The new Sony FX30 has been much leaked and much rumoured. I got to see one and briefly play with one at IBC and I actually think this is a rather exciting camera. From the outside it looks just like an FX3 and the overall specs and menu system is pretty much exactly the same. The body of the FX30 does use a bit more plastic than the FX3, but this does make it a bit lighter without it feeling flimsy.
The big difference is the sensor. The FX30 has an APSC sensor, so this means that for video it is the equivalent of a super 35mm sensor. Go back 5 years and Super 35 dominated the large sensor video camera market. But now Full Frame is the flavour of the day, so s35 seems a bit dated, but is it?
Super 35 was/is the most widely used frame size for film production. There is a vast range of lenses available for Super 35. If you want a parfocal zoom lens there are many more options for s35 than FF. s35 lenses tend to be smaller, lighter and more affordable that their direct full frame equivalents. There is a vast range of classic cinema glass out there.
One of the best things about Sony’s Venice and FX9 cameras is the ability to use either full frame or super 35 for 4K. So, you can pick and chose which to use depending on what you are shooting. For wildlife and natural history super 35 remains very popular. For news and documentary work the range of zoom lenses available for super 35mm means that many will chose s35 over full frame. And sometimes you just don’t want the extra shallow DoF that FF may mean (of course you can always stop down in FF).
An interesting proposition.
I think the FX30 is a very interesting proposition. For a start it’s a fair bit cheaper than an FX3 (half the price). In addition APSC lenses are cheaper than full frame lenses, so for a student or someone starting out it’s an interesting low cost option. It might also be an option for a crash camera or some other hazardous job where the camera may get damaged. APSC lenses are often smaller and lighter than full frame so this may allow you to get the camera into smaller spaces. If you have an FS5 or FS7 you will be able to use all of the same lenses.
Dual Base ISO and Speed Boosters.
Something I learnt about the FX30 is that the sensor it uses is a true dual base ISO sensor. As a result there is almost no difference in noise or dynamic range between its low base of 800 ISO and it’s higher 2500 ISO base (for S-Log3). I also feel that this is a more useful split between the two ISO’s. With the FX3 and FX6 the higher base sensitivity of 12,800 ISO is often a lot more than you really need and it is not a true dual base ISO sensor. Instead the FX3 and FX6 have two base sensitivity modes and this means the higher mode is noticeably more noisy and the dynamic range is very slightly reduced. This might be great for specialist low light work, but it’s not so useful for drama or short films. If you do need to work at very low light levels then you can add a speed booster to the FX30 and use full frame Canon or Nikon lenses. In fact, if you want the so called “full frame look” (something that doesn’t really exist and is a bit of an internet myth) then use a speed booster and full frame lenses.
26MP Sensor = Oversampled 4K.
If your sensor has 4096 x 2160 pixels, that’s only 8.8 megapixels. 26 megapixels on an APSC aspect ratio sensor, with 20MP used for video suggests that you’ll have around 6.2K x 3.6K of pixels when shooting video, so your 4K recordings will be nicely oversampled. Interestingly the raw output seems to be 4.7K x 2.6K, so I’m not sure quite what goes on when shooting raw (it appears to be oversampled or downconverted). Potentially the FX30 may deliver higher resolution images than possible from the 4.2K sensor in the FX3 and FX6 (remember a bayer sensor only resolves at around 70% of the pixel count). I need to test this! This oversampling also means that if you want to shoot anamorphic, even after you have made the necessary crops into the image, the resolution will exceed that of a classic open gate Arri camera, although the side crop will mean you will have a reduced field of view. For anamorphic the FX30 may be an interesting choice, don’t forget that most anamorphic lenses are made for 35mm movie film, not full frame. 1.2x anamorphic lenses on the FX30 will look great. Of the lower cost cameras I still feel that the FX9 is the best choice for Anamorphic.
The dynamic range is specified at around 14 stops, so perhaps a little lower than the FX3/FX6/FX9 etc. Perhaps given the smaller pixels this isn’t really much of a surprise. But again this needs to be tested and in most applications I doubt anyone will really notice.
Rolling Shutter?
Typically Sony’s APSC sensors have exhibited more rolling shutter than the full frame sensors in cameras like the FX3 and FX6. And this remains the case with the FX30. It does exhibit more rolling shutter than these two cameras. But it isn’t terrible, I’ve seen much worse. I didn’t get a chance to do any side by side tests so it’s is difficult to be precise, but I don’t thing the rolling shutter is any worse than an FS5 or FS7.
CineEI Mode.
With the same CineEI mode as the FX3 the FX30 will be a great camera for shooting Log. Like the FX3 it can shoot at up to 120fps in 4K. There isn’t much not to like about the FX30, especially for those on a tight budget or those that need a super 35 sensor.
Ultimate Webcam!
Another exciting feature is the ability to use the FX30 as a webcam. The FX30 supports the UVC and UVA standards allowing it to be plugged into a computer via USB to use it as a high quality web camera or streaming device. My brain is already thinking about things like using one to stream the Northern Lights from Norway next January.
If you like what you read here, please also visit and subscribe to my YouTube Channel by clicking here. Thanks.
About the oversampling. On some forums they say the camera is cropping. So how do you know if it’s oversampling or cropping? Just curious. Thanks!
I think I am mistaking the cropping when shooting stills compared to shooting video. It is still oversampling. That’s nice. I’m curious how evident the quality difference is with the FX3 of FX6 image at the same settings.
The FX30 has an APSC sensor that has a total of 27MP. 26MP are used for stills. In video mode you only use the 17:9/16:9 center section which is 20.1 MP which is more than double the number of pixels found on a 4K 8.6MP sensors.
Exactly how good the downsampling is remains to be seen and will need careful testing. What I would hope is that the FX30 might exhibit less moire than the FX3/FX6 as well as a slightly higher resolution image. Whether you’ll see this in most day to day shooting is questionable as there are not many every day natural textures that will show up that difference. Where you may see a difference is when cropping the image or doing complex prcesses such as green screen or anamorphic extractions. Another benefit of over sampleing tends to be a finer noise structure, but this will be offset to some degree by the smaller and very slightly less sensitive pixels.
So just getting it straight.
17mm lens for full frame equals 1.5x for apsc,
so becomes 25.5mm equivalent,
Then 1.6x for 4k 120, so that 17mm ends up being 40.8mm FF equivalent?
This is all wrong.
35mm film has been the standard for movie making for the best part of a century and cinematographers have understood excatly what feild of view a 17mm lens will deliver. They have never ever though in terms of other equivalnts because 35mm is normal.
A 17mm lens is a 17mm lens no matter what sensor you put it in front of and it never becomes the equivalent of any other focal length as it will exhibit the same DoF and CoC characteristics whether on a full frame sensor or any other sensor. A 17mm lens on a S35 sensor is not the same as a 25mm lens on a FF sensor, DoF will be different and more than likely many other image charteristics will also be different due to the different optical designs needed to achieve the different focal lenghts.
With the same lens the feild of view with a FF sensor is 1.5x wider than s35. So a 17mm lens will deliver a 1.5x wider FoV on FF than S35. If you want the same FoV, then you would use a 17mm lens on s35 and a 27mm lens on a full frame sensor if the difference between the sensors is 1.6x. But this doesn’t make them the same as other charcteristics will be different. Crop factors should be applied to the feild of view, not the focal lenght.
Very interesting camera indeed! Does it have full pixel readout without pixel binning? I can’t seem to find the information from sony website.
Unlikely to be pixel binning as it would be difficult to achive 4K resolution from 6K of pixels with 2×2 binning. Skipping would result in aliasing artefacts.
Sorry Alister I meant for HD 1080p mode. For audio, all modes are full pixel reading. Not sure what it is on the new fx30.
Any idea how the raw behaves in terms of sensor sampling, when it is just 4.7k
is it binning or skipping to get full sensor or doing 4.7k 1:1 in the middle
I don’t know for certain, but I believe it is a center crop.
How would low light / noise performance compare to the Fs7?
I know given cine ei and standard that’s a very open question but as a b cam looks interesting
Being an Ex fs7 user, and now using a lot of fx9 and fx6, fx3
Just my guess!
I would suspect this being about a stop or stop and a half better than the fs7 in lowlight. Fs7 being as good in lowlight as it was I think came down to some noise reduction that you could not turn off, which gave that camera a bit of the plastic look.
Where’s the new generation is a lot cleaner in the image excluding the fx3 which is internally noise reduced. this fx30 seems to be less noise reduced, which is why I find it very interesting, if 2500 is good and usable for me I am happy with that.
More than 1 stop better. You have 2500 ISO with less noise when you use the upper ISO plus it’s over sampled, so the noise is very fine, so you can shoot at 2500 EI on the FX30 without issue compared to having to shoot at 800 EI on the FS7 for a similar noise level.
Hi Alister,
Thank you for all the great stuff here, just wonder i’m using
still an FS7 with a Fujinon MK 1855 and this might be a good
B-cam for smaller spaces likes cars, do you think it is not to hard
to match these 2 cameras with there different color science?
The FX30 and Fujinon MK’s will be an amazing combo. I think you would find that the FX30 would be the “A” cam and the FS7 the “B” cam as the improved colour science makes the images so nice. And that does make it a little trick to match them as you will get slightly different results under different lighting, so sometimes matching will be easier than others.
might be a bit difficult to built some kind of a rig with the
long lense, also have an external viewfinder from zacuto
which i love to have under the always changing lighting
conditions…need to think about it