JVC HD100 vx. Sony FX1 sensitivity/noise

Ralph Oshiro

Ralph Oshiro

Veteran
I am actually planning on buying a Sony FX1 since I am inferring that the FX1 is more light sensitive that the HD100. My MAIN criteria in choosing my next camera is LOW LIGHT PERFORMANCE, and LOWEST CHROMA NOISE. The HVX200 is NOT a consideration, since I plan on buying Sony's new 1/2" XDCAM-HD camera when it ships next year. The Sony FX1 is an interim solution to be able to shoot some shorts before my XDCAM purchase. So . . .

Of all the native-16:9, 24p (including Sony's CF24) 3CCD prosumer cameras out there now . . . which one is MOST light sensitive and LEAST noisy? I guess that means, the HD100, the FX1/Z1, and the XL2. I guess it would be the XL2, since it's only an SD camera with larger pixels? But is it true that the FX1 is perhaps a full stop faster than the HD100 when both are in their respective 24fps "progressive" modes? I've seen CF24, and I don't mind it. And I DO NOT plan on shooting interlace and using Magic Bullet--the costs in workflow are too high. I'd much prefer the compromises of CF24 regardless, if, I GAIN greater sensitivity and/or lower chroma noise. Barry, any side-by-sides of these two cameras yet?
 
Nate Weaver has posted some HD100 low light stuff here:
http://www.dvinfo.net/conf/showthread.php?t=51927

It isn't very noisy (all at +0dB I think), a bit of FPN, but not too bad. The smearing is quite bad, but the FX1 has the same problem. A side-by-side would be good.

People seem to be saying that the HD100 is actually more sensitive than the DVX now. I think people are learning how to get the best from the camera.
 
Thanks, mmm. Just to make my motives clear . . . the SINGLE BIGGEST REASON I'm dumping my DVX100 RIGHT NOW is that I can't stand the NOISE. Yeah, its true 24p look and cinegamma curves are great, but it's the noisiest damn camera I've ever used and I'm sick of it! I'm just wondering if the FX1 is the low-cost, lower-noise, native-16:9 solution for me NOW. I'm not even considering the HVX200 since I'm planning on buying the 1/2" Sony XDCAM-HD when it comes out next year.
 
The DVX is indeed noisy to a sickening point, the stuff of the HD100 seems cleaner, but the XL2 is defenately the cleanest (imo...)
 
It is really starting to annoy me that the assumptions that the JVC HD100 sucks in low-light have suddenly turned into fact. This is not the only thread on the subject.

If you fire the thing up in the factory set Cinelike mode and compare it to a DVX100 or Z1 in low light it will appear to be less sensitive. This would be like buying a sports car and never taking it out of 2nd gear. The camera is considered pro for a reason: the ability to control the response curves in many different ways.
I've said this before, but I guess no one believes me. So here's some proof:

This is a quicktime (sorenson codec so you PC guys can play it) of a test shoot I did Sunday for a film I am prepping now.
The available light level was never any more than 5 footcandles, the average ambient light was probably around 2 footcandles.
For ƒ1.4, 1/48th shutter that would mean a sensitivity of 500 to 800ASA.

I want to stress that this was shot at 1/48th shutter at 0dB, except for the overcranked 576p50, which was shot at 1/50th shutter (I should have switch to 1/59.94 to avoid flicker.)

http://homepage.mac.com/timdashwood/.Movies/Ex-lover_test_shoot_assembly-sorenson.mov

The Camera Process settings were as follows:
MASTER BLACK... Normal
DETAIL... -1
BLACK... STRETCH3
WHITE CLIP... 108%
KNEE... MANUAL 85%
CINELIKE... OFF
GAMMA... STANDARD
LEVEL... MAX
 
I wanted to avoid talking again about the split... but I'm obduratted with it!! I can't say if I can really see it or if it's my imagination tricking me, I think I saw a split screen...

Poor girl!!
 
NBCshooter said:
I am actually planning on buying a Sony FX1 since I am inferring that the FX1 is more light sensitive that the HD100.
Click to expand...
I would be very surprised if this was true.

My MAIN criteria in choosing my next camera is LOW LIGHT PERFORMANCE,
Click to expand...
Then you don't want the Sony...

and LOWEST CHROMA NOISE.
Click to expand...
Then you may not want the JVC... the Sony is amazingly low-noise. But then again, so's the XL2.

Of all the native-16:9, 24p (including Sony's CF24) 3CCD prosumer cameras out there now . . . which one is MOST light sensitive and LEAST noisy?
Click to expand...
Probably the XL2.

But is it true that the FX1 is perhaps a full stop faster than the HD100 when both are in their respective 24fps "progressive" modes?
Click to expand...
Don't know. True progressive usually means a stop slower than interlaced, but the FX1 is two stops slower than a DVX, and JVC claims that the HD100 is one stop slower than the DVX. So I don't know how the FX1 could suddenly jump up two stops. The FX1 is really not very sensitive at all; we rated it at around 160 ISO.

I've seen CF24, and I don't mind it.
Click to expand...
That's surprising. Look closer. CF24 is awful.

I'd much prefer the compromises of CF24 regardless, if, I GAIN greater sensitivity and/or lower chroma noise.
Click to expand...
CF24 is just horrid. Spend a little extra and get the Z1, you'll get more features, better audio, and CF25 which is at least sixty times better than CF24.
Barry, any side-by-sides of these two cameras yet?
Click to expand...
Side-by-side between the HD100 and the FX1? Not yet, the HD100 is supposed to arrive tomorrow, and I'd have to arrange to get an FX1.

Interesting that at no point in the discussion do you mention anything about high-def. If HD isn't a concern for you -- the XL2 merits strong consideration. You'll get low noise, better sensitivity, true 24p, and an overall lower cost than either the HD100 or Z1 (ruling out the FX1 just because of CF24). And you should be able to pick up a used one for pretty darn cheap if you can find someone who's wanting to sell so they can upgrade to an XL H1...

... but that's only if you really don't care about high-def.

Something else to consider though -- do you plan on selling off the 1/3" camera when getting the XDCAM-HD? Because if you plan on keeping the 1/3", then the Sony, besides its drawbacks, will retain one big advantage -- it'll probably match much better with the XDCAM-HD camera. Its 1080i mode will be compatible, and its overall "look" will probably be compatible too.
 
mmm said:
People seem to be saying that the HD100 is actually more sensitive than the DVX now.
Click to expand...
I've seen those claims, and I've also seen claims that the HD100 has actually *more* dynamic range than the DVX.

Both seem unimaginable. I mean, even JVC themselves claims a sensitivity of F8 @ 2000 lux, vs. Panasonic's claim of F11 @ 2000 lux for the DVX. And JVC claims a contrast ratio of 300%...

It'll be easy enough to verify both claims this weekend...
 
Barry_Green said:
I've also seen claims that the HD100 has actually *more* dynamic range than the DVX.

Both seem unimaginable. I mean, even JVC themselves claims a sensitivity of F8 @ 2000 lux, vs. Panasonic's claim of F11 @ 2000 lux for the DVX. And JVC claims a contrast ratio of 300%...
Click to expand...

Remember that the codecs are completely different (mpeg 2 & DV).
The CCDs themselves obviously capture much more dynamic range than can be compressed and encoded onto tape. The guys who created the "Reel Stream Andromeda" have proven that. http://www.reel-stream.com/andromeda
Look at the comparisons of what the CCD can capture and what actually makes it to tape. http://www.reel-stream.com/cap_comp
They key is in the control of gamma curves. The DVX100A has some great Panasonic curves to choose from, but fact that you can customize the shape of the curves in the HD100 allows for a DP to customize the settings to allow for more dynamic range. This is also true with the XL2, and pro cameras like SDX900, F900, Varicam. The HD100 probably falls between the XL2 and SDX900 with the amount of control.
I haven't done my dynamic range comparisons yet, but will get to it as soon as I have a free day. In my limited testing of the HD100 though on a bright day, blue sky, white clouds, subject in the shadows (5 stop difference) and a knee setting of 80% and the black stretch turned on, I can get way more information on tape than I would ever have expected from the DVX's CINELIKE D or V curves.
 
athouguia said:
I wanted to avoid talking again about the split... but I'm obduratted with it!! I can't say if I can really see it or if it's my imagination tricking me, I think I saw a split screen...

Poor girl!!
Click to expand...

There is at least one shot in the test with a few frames of the split. It is intermittent and very telling of what is happening in the image processors. Flares on one side of the screen seem to trigger a slight shift in the blacks. I hope this isn't some sort of "auto knee" circuit that can't be shut off.
My other theory is that a flare on of of the CCDs (left or right) only extends to the physical limits of that chip and doesn't affect the other.
I'm still thinking about this, but now I know specifically what to test in the future.
 
dashwood said:
My other theory is that a flare on of of the CCDs (left or right) only extends to the physical limits of that chip and doesn't affect the other.
Click to expand...

Yeah... that would make sense. There are two separate chips aren't they.

Maybe it is nothing to do with the processing, but a flaring/peaking disturbance that effects the whole chip. It wouldn't normally be noticeable, but the abrupt edge of the chip makes it show up.
 
Barry_Green said:
CF24 is just horrid. Spend a little extra and get the Z1, you'll get more features, better audio, and CF25 which is at least sixty times better than CF24.
Click to expand...
Barry:

THANK YOU, for that true-to-form, thorough, fact-filled, Barry_Green-trademark, point-by-point post. VERY much appreciated! I think I see why everyone (that has a Z1) shoots the CF25 mode even when they live in NTSC-land (and then performs that famous "4% slow-down" in post). NOW, I think I get why choosing the Z1 over the FX may be significant (more than just XLRs). Also, is it true that another significant difference between the FX and the Z1 is that the FX does NOT allow the user to employ a 1/48th shutter speed in any CF mode, and that the Z1 DOES allow the user to set the shutter speed to 1/48th in CF24, CF25, and CF30 modes? Again, THANK YOU!
 
Barry:

Regarding the light-sensitivity properties of the Sony . . . where I said in the previous post, "My MAIN criteria in choosing my next camera is LOW LIGHT PERFORMANCE, " and you replied:

Barry_Green said:
Then you don't want the Sony...
Click to expand...
I've simply been inferring from what others have posted--something along the lines of, "The Sony at +3/6/9/12dB was as clean as 'camera X' at 0dB" Can't remember the exact gain values compared in that particular post.

Anyway, my reasoning for concluding that the Sony would be possibly more light sensitive than the JVC (and for that matter, over any other true progressive camera in its class), is that since the Sony is interlaced, I wouldn't lose that extra stop from the progressive read-out, correct? What am I missing here?
 
Barry_Green said:
Interesting that at no point in the discussion do you mention anything about high-def.
Click to expand...
EXACTLY! Either SD or HDV or ProHD whatever . . . I'm just looking for a quieter camera. I thought my problems were solved when Sony invented the DSR450WSL. That's $18K street price with a 20:1 Fuji. Then Sony announces that 1/2" XDCAM-HD camera for $25K MSRP with lens for next spring. $25K MSRP is just too close to $18K, so if I decide to throw that much cash at a camera, I might as well get one that records on optical media and does HDV as well. But next spring is a long time away, and I just want a cleaner camera in the meantime. My criteria for my next camera is as follows:

1. True 16:9 imager.
2. 24P (or CF24/25/30) capability.
3. Low cost.
4. Low light sensitivity.
5. Low noise.

You're probably right when you say the XL2 meets these criteria best. But for some reason, I just can't sell myself on the XL2. I was hoping to get out for only $3K with the FX, but as you pointed out, what I really want is the Z1 (for its CF25 capability). I think the Z1 may be my choice, assuming I can get a confirmation on my assumption about the "interlaced sensitivity advantage" the Z1 may have over the progressive-scan JVC. (Also, keep in mind, I own a ton of FX/Z1-compatible batteries and chargers already.)
 
Barry_Green said:
Don't know. True progressive usually means a stop slower than interlaced, but the FX1 is two stops slower than a DVX, and JVC claims that the HD100 is one stop slower than the DVX. So I don't know how the FX1 could suddenly jump up two stops. The FX1 is really not very sensitive at all; we rated it at around 160 ISO.
Click to expand...
And not to beat a dead horse, but . . . we all know (or should know) that these manufacturer-published sensitivity specs are ARBITRARY, due to the simple fact that the noise level in the picture that they are imaging at what they are calling "0dB" is HIGHLY variable. Therefore, the spec is NOT standardized, and using the spec as a point of comparison is meaningless, since there is no, "standard reference," correct? In a perfect world, manufacturer's sensitivity specs would be indexed to a standard noise level, THEN derived from there, yes?
 
(quotes aren't working for me right now, so I'll have to answer out of context):

I don't know about the 1/48th thing. I would think it would be not possible to do 1/48th in CineFrame 24 or CF30; CF24 is constructed out of 60i, so shooting at less than 1/60th would involve dropping fields, and that would destroy the ability for the camera to construct its CF24 picture (such as it is). Haven't used an FX1/Z1 in a while, so I don't recall if 1/50th is possible in CF25. I would expect that it is. I don't know whether 1/25 would be possible though -- CF25 is technically one field, but one field goes by in 1/50th of a second, so I don't know whether it can successfully generate CF25 from a 1/25 exposure. It could make a field-doubled 25p simulation from 1/25, but I don't know if that would retain the resolution advantage that CF25 has over straight de-interlacing.

As far as sensitivity: at zero gain the camera is a full two stops, maybe more, slower than the PD170. May be two and a half. So you would have to add +12dB of gain to get it to be roughly equivalent. Is +12dB noiseless? Probably not -- +18dB is most definitely not noiseless, it's noticeably noisy. So +3 or +6 is probably pretty safe, but if you're looking for sensitivity, needing to be tied to +12 to get equivalent sensitivity doesn't seem that good of an idea. It's possible that a +12 Z1 is as noisy as a +0 DVX -- and you've already said that the DVX is too noisy for you. Seems like you really need to get some hands-on time (the SonyStyle store, if there's one nearby, would let you use it I bet). I've also seen FX1's at Fry's as well.

Can't answer as to the interlaced vs. progressive sensitivity -- I think you're clinging to an idea that may not pan out. We shot the FX1 against the XL2 and DVX, and in all cases the FX1/interlaced was at least 1 or 2 stops slower than the other cameras when they were in 24p mode.

Regarding $18K for the DSR450 vs. $25K for the XDCAM-HD: also keep in mind that the XDCAM is 1/2" chips, vs. 2/3" for the DSR450. So it's not just a case of paying $7,000 more for optical recording and HD, you're also paying more to get a smaller-chip camera. Don't know if that'll affect your decision, but it's something to keep in mind.

(and, I know you've ruled out the HVX, but let me just say this: if you like low noise, wait until you see the HVX's picture! Even in its pre-production stage, the demo footage is showing some extremely low-light footage and it's clean clean clean clean clean, easily on par with the Z1).

Finally, regarding the nonstandardized spec: you are of course completely correct. I offered it up "for what it's worth", and it's only worth anything insofar as the manufacturers use a common plane of reference. And they don't.

With that said, I'll have an HD100 today (theoretically) so at some point in the next few days I'll be able to explore DVX vs. HD100 sensitivity, including screen grabs and noise levels. Don't have an FX1 on hand though so I can't comment on that, other than to refer back to the side-by-side stuff we did last year.
 
Barry_Green said:
(and, I know you've ruled out the HVX, but let me just say this: if you like low noise, wait until you see the HVX's picture! Even in its pre-production stage, the demo footage is showing some extremely low-light footage and it's clean clean clean clean clean, easily on par with the Z1).
Click to expand...

That would be completely unlike the DVX then... I don't really know - I share NBCshooters sceptisism because I know the DVX. Things like this are just to be said when we have better footage to judge on...
 
It's available to be seen right now at RESfest... next stop is Chicago.

Professionally-shot footage will be available in about six weeks...
 
You must log in or register to reply here.
Back
Top