Jvc gy-l300

philiplipetz said:
Skin highlight blowout when hand catches the light. For example :20. DR questions.
Click to expand...
I saw that too. I noticed it in the 'glass blowing' demo as well.

I often wonder; Is highlight-clipping inherent to a particular sensor or is it the processing from the sensor?

I don't think it's lack of dynamic range. JVC/Altasens claims approximately 11.3 stops. That's close to Canon's C300 at 11.4 and better than the GH4's 10.9 (according to one source)

If it's due to how the signal is being processed, maybe this is something JVC can tame in the next few weeks(?)
 
TFilipiak said:
I saw that too. I noticed it in the 'glass blowing' demo as well.

I often wonder; Is highlight-clipping inherent to a particular sensor or is it the processing from the sensor?

I don't think it's lack of dynamic range. JVC/Altasens claims approximately 11.3 stops. That's close to Canon's C300 at 11.4 and better than the GH4's 10.9 (according to one source)

If it's due to how the signal is being processed, maybe this is something JVC can tame in the next few weeks(?)
Click to expand...

the highlights were already present in that JVC Kill Bill style promo clip as well. I'd like to see some ISO range tests of the same image until the white clips.
 
TFilipiak said:
I saw that too. I noticed it in the 'glass blowing' demo as well.

I often wonder; Is highlight-clipping inherent to a particular sensor or is it the processing from the sensor?

I don't think it's lack of dynamic range. JVC/Altasens claims approximately 11.3 stops. That's close to Canon's C300 at 11.4 and better than the GH4's 10.9 (according to one source)

If it's due to how the signal is being processed, maybe this is something JVC can tame in the next few weeks(?)
Click to expand...


This is an excellent point. I believe that it is NOT the sensor, but something else that JVC is doing in the processing. I say this from my past experience with JVC cameras, which exhibit the same type of highlight limitations. I have a HM150 (I know it is not in the same league as the LS300) but it exhibits this identical problem. I bet if one looks at the other two cams in the soon-to-be released 4K line, the HM200 and 170, one would find it lacking in dynamic range just like the LS300.
 
TFilipiak said:
Is highlight-clipping inherent to a particular sensor or is it the processing from the sensor?
Click to expand...

Highlight clipping is inherent to all electronic sensors. The photodiodes really are like buckets. The photons come in, and they are translated into electrons. And the electrons sit in the bucket. (This is within the duration of a single frame. At the end of each frame, the buckets are emptied.) Some buckets have bright parts of the scene aimed at them, and they take more electrons than they can hold. They max out, abruptly. This is what we perceive as clipping.

Film also clips, but at the opposite end, in the blacks. We may yearn for the famous highlight roll-off of film, but it is little talked about that if you strongly underexpose film, the blacks are clipped. This is because of the way film works. When light strikes a film grain, that film grain becomes "exposed," and basically impervious to more exposure. As more photons pour in, then, there are fewer and fewer grains that can react to them. Like throwing darts at balloons on a wall. After a while the balloons pop, and the same amount of darts pop fewer and fewer balloons. This is why film highlights "slow down" in their roll off. They ease off. The number of photons hitting do not exactly equal the number of grains exposing anymore. Video chips clip precisely because they don't slow down. The number of photons hitting the sensor exactly equal the number of electrons, until suddenly the photodiode can't take it anymore, is full, and it clips. (I don't mean one photon equals one electron. Sensors differ. It may take four photons to equal one electron. But it stays at that ratio. Okay, the tech nerd voices in my head compel to again qualify that statement by saying that it may not always be exactly four photons. Some times maybe three, sometimes five. But the average in this example is four. It does not, in any case, roll down, like film does).

Anyway, what was I saying? Oh yeah, film clips, just like video, but at the opposite end, in the blacks. If too few photons hit an area of film, it stays unexposed. When it's developed in the lab, the unexposed areas are scraped off, and the film is clear at that point. No texture. When the film is clear, then a light from a projector will shine right through it, making an area that's solid white. But this isn't highlight clipping. This is a film negative. Black is white, and white is black. When transferred to a release print, the solid white becomes solid black. Hence, film clips in the blacks, and you need to be careful not to underexpose it too much, just as with video you need to be careful not to overexpose it too much.

Now what about video that has smooth highlight roll-off? Well, based on what I've said, it has to be done in post. Either "post" in the camera, right after reading off the sensor but before recording to media, or later in the editing suite, during color correction. You basically are artificially raising some of the bright-but-not-clipped parts to be a little brighter, and the parts just below them to be a little brighter than they really were, and the parts below them to be a little brighter than they really were, until you get down to the middle, to where you straighten out. But basically, yes, to roll off the highlights, you basically have to "eat" some of the almost-highlights, by brightening them, so that the transition from almost-blown-out to totally-blown-out is not sudden but smooth. So it pays to underexpose your video by a stop or two, because in post you are going to stretch the mids and mid-highs upward, changing mids into mid-highs, and mid-highs into highlights.
 
Last edited:
combatentropy said:
So it pays to underexpose your video by a stop or two, because in post you are going to stretch the mids and mid-highs upward, changing mids into mid-highs, and mid-highs into highlights.
Click to expand...

I work in a similar fashion. If I'm not working in RAW and lets say exterior wide shooting I expose for the whites but not clip. As we all know thr overall image still maintains its organic look. Then in post i boost lows and mids to mid highs. the more data in the codec the better the result. the Whiskey video states 1600, but I thought the LS300 goes by db increments?
 
combatentropy said:
Thank you, Jim. Fire your shooter. Take the camera, go out on the street at high noon, and record the people walking.

Test shoots should always be done of people, in high-dynamic-range situations.

People. Shoot people, not landscapes or still life, because we are good at judging what people look like. From Paul Graham:

In art, the highest place has traditionally been given to paintings of people. There is something to this tradition, and not just because pictures of faces get to press buttons in our brains that other pictures don't. We are so good at looking at faces that we force anyone who draws them to work hard to satisfy us. If you draw a tree and you change the angle of a branch five degrees, no one will know. When you change the angle of someone's eye five degrees, people notice.​


Retranslated into camera tests:

In camera tests, the highest place needs to be given to shots of people. There is something to this need, and not just because pictures of faces get to press buttons in our brains that other pictures don't. We are so good at looking at faces that we force any camera that records them to work hard to satisfy us. If you shoot a tree and you change the colors five degrees, no one will know. When you change the colors of someone's face five degrees, people notice.​


High Dynamic Range. Cameras these days can look just like each other under controlled circumstances. It's in broad daylight that their flaws are shown.
Click to expand...
We did not shoot this footage so I'm not going to fire our shooter......

Jim Martin
 
Thanks for sharing the live streaming event, Jim!
Perhaps that will help potential customers decide if the LS300 is the camera for them!
 
We did not record it (legal issues) so sorry about missing out on the stream.....for all, the wrap up was that people had a good time, learned a lot about the camera, and we ended up playing back the 4K footage multiple times at the end so people could really check it out on the JVC 84".....once the cameras arrive (last week of this month) and for those in town, we will have a free 'try and buy" program where you can rent the camera for 1 day (with deposit or insurance)....that way, you can put the camera through your own "paces", edit the footage, and make a qualified decision.

Jim Martin
 
Any words of encouragement from participants? We need definitive answers!!! Would be nice to get a quick review from someone that attended the event or watched on Ustream. Also, where's the JVC LS300 official video?
 
Last edited:
I'm working with JVC on getting the footage up as soon as possible.... so you can see it yourself! Again, DR looks good & colors look good (both subdued & standard Cinema Gamma). The reaction from the attendees was very favorable in regards to looking at the footage on the 84" 4K JVC monitor.

Jim Martin
 
You must log in or register to reply here.
Back
Top