I've seen detailed specifications for the Sony NEX-FS100 published on an Australian vendors website that states the signal to noise ratio for this camera as "54 dB (Y) (typical)". Could this be an error?


VideoGuys Australia | Sony NEX-FS100PK Digital Super 35mm Professional Camcorder (http://www.videoguys.com.au/Shop/p/24097/sony-nex-fs100pk-digital-super-35mm-professional-camcorder-65279-nex-fs100pk.html)

That's the same s/n ratio as the Sony EX1 which is quite a noisy camera and it's nowhere near the 63db s/n spec for the Sony F3.

I saw the FS100 today and I was playing with the gain at 30db. It was surprisingly cleaner than any conventional video camera I have seen 8K and under. It is less noisy than the EX1. Didn't bother comparing it to the F3 as I can't even justify the F3 in my line of work. The FS100 is surprisingly clean with the 54db/SN

Is that a real 30db or did they just rename 18db? Only asking because I've not seen a relatively cheap camera with gain that goes that high.

According to Alan Roberts' test of the F3, its signal/noise ratio is really about 48.5 db. Since it has the same sensor as the FS100, maybe we should expect the same.

Is that a real 30db or did they just rename 18db? Only asking because I've not seen a relatively cheap camera with gain that goes that high.
I don't know that answer. I do know the sensor is ISO 800 and applying gammas does knock it down like the F3. Whatever they are doing 30db is less distracting than 6db on my XH-A1 that uses noise reduction. I don't know how the F3 steps are in gain but the FS100 is 3db I just wish they used ISO in it as well. You win some you lose some.

I wouldn't use anything from Alan's report except the Zone Plates. His testing is so way off. He is trying to say that the F3 is better at 6db than 0db and 18db is best for programming.

FYI, for 10bits, a camera must have at least 60db of sensitivity with 6db needed per 1bit. So, at best, the FS100 could only provide 9bits; however, the S/N is probably a little less than what Sony has stated.

For this reason, I don't see any benefit even if the camera had a 10bit output although SDI would have been great.

I don't know that answer. I do know the sensor is ISO 800 and applying gammas does knock it down like the F3. Whatever they are doing 30db is less distracting than 6db on my XH-A1 that uses noise reduction. I don't know how the F3 steps are in gain but the FS100 is 3db I just wish they used ISO in it as well. You win some you lose some.

Thanks for the reply. If it is real I fully expect some side by side comparisons showing the FS100 12db brighter than the F3 at the same F-stop. If not I call BS.

I wouldn't use anything from Alan's report except the Zone Plates. His testing is so way off. He is trying to say that the F3 is better at 6db than 0db and 18db is best for programming.

Allan knows more about camera testing than just about anyone in the world. But, he isn't the best writer. Actually, he is a terrible writer as are many engineers. Nevertheless, one would really have to be a 2 year old to assume he was RECOMMENDING you shoot at 18dB. What he means is you can feel safe using 18dB gain for shooting programming. He is correct. 18dB is 3 stops of gain. 3 stops can be used with all NEX cameras with near zero visible gain noise. Only at 21dB and above does noise begin to show.

Given this, I see no need to compare noise "measures" on different cameras unless the SAME test is used.

PS: Only after reading years of Allan's BBC whitepapers do you see he assumes that when he writes a short sentence on "DR and Knee" YOU remember the report 5 years ago when he explained DR, Knee, and how he tests. He also tends lump cameras into two camps: $50K+ that he assumes will be usually be used by those creating for the BBC and the whole camp of "cheap" cameras. I think in the first few paragraphs he makes it clear the F3 is not in the former camp.