We’re currently looking into if there’s any benefit of shooting RAW with the ARRI Alexa camera in an ACES workflow compared to using Prores 4444 XQ (in AWG/LogC).
Basically, do we get any extra color information by using RAW and “developing” directly into ACES instead of converting the Prores files into ACES colorspace?
Has anyone done any tests with this? I assume there’s no extra information, as it would be kind of strange if ARRI didn’t use the full potential of the sensor with it’s own colorspace?
- From what I remember you can’t shoot directly in ACES colorspace when using Prores.
ARRIRAW is uncompressed data and ProRes is compressed and this is what they say on their web site:
“For the absolute best in image quality, for the greatest flexibility in post, and for safest archiving, the 16-bit (linear) raw data stream from the sensor can be recorded as 12-bit (log) ARRIRAW files. Nothing is “baked” into an ARRIRAW image: Image processing steps like de-Bayer, white balance, sensitivity, up-sampling or down-sampling, which are irreversibly applied in-camera for compressed recording, HD-SDI outputs, and the viewfinder image, are not applied to ARRIRAW. All these parameters can be applied to the image in post.”
I haven´t done any tests in ACES but according to ARRI the difference between ARRIRAW and ProRes 4444XQ is minor.
“Recording clips in Apple ProRes 4444 XQ can preserve the full sensor quantization in logarithmic encoding, with the same range of colors available in ARRIAW. Images recorded in a 4:4:4 codec are almost indistinguishable from uncompressed HD or UHD material. This makes internal recording attractive even for feature film productions for the big screen.”
Seems they use the same primary colors…
as the white balance is already set in a ProRes file and not in the RAW file this might be the biggest visual difference that you can expect.
I had this once on ALEXA-MXF night car driving footage where it is anyway hard to say what this the “right” white balance between the warm street lamps and the cold LED lights from the (in this case white) car.
I compared the default white balance setting, convert to ACES EXR files and then try to balance the shot in Nuke vs. balance already in the ARRIRAW Converter and compare the two images.
There were slight differences in the high values in LED lamps front and back of the car. That was it.
Apart from the the slight compression that a ProRes based codec would apply, no I would not expect much extended colour range from an ACES- based workflow. However, this is not really the point. ACES apart from encoding the entire spectral-gamut provides you with a consistent and expected output in DI… It is non-trivial to derive your base look across many output devices and standards and ACES provides your the base framework to allow this to be seamlessly accessible and consistent… Tell a DI that you’re using ACES, and they’ll immediately know how to capably handle your footage… tell them anything else and everyone involved will have to reinvent the the wheel for your specific production.
So my advice is, please don’t reinvent the wheel unless you know why the wheel exists.