Sony FX6 Lab Take a look at – Exterior ProRes RAW vs Inner XAVC-Intra

Since most newer full-frame cameras can record ProRes RAW via external recordings, we can take a look at how the pristine "raw images" are compared with the internally processed signals. So I was very curious to see how the new Sony FX6 is constructed and how it compares to its smaller brother, the Sony a7S III – after all, they seem to have the same sensor, but the Sony FX6 is part of the “Cinema Line” ”. Therefore the signal processing can be different. Curious about the results? Then read on …

If you haven't read our Sony FX6 review, please take a look here. You can find our Sony a7S III laboratory test here. As a quick reminder, the Sony a7S III was hampered in our lab test by the heavy noise reduction that resulted in ugly, large patches of chroma noise. Will the Sony FX6 be any different? Spoiler alert – yes!

But one by one. Let's look at the results of the rolling shutter first.

Short side note: 12-bit 4K ProRes RAW files on a 4K timeline in Premiere Pro are no problem for our NVIDIA GeForce RTX3090 graphics card – they run buttery smooth with a load of around 8% and 5.8 GB of GPU memory. This card is on loan from NVIDIA – thank you!

Sony FX6 – roller shutter

With our 300 Hz flash, it was easy to see that both the Sony a7S III and Sony FX6 produce the same results. The Sony FX6 shows an excellent rolling shutter result of 8.7 ms in full screen, UHD recording from 24 to 100 frames per second (yes, even 100 fps are without cropping – a slight cropping is only used at 120 fps).

Sony FX6 Rolling Shutter in UHD 25fps full screen mode. Image credit: CineD Sony FX6 Rolling Shutter in UHD 120fps full screen mode. Image credit: CineD

Sony FX6 Dynamic Range Shot at ISO 800

ProRes RAW HQ:

We used the Atomos Shogun 7 for the ProRes RAW recording. The files were developed in the latest Adobe Premiere Pro version 15.2 with SGamut3.Cine / SLOG3. Please stop by here to see how we test dynamic range. The internal noise reduction was set to "OFF" for all recordings. Oddly enough, the ProRes RAW files recorded on the Shogun 7 have a resolution of 4128 × 2192 (as opposed to the 4K DCI 4096 × 2160 resolution claimed by the Atomos website).

So let's take a look at the ProRes RAW waveform graph first. About 12 stops can be seen above the noise floor, and a 13th can be seen within the noise floor:

Waveform diagram of the Sony FX6 in ProRes RAW at ISO800 UHD 25fps. Image credit: CineD

IMATEST calculated 11.4 stops at a signal-to-noise ratio of 2 (SNR = 2) and 12.6 stops at SNR = 1.

The Sony FX6 shows a very nice organic noise floor, and IMATEST identifies about 3 more stops within the noise floor (above the blue SNR = 1 line in the middle graph below):

MOST IMATEST result for the Sony FX6 in ProRes RAW at ISO800 full screen UHD 25p. Image credit: CineD

If we compare that to the Sony a7S III (at ISO 640), which again uses external ProRes RAW footage (this time via HDMI on the Atomos Ninja V), we get a pretty similar result of 11.6 stops at SNR = 2 and 12.8 stops at SNR = 1 (more details on this in a future post). As a side note, I found it interesting that the Sony a7S III ProRes RAW images show a resolution of 4264 × 2408 (as correctly mentioned on the Atomos website here) compared to the 4128 × 2192 ProRes RAW on the FX6 with the Shogun 7 -. So is it really the exact same sensor?

Internal XAVC-Intra 4: 2: 2 SGamut3.Cine / SLOG3::

At ISO 800, IMATEST calculates 11.7 stops for SNR = 2 and 12.8 stops for SNR = 1. Compared to the ProRes RAW recording, it becomes clear that there is little image processing / noise filtering for the internal recording. It is a very good thing!

IMATEST DR result for the Sony FX6 at ISO800 with XAVC-I SG3.C / SLOG3 recording. Image credit: CineD

The Sony a7S III dynamic range in IMATEST shows 12.4 stops at SNR = 2 and 13.6 stops at SNR = 1 with the internal XAVC S-Intra 4: 2: 2 codec. Here you can see the rather significant internal noise processing of the a7S III (compared to external ProRes RAW), which pushes the values ​​for SNR = 2 and 1 up – which unfortunately cannot be switched off. Personally, I prefer the internal recorded images of the Sony FX6 with noise reduction “off”. It has that nice, organic feel to it.

Sony FX6 Dynamic Range Admission to "high ISO value, 12800

ProRes RAW::

The RAW waveform diagram of the Sony FX6 ProRes becomes much louder, yet around 12 f-stops can be identified above the noise floor:

Waveform diagram of the Sony FX6 at ISO 12800 in ProRes RAW at UHD 25fps. Image credit: CineD

IMATEST calculates 10.5 stops at SNR = 2 and 12.3 stops at SNR = 1. In a similar way as mentioned above for ISO 800, about 3 additional stops are identified within the noise floor, which can be worked out with advanced noise processing techniques in the mail .

IMATEST result for the Sony FX6 at ISO12800 full screen UHD 25p. Image credit: Cined

For comparison: The Sony a7S III ProRes RAW image recorded with ISO 12800 shows 10.2 stops at SNR = 2 and 12.1 stops at SNR = 1. Here, the FX6 now takes the leading position with a small margin.

Internal XAVC Intra 4: 2: 2 recording:

For the internal recording at ISO 12800 with SGamut3.Cine / SLOG3, IMATEST calculates 11.2 stops at SNR = 2 and 12.6 stops at SNR = 1. In total, about 0.5 stops less than with ISO 800 and compared to ProRes RAW something internal At this ISO there seems to be a noise filtering that cannot be switched off. As can be seen later in the latitude test, this internal noise filtering has an advantage over post-production techniques, probably because it is tailored to the specific sensor.

IMATEST DR result for the Sony FX6 at ISO12800 with XAVC-I SG3.C / SLOG3 recording. Image credit: CineD

For comparison, the internally recorded XAVC S-I 422 Intra from Sony a7S III at ISO 12800 shows 12.1 stops at SNR = 2 and 13.3 stops at SNR = 1. This is also a result of the strong noise reduction that Sony does not a7S III can be switched off – especially at higher ISO values. This strong noise reduction results in larger patches of chroma noise that make shadow restoration difficult.

Dynamic Range conclusion:

All in all, the ProRes RAW results are good for a full frame sensor and are on par with most other cameras that use 12-bit consumer A / D sensors. The differences we see from the internally captured images are related to how the image is set up in terms of noise processing (and other factors, of course).

And of course I am in favor of the way the Sony FX6 processes the internally captured images compared to the Sony a7S III.

When it comes to internal signal processing, the Panasonic S1H (and S1 / S5) with internal codecs at SNR = 2 show about 1 stop more dynamic range than the FX6 – see our article here – with a very pleasant way to filter the noise – While in ProRes RAW the S1H shows very similar results as the FX6 (this will be the subject of a future lab review). So it's all about internal signal processing!

And while it's a little unfair, refreshing your memory with the current ARRI ALEXA Mini LF that we recently tested in our lab shows 13.5 stops at SNR = 2 (in ARRIRAW), so 2 stops more.

Sony FX6 Lab test – latitude test

As mentioned in previous lab tests, the margin is a camera's ability to retain colors and detail when overexposed or underexposed and (retrospectively) brought back to base exposure.

As usual, we have our (arbitrarily) selected CineD base exposure of the studio scene with a maximum of 60% luma value of our test subject's forehead, in this case my dear colleague Johnnie.

As already mentioned, all ProRes RAW recordings in Premiere Pro were developed using SGamut3.Cine / SLOG3 as the gamma curve and using an additional SG3.C / SLOG3 for REC709A LUT. Premiere Pro only has the option to choose exposure and gamma, nothing more (no white balance, no hue, etc.). However, the exposure compensation works fine.

These recordings were then exported with ProRes 4444XQ (12 bit) and brought into DaVinci Resolve 17.1 for further analysis.

Sony FX6 base exposure of our standard studio scene with ProRes RAW. Image credit: CineD

From this basic exposure we open the iris of the lens until the forehead begins to pinch, which happens 4 stops above our basic exposure scene:

4 stops above the basic exposure, ProRes RAW, and pushed back. Image credit: CineDRGB waveform of the 4 stops over the scene – the red channel cut off Johnnie's forehead. Image credit: CineD

As can be seen, the red channel cuts on Johnnie's forehead (see also the RGB waveform above) – so we went too far and cannot count this result. Unfortunately, ProRes RAW does not have a highlight restoration option that could be used to reconstruct possibly clipped color channels (e.g. Blackmagic RAW offers this option). Otherwise, this recording may have been saved.

The same also applies to internal XAVC-Intra 4: 2: 2:

4 stops above base exposure, XAVC Intra, and pushed back. Image credit: CineD

Now we close the aperture of the lens one after the other and from F8 onwards we reduce the shutter angle to 180 °, 90 °, 45 °, 22.5 °.

At 3 f-stops below our base exposure, pushed back, noise appears in the image:

3 stops below the base exposure, ProRes RAW, and pushed back. Image credit: CineD3 stops below base exposure, XAVC Intra, and pushed back. Image credit: CineD

Both the ProRes RAW and internally captured XAVC intra samples look very similar at this point (but the colors of the internally captured images are muted a bit) but things change when we go down 4 stops, pushed back:

4 stops below the base exposure, ProRes RAW, and pushed back. Image credit: CineD4 stops below the base exposure, XAVC-I, and pushed back. Image credit: CineD

Noise affects the images a lot, especially in ProRes RAW – where horizontal stripes appear in the images. Noise reduction helps in both cases, and both images can be cleaned up:

4 stops below the basic exposure, ProRes RAW, noise reduction and push back. Image credit: CineD4 stops below the basic exposure, XAVC-I, noise reduction and push back. Image credit: CineD

What can be noticed so far is that the ProRes RAW files are a bit louder than the internal XAVC-I files – and I found it difficult to reduce the noise in DaVinci Resolve without spoiling the image too much.

At 5 apertures underexposed, this difference becomes very evident – to the extent that I haven't found a way to use noise reduction to bring the ProRes RAW file back without slowing it down. But let's take a look around first:

5 stops below the base exposure, ProRes RAW, and pushed back. Image credit: CineD5 stops below base exposure, XAVC-I, and pushed back. Image credit: CineD

The strong noise in the 5 stops under the ProRes RAW file damages the image so badly that it cannot be saved. While the 5 stops can still be saved under the XAVC-I file.

5 stops below the base exposure, ProRes RAW, noise reduction and push back. Image credit: CineD DaVinci solve the noise reduction settings for the 5 stops under the ProRes RAW file of the Sony FX6. Image credit: CineD

As can be seen, having a lot of noise cancellation doesn't help in DaVinci Resolve – the point of no return. Noise completely corrupts the image, and horizontal stripes also become clearer.

5 stops below the basic exposure, XAVC-I, noise reduction and push back. Image credit: CineD DaVinci loosen the noise reduction settings for the 5 stops under the Sony FX6's pushed back XAVC-I file. Image credit: CineD

The internal XAVC-I 5 stops under the file can be saved by noise reduction (much less of it in terms of the temporal and spatial threshold). But it's also about to break up, and horizontal stripes also appear across the image.

What needs to be mentioned, however, is that the 12-bit ProRes RAW files retain colors much better despite the noise.

Hence we can conclude that:

• From the results, we can infer about 8 stops in the exposure margin
• The externally recorded ProRes RAW HQ files show more noise than the internally recorded XAVC-I files
• Strange but true, the ProRes RAW HQ files dissolve on the 5 apertures that are underexposed and pushed back. Although it's 12-bit, moving up shadows with a 12-bit codec should theoretically produce better results. Not because of the noise
• The internally recorded XAVC-I 4: 2: 2 10-bit files with 5 underexposure stops that were pushed back to the base can still be saved using noise reduction. You are about to break but are still okay

Summary

All in all, the Sony FX6 is a solid performance in our laboratory test – excellent roller shutter values ​​and good dynamic ranges are offered. The exposure latitude is about 8 stops (3 over, 5 under) – but more for the internal XAVC-I codec, not quite for ProRes RAW.

I'm sorry to say, but I don't see the point of ProRes RAW HQ from these lab results. It is not supported by DaVinci Resolve. You only have rudimentary controls in Premiere Pro (exposure and gamma curve, no white balance, no hue, etc.) and the files are louder – even DaVinci Resolve's noise reduction can't save them to a certain extent. Additionally, there are no highlight recovery options available – one of the benefits that other RAW codecs (such as BRAW) have.

But I like the way the FX6's internally captured image is set up – very organic and minimal (but effective) noise filtering, very “cinema line”. Completely different to the Sony a7s III – which looks overworked to my eyes, is too noise-filtered and thus affects the skin tones.

Have you worked with the Sony FX6? What is your experience Please let us know your thoughts in the comments section below.