Sony PXW-FX9, Maxi FS-7 or Mini VENICE?

Sony PXW-FX9

Every new camera presents novel, distinct features that make it stand out from competitors. By defining what tasks they are best at we get to determine their target audience. Let us see what nice surprises are awaiting us this time.

Lab test perfomed by Luis Pavía


The camera we are bringing today to our lab has been for some months now on the market, and we were considering to wait a bit more so as to be able to test it with the new features that will be included by means of a firmware update, as a good part of its novelties will only be available as from version 2.0 of said firmware, which is scheduled for release –free of charge- during the coming autumn.

But we have finally decided not to delay matters any longer so we can share with you the features that make this camera something as different as the new S-Cinetone colour technology that comes standard does. A legacy from the renowned VENICE that really places this device on a different level as compared to its natural competitors.

But, please let me proceed in a certain order. Very similar in appearance, size and weight to a FS7 and with no intention to replace the latter, this camera features a number of significant differences such as:

– The full-frame backlit Exmor-R sensor, with a 6K native resolution. However, images recorded will never exceed 4K in any of the available formats.
– The hybrid autofocus system, which is extremely quick and accurate, with expectations for eye-tracking focus as from firmware v. 2.0 onwards.
– The above-mentioned S-Cinetone colour technology, that really impressed us and which deserves some paragraphs below.
– The first full-frame format variable neutral-density filter, featuring continuous variation.
– Its dynamic range, slightly above 15 f-stops in latitude.
– Dual ISO, with bases in 800 and 4000, capable of providing excellent results within a very wide range of lighting conditions.
– Built-in WiFi and 12G SDI, together with timecode and Genlock input-output connectors within the camera’s own body, although we will still need its optional XDCA-FX9 adaptor in order to record internally in RAW format, this time up to 16 bits.
– And other minor features that we will examine in due course throughout this lab feature.


So, let’s get going.

Fitting a 6K sensor in a camera that does not offer –at least for the time being or as foreseen in the short term- any possibility for recording in resolutions higher than 4K DCI may seem paradoxical. We do not know if this is an option that could be offered in the future, but certainly there is no information that would enable us to tell. Does it make sense then? Sure it does. The sensor features a native resolution of 6008 x 3168 pixels in academic format 17:9, for a usable area of 35.7 x 18.8 mm.

Having all that extra resolution available provides several significant improvements. On the one hand, space resolution is greatly increased, thus minimizing the restrictions inherent to the Bayer sensor. As not all pixels have all the colours, the higher the number of pixels, the more accurate the information gathered at capture both in terms of resolution and in colour and noise reduction.
And, on the other hand, this increased sharpness in capture immediately translates into data that are more accurate for the autofocus hybrid system. Therefore, by combining the contrast detection and phase detection technologies, and thanks to that larger collection of more accurate data, an autofocus system that is significantly faster and more accurate than what we had known until now is achieved. And the difference is clear and pleasantly noticeable.

Add to this the well-known face tracking feature, which is not limited to recognition of human faces but goes a step further and is now capable of distinguishing and identifying specific individuals, and we will be able to sort out a lot of situations much more easily than in the past. The camera is capable of recognizing and following our main character, identifying the same person again after they have left the framing. There are options available for managing the behaviour of autofocus for objects in general, only human faces in general, or a specific face that has been memorized. This facilitates the tracking of individuals in action and therefore versatility of use for highly varying environments is increased.

Save for situations in which we rely on heavily marked action positions and focus, we are convinced that in quite a few instances the camera’s automation will provide better results than many of us would attain when dividing our attention between focus, exposure, framing, etc.



Sony PXW-FX9



Further analyzing the autofocus system and pending release of new firmware version 2.0, for which the eye tracking feature is announced, having the certainty that the autofocus will be able to follow-up this detail will facilitate the shooting of much closer takes with a precision level that we are eager to test. Not only the face will be on focus, but also the eye selected. Precision will be increased even in awkward takes in a much more flexible way.

Why are we so blunt in stating the significant difference in speed and precision of the hybrid autofocus? Simply because our test unit came equipped with a FE PZ 28-135 F4 G OSS optic, identical to the one provided with the kit for the FS-7, which we had the opportunity to try out in due course and of which we will not reveal anything new.

Placed in front of the sensor is the variable neutral-density filter in a ND2-ND7 range (1/4 to 1/128), with two important novelties. First, it is the first full-frame format variable neutral-density filter. Secondly, when removed a transparent glass is placed instead in order to prevent any kind of back-focus effect.

The filter can be configured for response on a continuous mode or with three different levels of intensity, in the traditional style. Each of them can be set at any value by means of the menu.
We have already mentioned in other lab features how much we like this solution. Being able to adjust exposure without touching ISO, iris or speed and, therefore, without altering the visual narrative –as the relevant noise, field depth or feeling of movement levels remain unchanged- seemed an extraordinary feat to us from the very first time in which we had the opportunity to try this in a FS5.

The use of a filter does not alter colour fidelity at all and it does not add any dyes; and no differences in filtering uniformity are noticeable. Furthermore, the variation is progressive enough, so no visible effects are shown. This enables using the camera with the assurance that images will not be altered and no delicate post-production processes will be subsequently required.

In front of the filter we have the E-Lock mount. Introduced in the FS7-MII, it requires some practice, much in the PL style. We will need both hands in order to mount and detach the optics safely, but fixation –conceptually similar to the above-mentioned PL mounts- is much more robust and reliable than the traditional bayonet mounts. It has been designed for mounting and detaching optics without twisting them, which makes handling of follow-focus type accessories, remote controls or mate boxes a quicker and easier operation. As the versatility of Sony’s E-mount is maintained, we can make the most of the optics we may already have.

In this regard it is worth highlighting that, at least until the time of drafting these lines, it would seem that third-party adaptors do not provide the expected result as some functionalities are missing or somewhat limited. This is nothing to be really worried about, as in these times in which even optics and adaptors come with internal firmware, we are convinced that the relevant updates will put everything back in place.

This is due to all the exchange of information that takes place between the camera’s body and the optics as metadata even store details on stabilization that may be used for stabilizing takes during post-production. More on this further below.

On the body we find plenty of controls, most of them located on the left-hand side and handgrip. Everything is quite familiar for those used to this brand. In addition to the ND filter selector and its control knob there is a new rotating push button that is very accessible and useful for various purposes, such as iris control or access to and operation of the direct menu. Also available is a direct control for the four audio channels and 10 customizable buttons distributed through body and handgrip.

The three record buttons located on the body itself, handgrip and in the upper handle, also facilitate operating the camera in quite different shooting situations.

In view of the fact that possible combinations of resolutions, formats and frame rates will be expanded through new firmware, we refer to the tables already published for the sake of avoiding unnecessary duplication.

Audio channels 1 and 2 get in through the already classical XLR connections, with the option of selecting for each channel whether the signal of these connectors or that of the built-in internal microphone is to be recorded. This function comes handy for attaining a synchronized audio of reference whenever recording is made on external devices. Access to audio channels 3 and 4 is made through the smart pad, but with an interesting novelty. Up to now wireless audio from belt packs would be transmitted between issuer and receiver by means of digital technology. This signal had to be converted to analogue output in order to get it in through the camera’s XLR connectors and then converted back to digital. Now, by means of the UWP series wireless microphones and the new SMAD-P5 adaptor, the audio –digitalized in the transmitter- enters the camera directly in digital format, thus making any intermediate digital-analogue-digital conversions unnecessary.

The viewfinder has also undergone improvements on resolution, contrast and colour, offering now 1280 x 720 pixels. It has the tenth customizable function button. Support bars are now shaped so as to prevent it from turning by accident as it would be the case with some of the first few models of the FS series. The next firmware update is also expected to come with touch screen features that, for the time being, we have not been able to enjoy. But there is an aspect that we think it can be improved regarding the eyepiece. Although better than in past models, the fact that the push buttons used for opening it or removing it in full are so close together and soft, makes us fear that the eyepiece may easily come off by accident. We must be careful with this.

As for the handgrip, very similar to those known from the FS5 and FS7 models, here we find customizable buttons 4, 5 and 6, as well as the control knob which is also customizable and very useful, for instance, for adjusting the variable ND filter. The connector is no longer a 2.5mm mini-jack as in previous models, but a mini USB 3.0. Also changed is the design of the mini-joystick type button used for scrolling through the menus. We do not know yet what kind we like best, the former or the new one. But we think this is a matter of getting used to the new feel, which is certainly very different.

We love the handgrip notion in the adjustable arm. We have liked it ever since it first appeared in the first FS7 model. And handling of the equipment once fine-tuned is excellent. But everything that makes this camera so good in operation turns into a drawback when leaving it sitting somewhere if a tripod is not used. We know that there are third-party solutions that make the turning of the arm something much quicker and safer than with the current design. And we strongly recommend them if the camera is going to be used in ENG-style environments.

In addition to the usual SDI, HDMI, TC and Genlock connections –already mentioned- that do not require an optional external adaptor, it was surprising to see that the classical USB port that allows for direct connection of the camera with a computer to transfer content is apparently missing. However, thanks to the varied and illustrative information provided by Álvaro Ortiz – a Sony’s product specialist- we were able to find out that this port is now the very USB 3.0 located on the handgrip.

Completing our review of the camera’s physical side, there is little else to add that is not already known based on its likeness with the model’s predecessors. 2 sockets for XQD cards and a very deep hole for BP-U batteries. Given that power consumption is higher in this camera than in the FS7, bigger, more powerful batteries have been developed.

Moving on now with the software review, it should be mentioned that explaining the menu options in depth would require a whole feature article, so we will only highlight the things that we have found most interesting.

First of all, syncing of metadata between the optics and the camera’s body, making it easier –just as an example- to be able to subsequently use stabilization data in some software applications such as Catalyst.

A feature that we like a lot, but one that often goes unnoticed, is ‘pre-record’, or the possibility of ‘recording the past’. Depending on the format and codec selected a time slot of the past –lasting between 4 and 28 seconds- can be retrieved, provided, of course, we had been pointing to the right direction. When this function is enabled, a cache memory area is permanently kept on a loop whose content is transferred to the memory card when the record button is pressed. The camera keeps recording until the button is pressed again, as in every regular recording.

The truth is that with regards to recording formats and frame rates this first version does have some limitations that will be removed when updated to the new 2.0 firmware. For example, the camera only records in HD and UHD, with some limitations as for available combinations between video format (HD or UHD), sensor’s (FF or S35) scanned area, codec (XAVC-I or XAVC-L, in addition to MPEG2) and frame rate (23.98, 25, 29.97, 50 ó 59.94), being at present 120 fps for HD and 60 fps for UHD the highest frame rates possible. As for data rates, as usual, based on the parameters selected, content will be generated with rates ranging between 25 and 600 Mbps. With the expected firmware update we will have at our disposal 4K DCI at 17:9 and frame rates up to 180 in HD as the most significant enhancements.

What has really struck us is the unavailability of recording at 24 fps. And this is no technical limitation, we are sure, for the camera is capable of recording both at 23.98 and 25 fps. We would rather believe that this is one of the points on which the purpose of this camera lies as it is targeting more ‘broadcast’ or ‘video’ environments that a pure ‘cinema’ environment. For the latter task we have the VENICE, with all its excellent qualities.

This is probably a marketing-led decision, rather than one based on engineering, that has both advocates and detractors. Sure this could be the subject of another feature article.

As for both base ISOs, let us see what they really mean. What could be the point of changing ISO base when ISO can be subsequently adjusted to sort out various lighting situations? To make things simple, shooting a take at ISO 1600 from an ISO 800 base and shooting the same take also at ISO 1600 but from an ISO 4000 base will certainly not yield identical results. In each case and always setting at the same signal level the average grey of reference, different image textures –regarding both grain and noise- will be obtained.

As we have done so many other times, we kindly invite you to evaluate by yourselves the contents recorded to this purpose as not all our eyes share the same demands, and opinions without the proof of one’s own experience in an aspect as critical as this may lead to error too easily.

When it comes to recording, we have, of course, S-Log2 and S-Log3 curves, in addition to all other usual standards. But the great news really is the S-Cinetone curve, which certainly deserves a word.

When Sony’s landmark product, the VENICE, was developed not so long ago, great care was taken to obtain a particularly accurate reproduction of colour, in addition to a smooth treatment of skin hues. Work was made in close cooperation with renowned photography directors, thus attaining the well-known result that has turned the VENICE into Sony’s product of reference for top-notch productions.

But in order to achieve such excellent results, proper processing of images recorded with log curves is required. And this always involves a critical colour grading process during the post-production stage, something that requires great deal of resources and has significant impact in delivery times of the finished product. And this quality is adequate, naturally, in said top-notch productions but less advisable when both deadlines and budgets grow tighter.

Working with the FX9 in S-Cinetone format allows for close resemblance of the results obtained in video aspect relating to hue and colour with the colour-graded cinema style, with no need of further subsequent processing. But let me be clear: I mean ‘close resemblance’ within the video environment, not ‘equal’. Making good use of the S709 LUTs (not to be confused with the R709 colour space) to get the result of a RAW recording under S-Log3 closer to preview monitors, this reference has been taken as starting point for developing the S-Cinetone, thus being capable of creating content with an excellent final look directly in the camera with no need of resorting to a painstaking, tricky and expensive post-production process.

By gradually decreasing contrast in highlights exceeding 70% of video level up to the upper limit, these areas show a smooth look while maintaining the degree of detail. However, by increasing it in the shades while also increasing saturation, a general aspect that is easy to adjust is achieved as well by just controlling exposure in the camera. Black level is decreased to 1.5% as compared to 3% in traditional video. As skin hue is the most important and critical issue, simply by changing exposure levels, its look can be modified quite easily.

One of the distinct features which take this camera to a different level is the availability of this kind of system, which allows controlling the look of skin by just adjusting the exposure level in the camera with no need of further processing while offering an exceptional result in an easy and immediate fashion.
The dynamic range offered by this curve reaches 460%, which although lower than what a S-Log3 curve can achieve, still provides an excellent balance between dynamic range and noise level for video images. The purpose is not to leave aside the results obtained through the full process, but to achieve high-quality output by means of an extremely efficient easy-to-control workflow. Furthermore, it is suitable for use both in recording and for viewing on a monitor, on output connections or through wireless monitoring, which means that we have direct preview with no need for LUTs.
Still in regard to software, worth noting is the fact that -in keeping with the latest trends- the camera has its own built-in WiFi adaptor. It supports both 2.4GHz and 5GHz networks, in addition to the dual link capability offered by the XDCA-FX9. WiFi can be configured either for connecting to the net through an external router or for managing the net autonomously and being able to provide services to remote devices that are connected to the camera. In this fashion and thanks to the Content Browser Mobile software, both remote control and monitoring can be made available to a smartphone or a tablet with a quite reasonable delay –about one second- that varies depending on network and propagation conditions of our connection.

And what about user experience? An extremely familiar one. Users acquainted with the cameras of the FS5 and FS7 ranges will find here a tool with which they will feel “at home” from the very beginning and their hands will find everything they need with hardly any need to look. In this sense, we find it a good idea that some similarities may facilitate transition between models, at least for the same manufacturer.

The camera feels somewhat heavier than the FS7. But the result achieved through the combination of the full-frame 6K sensor; a much quicker, accurate autofocus; the versatility provided by the variable neutral-density filter and the simplicity of working with the new S-Cinetone curve, make work with this camera a bliss, because we can attain excellent results with astonishing ease.
Before ending this feature, find below a reference of some of the novel features that will be available starting October 2020 free of charge, once version 2.0 of the firmware is released:

– 4K 50/60p full-frame recording.
– HD at 180fps and 4K DCI 17:9 recording.
– 16-bit RAW output and support for DWX audio receivers through the optional XDCA-FX9 adaptor.
– Autofocus capability with eye-tracking for the person of our choice.
– 6G-SDI connectivity.
– HDR capabilities with Hybrid Log Gamma.
– Touch-screen functionality for the viewfinder.
– Increase in the upper limit of gain.


Thanks to its extremely wide ranges for capture and configuration and customization possibilities, this camera offers sufficient versatility as to tackle high-level assignments, even as a ‘B’ camera alongside a VENICE. But at the same tine, it is capable of providing good performance in ENG-style single-operator environments, in view of the efficiency shown by its options and automation.
Our conclusion: this is a tool that will allow us to concentrate all our efforts and creativity in telling the best stories without having to worry much about all other constraints, thanks to all possibilities offered by its technology.

Grass Valley’s ‘
Imagine Communicatio