As an Amazon Associate I earn from qualifying purchases from amazon.com

Sony VENICE 2 Lab Check – Rolling Shutter, Dynamic Vary, and Latitude


The brand new Sony VENICE 2 with the 8.6K sensor was within the CineD headquarters for a fast go to – so we jumped on the possibility to run it by way of our lab take a look at procedures. Curious the way it fared? Then carry on studying!

It at all times feels particular to me when a “actual” cinema digicam arrives for me to run it by way of a lab take a look at – there’s at all times the hope {that a} new benchmark is about for our lab take a look at outcomes, because the final time this occurred was with the ARRI ALEXA Mini LF and Traditional (in case you missed it, you may examine it right here). So I used to be tremendous curious what Sony have up their sleeves with the brand new VENICE 2 cinema digicam.

Earlier than we begin, a giant “thanks” goes to my pricey CineD colleague Florian Milz who helped me with taking pictures this lab take a look at!

Sony VENICE 2 – Rolling Shutter

Effectively, first massive constructive shock: in full-frame 8.2K mode our normal process to check the rolling shutter reaches its limits as a result of the rolling shutter is so low: we use a 300Hz strobe mild which generates patterns of pairs of black and white traces. Every pair represents 3.3ms. With the Sony VENICE 2, the readout velocity of the sensor is so quick that we weren’t in a position to confidently establish one full pair.

Therefore, for the primary time, we’re supplying you with an approximate outcome: the readout velocity is about or barely beneath 3ms (decrease is best)! That is the fasted readout velocity for a full-frame digicam that we’ve measured up to now! Wow – right here you go, a brand new benchmark!

Rolling Shutter for the Sony VENICE 2 in 8.2K full-frame mode: barely beneath 3ms.

Curious to seek out out extra about how we take a look at the dynamic vary of cameras? Take a look right here.

Let’s begin first with the brand new inner X-OCN format. Little or no technical data may be discovered on-line about it, however in DaVinci Resolve 17.4.6 (construct 4) all typical “RAW” controls can be found, exhibiting the file as “Sony RAW”. Therefore, I developed the X-OCN XT information with SGamut3.Cine / SLog3 settings in Full Res – Sony:

The Sony VENICE 2 encompasses a twin native ISO sensor, the 2 base sensitivities are ISO800 and ISO3200.

At ISO 800 utilizing 8.6K X-OCN XT Open Gate mode, we get the next waveform:

XYLA 21 Waveform plot of the Sony VENICE 2 at ISO800. Picture credit score: CineD

A stable 13 stops are seen above the noise ground, with a visual 14th cease contained in the noise ground as effectively. The information are a bit noisy, like many RAW codecs these days.

IMATEST calculates 11.7 stops at a sign to noise ratio (SNR) of two and 12.9 stops at SNR = 1:

IMATEST outcomes for 8.6K X-OCN XT at ISO 800. Picture credit score: CineD

Above the blue “12.9” line within the center graph above, one other 2 if not 3 stops are recognized contained in the noise ground. Let’s see if these are usable in a while within the latitude part.

At ISO3200 (the second native ISO setting) the outcomes are nearly the identical – as they need to be for an actual twin native ISO sensor! IMATEST (see beneath) calculates 11.1 stops at SNR = 2 and 12.9 stops at SNR = 1.

XYLA 21 Waveform plot of the Sony VENICE 2 at ISO3200. Picture credit score: CineD
IMATEST outcomes for 8.6K X-OCN XT at ISO 3200. Picture credit score: CineD

The VENICE 2 additionally options internally downsampled 4K ProRes HQ recording – a really engaging various for quicker turnarounds in a well-established ProRes pipeline. And right here, the outcomes out of the field are even higher, as a result of a correct downscaling from an 8.2K portion of the sensor, however some inner noise discount appears to occur as effectively. Extra on that additional beneath.

Utilizing SGamut3.Cine and Slog 3, the next waveform plot is obtained:

Xyla 21 Waveform plot for ProRes HQ 4K. Picture credit score: CineD

Fairly surprisingly, the code worth distribution modifications. SLog3 normally clips across the code worth of 896 (or luma worth of 87). Within the waveform plot above the sign clips round 960 code worth (or 94 luma worth). Additionally, the noise ground sits at a decrease code worth. We’re reaching out to Sony with this query and can replace the article accordingly at a later stage.

IMATEST calculates 13.2 stops at SNR = 2 and 14.2 stops at SNR = 1:

IMATEST outcomes for 4K ProRes HQ at ISO800. Picture credit score: CineD

These are very good outcomes, the second-highest values ever measured within the lab! The kings of dynamic vary are nonetheless nonetheless the ARRI ALEXA cameras, with the ALEXA Mini LF coming in at 13.5 stops at SNR = 2 and 14.7 stops at SNR = 1 (Lab take a look at right here)

At ISO3200 in 4K ProRes HQ mode, 13 stops at SNR = 2 and 14.3 stops at SNR = 1 are measured. By the way in which, if I downsample the ISO800 8.6K X-OCN XT file to 4K after which run it by way of IMATEST, 12.4 stops at SNR = 2 and 13.6 stops at SNR = 1 are obtained – decrease values. Therefore, confirms my suspicion that in ProRes HQ extra noise discount is utilized internally to the information.

One remaining observe to ProRes HQ – I don’t know if it’s a DaVinci Resolve or Sony concern, however there are not any Metadata seen in Resolve. For X-OCN although all metadata are proven accurately.

Sony VENICE 2 – Latitude take a look at

Latitude is the aptitude of a digicam to retain particulars and colours when over-or underexposed and pushed again to base publicity. A while in the past, we’ve chosen an arbitrary worth of round 60% luma worth (within the waveform) for our topics’ brow in our normal studio scene. This CineD base publicity ought to assist our readers to get a reference level for all of the cameras examined, no matter how they distribute the code values and which LOG mode is used.

Commonplace CineD studio scene. Picture credit score: CineD

For the Sony VENICE 2, we put the bottom publicity a tad beneath 60% within the luma waveform to permit overexposure of as much as 4 stops. Once more for X-OCN XT, we developed the information utilizing SGamut3.Cine and Slog3 within the DaVinci Resolve “Digital camera RAW” tab. Additionally, we used the official Sony SLOG3-S-GAMUT3.Cine TO S709 V200.CUBE LUT to deliver the picture into the REC.709 house (may be downloaded for the VENICE 2 right here). Out of the field, the picture seems very good – I added a tiny little bit of saturation, nonetheless.

Taking a look at 4 stops of overexposure, and bringing again the picture to base publicity we get the next picture:

VENICE 2 4 stops overexposed and introduced again to base. Picture credit score: CineD
RGB Waveform for the ungraded 4 stops overshot: the crimson channel on Johnnie’s brow is near clipping. Picture credit score: CineD

The crimson channel (on the ungraded shot) on Johnnie’s brow is near clipping, however nonetheless intact – correctly.

From right here on we underexpose the photographs by closing the iris of the lens in one-stop increments till T8.0, after which we’re additional halving the shutter angle successively.

At 3 stops beneath base publicity, issues are getting fascinating, as already fairly some noise reveals up within the picture:

3 stops below, pushed again to base. Picture credit score: CineD

Nothing dramatic, so let’s transfer on to 4 stops beneath base, introduced again:

4 stops beneath base, pushed again. Picture credit score: CineD

Now, heavy noise is affecting the picture. Additionally, nasty speckles of chroma noise (inexperienced and crimson) are all around the picture. Utilizing heavy spatial and temporal noise discount in DaVinci Resolve, we will save the picture however on the expense of inexperienced and pink chroma noise – see beneath:

4 stops beneath base, pushed again and utilizing noise discount. Picture credit score: CineD
DaVinci Resolve 17.4.6 noise discount settings for the 4 stops beneath base & pushed again picture. Picture credit score: CineD

The picture turns greenish, and we’re getting bigger areas of greenish and pinkish blotches hovering across the picture, which I discover quite distracting in a shifting picture. However I’d say, nonetheless useable.

Let’s take a look at 5 stops below base, pushed again:

5 stops beneath base, pushed again. Picture credit score: CineD

The picture reveals a really sturdy inexperienced solid, and noise is getting atrocious. Sport over? Certainly, sport over. Take a look at my try to make use of noise discount beneath:

5 stops beneath base, pushed again and utilizing noise discount. Picture credit score: CineD
Noise discount settings for five stops below, pushed again utilizing X-OCN XT. Picture credit score: CineD

As quickly as I improve the chroma noise slider within the Spatial Threshold part the picture turns brown. Utilizing the Luma slider simply turns the picture right into a plasticky, delicate mess. Therefore I fully prevented that.

Sport over. Therefore, the Sony VENICE 2 8.6K in X-OCN XT reveals 8 stops of latitude. For comparability, the ARRI ALEXA Mini LF has 2 stops extra latitude (see the take a look at right here), yet one more within the highlights, and yet one more within the shadows (5 over and 5 below).

On a aspect observe, I attempted numerous methods in Resolve to see if I might squeeze out higher outcomes:

  • utilizing the publicity slider (therefore growing the ISO) within the Digital camera Uncooked tab: right here, outcomes have been even worse, with a a lot stronger colour solid. So I set it again to ISO800.
  • then, utilizing the Elevate / Achieve and Distinction sliders within the Digital camera Uncooked tab: with these controls, I used to be not in a position to accurately reconstruct the picture (by evaluating it to the bottom publicity waveform)
  • Lastly, I did what I at all times do for every digicam: utilizing the “Elevate, Gamma, Achieve” Primaries within the Coloration Wheels tab in a node earlier than the ultimate LUT node (by the way in which, noise discount is utilized on the primary node at all times).

Now I used to be curious if ProRes HQ would behave in a greater approach. For 4K ProRes HQ, I used to be getting this picture at 4 stops below:

4K ProRes HQ at 4 stops beneath base, pushed again. Picture credit score: CineD
4K ProRes HQ at 5 stops beneath base, pushed again. Picture credit score: CineD

Slightly related, however much less noisy. Therefore, I attempted once more noise discount on the 5 stops beneath base picture:

4K ProRes HQ at 5 stops beneath base, pushed again and making use of noise discount. Picture credit score: CineD
Noise discount settings for five stops below, pushed again utilizing 4K ProRes HQ. Picture credit score: CineD

The static picture seems fairly okay, however sadly within the shifting picture there are nonetheless these blotches of pink and inexperienced chroma noise however to a much less apparent extent – I do discover them very distracting however relying on the person style you might argue that we see 9 stops of latitude.

So all in all, in 4K ProRes HQ you do get higher outcomes – most likely because of the inner costumed sensor-specific fine-tuned noise discount.

I’d actually ask you to match the Sony VENICE 2 latitude outcomes to the ARRI ALEXA Mini LF latitude outcomes right here. Not solely does the ALEXA Mini LF have a full cease and extra latitude within the highlights, however additionally it is eye-opening to see how good the underexposed Mini LF photos nonetheless look in comparison with the Sony VENICE 2, even with out noise discount. Sony has these giant nasty patches of chroma noise within the shadows, whereas the Mini LF has a really tremendous grain.

Abstract

The Sony VENICE 2 units a brand new benchmark for rolling shutter: lower than 3ms for a full-frame 8.6K sensor – Wow! Within the dynamic vary division, issues get a bit extra difficult. In Sony RAW, X-OCN XT the photographs are fairly noisy – resulting in fairly common dynamic vary outcomes of 11.7 stops at SNR = 2.

In downsampled inner 4K ProRes HQ, we get the third-best outcomes ever measured, 13.2 stops at SNR = 2. Clearly, Sony is making use of some personalized inner noise discount right here.

Within the latitude part, between 8 (Sony RAW – X-OCN) and 9 stops (4K ProRes HQ) are obtained. Sadly, excavating data from the noise ground isn’t as simple as a result of nasty chroma noise is destructing the picture. Even with superior noise processing in DaVinci Resolve the bounds within the latitude part are reached fairly early. 8 stops (for Sony RAW) is definitely alongside the traces of full-frame client cameras just like the Panasonic SH1, or Sony A1 that even have 8 stops of latitude in our normal CineD studio scene. The ARRI ALEXA Mini LF stays the chief of the pack with 10 stops of publicity latitude, regardless of if filmed in ARRIRAW or ProRes.

What do you consider this Sony VENICE 2 lab take a look at? Do you already use this new digicam? If sure, please share your expertise within the feedback part beneath!



We will be happy to hear your thoughts

Leave a reply

realdealdeals
Logo
Enable registration in settings - general
Compare items
  • Total (0)
Compare
0
Shopping cart