Interesting low light test
 

I was doing some low light tests with my new EA50 last night. It does reasonably well using the stock lens, and the image looks clean even at ~24dB gain.

For the sake of curiosity, I did a quick comparison between my JVC HM600 (1/3" sensor) and the EA50. Both cameras were set to 1/50 shutter, 0dB gain, f3.5.

You can see from the image below that the HM600 does a far better job. How is this possible considering the HM600 uses such a small sensor in comparison? My technical knowledge of the inner workings of cameras is fairly limited.. so I'm guessing that even though the HM600 is set to 0dB, there's still some digital gain or image enhancing trickery happening?

Re: Interesting low light test
 

Hi Jody,

I have to say that never in my life have I ever used anything more than 9db gain, but with the NEX50 24db gain is expectable, yet all of my other Cameras are either 1/3 or 1/2 Chip.

Re: Interesting low light test
 

Well the game is different is with the DSLRs or at least with this camera. I compared it against my AG-HMC 150s the results I got were at 0bb the Panasonic blew the Sony away BUT when I started gaining up I got a lot more light per gain up vs the Panasonic and a lot less grain. At the end the Sony was at 21 db gain and was equal in light to the Panasonic at 12db BUT the Panasonic was noisy but I couldnt see any noise on the Sony.

Re: Interesting low light test
 

I just had a thought. Could it be because the HM600 (and most small sensor cameras) have native 1920x1080 sensors, therefore the pixels are physically bigger and let in more light. The EA50 has a 16MP sensor so the pixels would be tiny in comparison.

Still, would have thought the EA50 would blow the HM600 away in low light.

Re: Interesting low light test
 

I wouldnt The JVC is the closest thing to the "old days" of DV cameras out there. Its the only one that gives you a honest low light spec. plus if your using the stock lens of the Sony your comparing a camera that is at f-1.4 vs f- 3.5 at best.

Re: Interesting low light test
 

It comes down to the native ISO of the chips. The EA50 being based on a stills camera probably sits at ISO 100 whereas the HM600 probably sits at around 600 .. Gain the Sony up to match and it should still be cleaner than the the 600 ungained.
A 2 megapixel 1/3" chip actually has smaller or very similar sized pixel sites to a 16 megapixel super 35mm sized sensor. You can fit between 8 and 9 of them into a S35 chip. I think the JVC is 3 x 1/3" CCD as opposed to a bayer pattern single chip CMOS so not a like for like comparison..Could be confusing my cams tho - i own the 700..

Re: Interesting low light test
 

Hi John

Quite correct..the base ISO of the EA-50 is actually 160 not 100 but still way under the traditional 3 x 1/3rd chip cameras native ISO. Are you sure the JVC has a native ISO as high as 600? My 1/3rd chip Panny's were a LOT lower than that and I remember my HMC82's had a miserable ISO of just 40!! I know Sony's EX series have a base ISO around 600 or 700 but those are 1/2" chips.

Regardless of that fact, even with the stock lens at F3.5 I still get a way cleaner image than I used to get with my Panasonic AC-130's .... I used to have to use a video light on the 130's almost at full output during a reception but pop a constant F2.8 lens on the EA-50 and most of the time I can turn the light right off ..I normally only use it when the dancing starts and the venue virtually plunges into darkness.

Chris

Re: Interesting low light test
 

Jody

Using colour curves push the brightness up on both clips. See what happens ...

I would of thought you can colour correct / push levels up on the EA50 better.

Re: Interesting low light test
 

Chris - I'm not sure what the native ISO of the HM600 chips are - I checked - it is 3 x 1/3" CMOS camera - whereas my HM700 is 3 x 1/3"CCD. In general up to date CMOS cameras are faster than the older CCD based chips but the ISO could be anything and I don't think JVC gives that as a spec..I was guessing from the pics supplied.

The size of the chip has nothing to do with it's native ISO - and being "slow" doesn't necessarily equate into being a poor low light performer. The EA50 being natively slow could actually be a good thing in some situations as you may get away without ND filters (not bright sunshine but maybe overcast).

The proviso is that at the same ISO as a similar camera (lets say the FS700 because I have that cam and I happen to know that it is natively ISO 600 on most of its Pic profiles and that it is a S35 based sensor as well) that the noise profile is similar - in other words the EA50 should still be very clean at ISO 600. I don't expect it to be as clean but hopefully it's close.

Re: Interesting low light test
 

Hi John

In fact the EA-50 is remarkably clean at ISO640 which is only 12db gain!! I can shoot receptions at 21db gain and the image has no noise ..it does tend to start appearing at 24db ... At ISO5000 which is 30db it's decidedly noisy but that's quite a gain boost indeed. I'm pretty happy if I can keep mine at 21db and I also limit it there

I think the general technical opinion is that fewer pixels on the chip means better light collecting capability but lower resolution ... in the good ole days where the image was an SD 4:3 manufacturers could manage a rather good low light chip with fewer pixels as less resolution was needed for the smaller image. The outstanding low light performance of the old Sony VX2100 is an excellent example.

Chris