In the midst of the expansion of AI, with this camera comes the time to knock down another of those concepts that until now were part of our “great immutable truths”

*This article is included in the February issue of TM BROADCAST. You can access the full magazine here.

Technological evolution never ceases to amaze us. Automatic functions have been an invaluable help for many years, although we have always said that a good professional outperforms any automatism. And that, at least for the time being, is still a totally valid statement.

Needless to say, as strong advocates of innovation and technology, we always welcome all relevant advances and developments with enthusiasm. And while this increasingly facilitates participation of professionals at different levels, we would like to emphasize that a well-prepared professional will always enhance any tool and will not just be content with relying on it.

It is true that autofocus systems are faster and more accurate than most of us, but in general they do not know how to correctly identify the point of interest in each scene. This has improved since these systems are now capable of recognizing faces, eyes, and even specific people to keep the focus on them and not on someone else.

And although at the moment they still lack the ability to bring the focus to the take of narrative interest in certain situations, the ability to select the desired point or area of focus at any time on the touch screen is a great advance.

Something similar is provided by exposure automatisms, capable of very precise adjustments, but only as long as the lighting is not complicated and we manage to “tell” them in some way what the critical area is for our content to be really useful for our purpose. Which, again, can probably be more dependent on narrative needs than on the quantity, quality, or direction of the existing light.

It is even a frequent and common feature that PTZ-type cameras are capable of tracking people, although the innovation we discussed in our recent BRC-AM7 lab test (also from Sony) with the ability to maintain a certain frame by specifying the air on the contours, on subjects who move quickly and in wide spaces is indeed impressive, thanks to its integrated movement system.

But, back to our headline… how is is possible that a handheld camera can properly frame, if does not have the ability to move by itself? This is one of those immutable truths that we must now revise.

Let’s see what all this is about, and although the answer is simple and it does have limitations -like all automatisms- there are situations in which this will help a lot. But we continue to insist on this great truth, which we would wish to remain unchanged: it will always be the hands of a good professional that will make any tool really effective and efficient.

As usual, before delving into the subject, let’s review the other features in order to provide proper context. It is a handheld camcorder, with integrated 20x zoom optics, a single 1.0-inch type R-CMOS sensor with native 5K resolution, and Ethernet + WiFi connectivity.

It seems to be designed to cater to needs mainly in Broadcast and ENG areas, and in view of other features, it will also yield excellent results in areas from web creation to aspirational filming.

It is a handheld camcorder, with integrated 20x zoom optics, a single 1.0-inch type R-CMOS sensor with native 5K resolution, and Ethernet + WiFi connectivity.

The integrated zoom optics cover an optical range of 24-480 mm in equivalence to full-frame focal lengths. In combination with the Clear Image Zoom -a built-in function that uses additional processes to avoid the characteristic artifacts of digital zoom- the initial 20x can be increased up to 30x without noticeable loss in image quality in UHD; and up to 40x when working in HD, thus reaching a truly extraordinary 24-960 range.

Brightness starts from an excellent f/2.8 at an angle, which goes down to f/4.5 on telephoto. The focusing distances decrease to only 1 cm at an angle, being even capable to focus inside the parasol, and grow to a minimum distance of 1 meter with telephoto at maximum range.

That is, although the optics are not interchangeable, it has sufficient qualities to solve most situations, with the comfort of being perfectly integrated and forming a solid and compact whole with the body.

It comes with a built-in optical stabilizer with two operating modes. The two independent rings for focus and zoom are complemented by an iris control that is carried out by means of a knob placed in a comfortable and natural position for any operator, although already on the body and not on the optics barrel.

Its position, close to and a twin of the knob that controls the variable neutral density filter -already characteristic of the brand- makes it easier for us to always have the best possible exposure control.

By the way, and according to information provided in the technical documentation, it is possible that some upcoming firmware update will allow the iris control function to be assigned to one of the main control rings. We find that this possibility not only facilitates the customization of the equipment, but also increases its versatility by enabling to tackle different needs with maximum ease for different operators or various scenarios.

The variable ND functionality, which can also be configured at fixed levels like the classic filters and not necessarily in progressive order as in the rest of models, can be set up to run on auto mode. In this way, the selected exposure will be maintained, compensating for the differences in lighting when moving from one environment to another without affecting the visual narrative.

This ability to control the exposure level without modifying diaphragm, shutter speed or sensitivity, means that depth of field, movement or noise in the image will not be modified even if lighting levels vary by up to 6 diaphragm steps, which is its operating margin between 1/4 and 1/128 exposure, and always without causing any chromatic alteration.

For these reasons, variable ND is one of the features that has seduced us since it appeared years ago, and we are very happy to see that it is a natural part of all the new equipment that Sony launches on the market.

The Exmor RS CMOS sensor, 1.0 inch-type, has an effective 14 Mpixels that allow for 5K resolution. In addition to resolution, there are dedicated BIONZ XR processors and a specific AI module. This set, in addition to contributing significantly to the reframing procedures that we will see in detail below, facilitates greater focusing precision by combining contrast detection and phase detection techniques, thus optimizing both speed and accuracy.

This processing speed of the sensor assembly is what allows recordings with high frame rates, which reach 120p in UHD format (3840 x 2160) or 240p in HD (1920×1080).

Color response can be configured in different spaces including S-Cinetone, S-Log3, HLG, and various variants of 709, thus facilitating integration with different cameras, environments, and workflows. That is, it will always be easy for us to match the chromatic response with both Broadcast cameras and Cinema Line range cameras.

Operating margins in the classic image capture parameters are very broad: shutter speeds between 1/8000 and 64f; white balance between 2,000 and 15,000 ºK; gain from -3 to 36 dB in 1 dB steps. ISO between 250 and 16,000, with the possibility to configure lower and upper limits if automated.

The autofocus has up to 475 detection and recognition points for people, with subject detection. That is, you can identify a person to track focus on a particular individual. And we’ll see that it’s not just the focus…

For visualization we have a viewfinder and screen, as is also usual in this type of equipment. The 1 cm XGA-OLED viewfinder features 2.36 Mpixel and proximity detection if we are using it, in which case it turns off the screen. As it is logical, this mode is configurable so as to prevent it from being accidentally turned off if it is the hand or the body which covers the eyepiece.

The 3.5-inch 2.75 Mpixel LCD display offers excellent resolution and image quality, easily visible in broad daylight, and features touch functionality for ease of operation. We have a folding and removable sunshade as standard and three buttons on the side always accessible with zebra, peaking functions, and a customizable button.

The amount of information that can be overlaid on the image during recording can become overwhelming. Fortunately, all of this information can be turned on and off with a simple tap on the assigned button. This information includes not only data in the form of text or figures, but also graphs, such as histogram, vectorscope, or waveform.

By configuring this through the menu, LUTs can be used, both on-screen and in one of the outputs, so that our display is as close as possible to the final result that we will obtain once the color-grading of the filming is completed.

Color response can be configured in different spaces including S-Cinetone, S-Log3, HLG, and various variants of 709,

The LCD support has been newly designed, and although it feels robust and inspires confidence, due to the position in which the screen remains once folded, it gives the impression of being somewhat vulnerable to not very delicate handling. Only the qualities of the new materials and the experience of use over time will judge the benefits of the new solution.

Regarding operation, another feature that we really liked are the various methods to control all its functions and capabilities: on the one hand, the buttons and selectors with dedicated functions, in the style of classic ENG cameras, in the usual places.

In addition, most of the 12 customizable buttons distributed throughout the body offer a different touch to make it easier to locate them based on touch, and through the menu we have a range of up to 63 functions among which we can choose for each of them. In this way, we will always have maximum efficiency when dealing with any type of project.

On the other hand, we have access to the menu in two modes: summarized mode, so as to access the main configuration functions (with a short press), and extended mode (with a long press) to control every single detail of each parameter. These two modes are available on the two menu buttons, one of them being very accessible on the handle itself for use during filming if necessary.

And finally, the combination of an on-screen quick menu with the joystick to switch between parameters and adjust each one directly. These parameters are the ones we usually find as information displayed on the edges of the screen during filming and, in fact, this mini joystick is duplicated on both handles (side handle and top handle) next to the two recording buttons.

Regarding audio, exceeding what has been customary recently, we have up to 6 physical inputs for 4 audio channels, always recording audio in LPCM 48 kHz format at 24 bits.

We have two inputs through the XLR connectors, with the possibility of power supply, and classic direct controls for the input and level selector. Two others through the MI-shoe, capable of carrying both analog and digital signals and whose controls are accessed through the main menu. And in addition to the stereo input provided by the built-in microphone, we have another 3.5 mm TRS mini-jack input for an unbalanced external microphone.

Regarding video recording formats, it is important to note that it will invariably be done on AVC, intraframe or long-GOP codecs, and always in progressive. Internally, we no longer have MPEG2 codecs or interlaced mode. Pay attention to this, if for any reason there is a need to generate interlaced content.

In summary, within the wide range of options available in terms of resolution and frame rate, we can select from a maximum of 500 Mbps in intra UHD at 50p, up to a minimum of 50 Mbps in long-GOP HD, either 420 or 422 and 25p or 50p. It seems that the reason for maintaining the minimum of 50Mbps is to always meet the standard supported for broadcast. As for proxies, these are made up of 16 Mbps.

The recording of MXF or MP4 containers is carried out on type-A CFexpress cards or SD UHS-I/II cards, in either of the two slots with configuration flexibility: different formats in parallel, the same one in relay, and even an independent control for each of them from each of the two recording buttons.

Obviously, all the formats that the camera is capable of generating can always be limited by the actual speed of the card being used. With the option to choose between SD or CFexpress, let’s keep in mind that not all of them can support the sustained data rate in recording required for the most demanding formats.

Being all very aware that capacity affects recording time, we must make sure about the sustained data rate in recording, since many manufacturers only advertise the read rate, which is always higher. Be careful, please, with this.

Moving on to connectivity, in addition to the audio inputs already described, we have two BNC connectors: one configurable as input or output for the time code, and the other for the SDI output. Another HDMI connector provides a second video output, as both SDI and HDMI outputs can be configured with different resolutions or overlays.

We have a 3.5 mm jack for stereo headphones, and a small mono speaker for monitoring recorded content. What is not so common is having a menu “reading” system that allows us to listen to options while operating the device. However, this feature is not available -at least for now- in all languages in which the camera can be configured.

As it is also common in this type of camera, we have network functions, both Ethernet 10/100/1000BaseT and Wi-Fi on 2.4 and 5 GHz bands, thus enabling direct streaming or uploading of files to any FTP platform, as well as dedicated cloud services such as Sony’s proprietary Creator’s Cloud C3 Portal, with QoS quality controls and real-time data flow.

Here we must make a quick mention of a new optional PDT-FP1 (Portable Data Transfer) device. Similar to the well-known external recorders, this one offers us a few more functions: a larger screen with 4K display and remote control functions on the camera, possibility of internal recording and, most interesting of all: additional connectivity featuring a stable data transfer at high speed with a multitude of bands, such as LTE, 5G Sub6 and mmWave. It connects via USB-C and releases the camera’s network functions for any other purpose.

These network capabilities allow, in addition to the 2.5 mm LAN-C connector, remote control of the camera from phones or tablets via Wi-Fi or Bluetooth, with the possibility to become its own hot spot in those places where there is no network available.

A USB-C port provides high-speed connectivity -up to 5 Gbps- with USB 3.2 protocol suitable for various uses, from content dumping to remote control.

Streaming is done using RTMP, RTMPS and SRT protocols with configurable H.264 and H.265 codecs in 720p, HD or up to UHD resolutions with 2 16-bit audio channels at 48 kHz.

The batteries are BP-U type, with the already known capacities that provide either lightness in the smallest BP-U35 model or long operating time in the highest-capacity BP-U100. The feeling is that performance has improved overall as compared to previous models, with up to 90 minutes of continuous recording with the smallest and lightest battery, according to the catalogue.

Keep in mind that consumption will always be determined by actual use, and that this can vary significantly if you use, for example, network services or not, the display or the viewfinder, prolonged viewing of content, etc., plus all their possible combinations. This is why we suggest that each operator carry out some tests in their environment and under actual conditions so as to obtain really accurate data for their working conditions, and thus have all the peace of mind that they will need when the time comes.

We end with the description of the set, confirming that all these qualities are in a comfortable and very manageable body, in which good care has been taken in details such as the fact that ventilation is designed so that the air flows from the left and down, to the right and up, moving the hot air away from the person operating the camera and thus providing greater comfort of use.

It has been manufactured by combining light metal alloys with new synthetic materials, in search for robustness, lightness and reliability together with the best balance with environmental commitment. In this way, the whole set weighs around 2.4 kg (little over 5 pounds) in operating conditions, battery included.

We have needed this entire lengthy block just to list and briefly comment on the large array of features we have at our disposal, as well as some of the benefits. But let’s get now into that new, fascinating autoframe feature.

Streaming is done using RTMP, RTMPS and SRT protocols with configurable H.264 and H.265 codecs in 720p, HD or up to UHD resolutions with 2 16-bit audio channels at 48 kHz.

As we can imagine in a camera without PTZ functionalities, the autoframing method is conceptually very simple. And it has its limitations too, as is the case with all automatisms, although this does not preclude a surprising result.

Basically, what we do is establish the cropping margin that we will allow so that, once the subject is identified in the same way that we do to maintain focus through identification of people, our protagonist remains in a prominent place in the image and not lost in a corner.

Right. The reframing is done by cropping and scaling the image, in the same way we would do in editing, but in real time during the recording itself! And while it is true that this can lead to compromising image quality if we are too permissive or aggressive with the setting, it can be tremendously useful in a multitude of situations.

Considering that the sensor is 5K, in an HD shooting scenario like the one still used in many news pieces, events or broadcasts, we can make a larger x2 magnification without loss of quality of any kind. That means that an operator in the middle of a crowd of operators in a news piece -even if not being able to aim carefully- will succeed in achieving a recording or a more than reasonable live feature in real time.

By combining the speed of image processors with AI functions, the versatility of subject identification with the effectiveness of autofocus, we achieve a result in which our subject of interest will be on focus at all times, and will stay as focused as possible within the frame of reference.

We have two parameters to control system response: on the one hand, the range of reframing or enlargement that we want; and, on the other, the sensitivity or speed of response.

This is where the skill of the relevant professional will make all the difference. Imagine a general take in an auditorium, where the subject will move within a limited space: the reframing will follow the speaker, thus offering a closer shot on the person, but the first decision will be to establish the limits of the general shot, since the reframing, obviously, can never exceed them.

Then, if we force the enlargement too much, we could compromise image quality. If we magnify too little, the result of the tracking will be insignificant. If we exaggerate sensitivity, the protagonist’s small movements could lead to unnatural tracking. And, if we reduce it excessively, we risk having to get too close to the limits before reframing.

For all these reasons, being a system that generally works quite well, unwise tweaking can lead to an undesired result, far from our expectations and from the true capabilities of the system.

Another example of very practical use is when a single operator does an interview, in which we will devote all our attention to the content without also having to be aware of the framing of our protagonist while carrying out the interview.

On the other hand, with a correct setup, and if we also add optical stabilization, the shot achieved from a crowd that keeps us moving while we follow the relevant news piece, can end up looking like a perfectly planned and executed shot.

As with the subject identification system, if our protagonist exits the take we will lose the tracking, but we will recover it as soon as the person re-enters the scene. In this case, having configured autoframe activation/deactivation in one of the customizable buttons will facilitate direct and instant switching of the function.

Finally, just mention that together with this camera a similar model  (NX800) has been presented at an even lower price which, sharing the same body and most of its features, only cuts on some specific broadcast functions such as the BNC connectors with the SDI output, the timecode and it only records in MP4 formats.

In short, a new progress in content creation possibilities that provides the most sophisticated features currently available at a highly competitive price. Summing up, and insisting on our initial idea: being delighted with all the advances, we continue to argue that a good professional will provide unquestionable value, by enhancing the best tools instead of being replaced by them.

The reframing is done by cropping and scaling the image, in the same way we would do in editing, but in real time during the recording itself!