OM-Log400's base sensitivity is ISO 400 which is good in that it's very clean and you don't have to go so crazy with NDs. I noticed that I only ended up using an ND8 and ND64 - even in relatively bright situations, ND500 was a bit strong.
A couple of things that I noticed once I got the footage back to my computer at home. OM-Log400 seems to clip highlights at a value greater than 100IRE. Also, it seems to also retain values below zero (at least that's what my waveforms are telling me at the time of exposure as well as through my scopes on Davinci Resolve). I was able to pull back highlights from a lot higher on my waveform than I expected to so perhaps I could ETTR just a touch further.
I haven't done any scientific tests - others will I'm sure. I found that I was exposing around 1.3 - 2 stops over the exposure I was getting for Rec 709, and in post, I was able to pull down the file quite easily using the gain wheel. All that said, there is a very good improvement in dynamic range by having access to a log gamma profile.
As far as log gammas are concerned, it's not the flattest I've ever come across. In saying that, it's probably the most responsible 8-bit log gamma I've tinkered around with. Let me explain...
I've graded other camera's log gammas in 8-bit in the past and whilst some technically do have more dynamic range, the file's malleability is particularly compromised due to the combination of a wider dynamic range vs a limited number of values that could be assigned to address it i.e. 0-255. I found that I was able to stretch the Olympus file a bit more without the file breaking down so quickly compared to other 8-bit log gammas I've used in the past.
Definitely possible to get some really good footage from this camera. Now that I have a bit of a better idea of what I'm looking at exposure-wise, I'll probably post something more interesting going forward. This was a bit more of a stress test and you can definitely see that the log footage does capture more data than the default standard Rec 709 profile. So stay tuned in the near future.
Комментарии