View Full Version : Dynamic range and problems with red
disjecta
10-11-2005, 02:28 PM
I apologize if this has been asked a million times but in the latest article, there is a mention that the HVX spans at least 7 stops...what's the range of the DVX100a? Just curious to put it in perspective.
Is the problem with shooting reds (notorious in DV for the pulsing effect) eliminated shooting with this camera?
Barry_Green
10-11-2005, 03:30 PM
The problems with reds are due to two things: NTSC in general, and 4:1:1 color sampling. Both issues will be totally eliminated when shooting 4:2:2 HD. However, if converting from HD back to SD, the red issue may rear its head again, 'cause NTSC just doesn't like red.
In my testing I'd place the DVX at around 7 stops.
Knock Out Films
10-11-2005, 03:55 PM
......
In my testing I'd place the DVX at around 7 stops.
The article from your trip to LA says the HVX appears to have at least 7 stops. Do you think that basically it will be the same as the DVX in range, versus a marked improvement (range that is)?
I know it's may be too early to tell, but would you hazard a guess that the DOF will be any shallower? measurably different? on the HVX
Cheers,
Chris
I am not as bothered about lab style latitude as I am about highlight handling. I hope it is good at that.
Knock Out Films
10-11-2005, 04:40 PM
I am not as bothered about lab style latitude as I am about highlight handling. I hope it is good at that.
do you mean on the potetial blow-outs and bright whites? i hate my DVX when you get into a tough situation with highlights, the whites start to look a bit purple-ish
i know it's all about lighting, but in the real-world it can be tough
cheers,
chris
do you mean on the potetial blow-outs and bright whites? i hate my DVX when you get into a tough situation with highlights, the whites start to look a bit purple-ish
i know it's all about lighting, but in the real-world it can be tough
cheers,
chris
Yeah, kinda.
One of the big giveaways of video is that highlights tend to clip and blow-out rather than roll off nicely like with film. I don't like that look much.
Barry_Green
10-11-2005, 05:12 PM
The article from your trip to LA says the HVX appears to have at least 7 stops. Do you think that basically it will be the same as the DVX in range, versus a marked improvement (range that is)?
What the article says is that Panasonic claims that they've measured 7 stops. Jarred and I haven't had a chance to put an HVX on the test bench and verify that.
At 7 stops, it would about match the DVX. We had been expecting significantly narrower latitude from these HD cameras (as exemplified by the JVC HD1/HD10 and the Sony FX1) but the newest ones are turning out to be surprisingly comparable to the equivalent SD cameras -- the JVC HD100 looks like it can deliver 6 to 7 stops, and Panasonic's claiming 7 stops. That is unexpected good news!
What is the limitation and/or reason for these cameras blowing out like they do? Isn't there a way they can roll off the highlights better? Or is it inherent to either latitude or digital sensors in general?
Knock Out Films
10-11-2005, 06:37 PM
... We had been expecting significantly narrower latitude from these HD cameras ........That is unexpected good news!
I somewhat understand that the more pixels you cram onto a CCD can narrow the range, but I can't honestly see how these cameras are worth squat if the range is even less than what we have been getting out of the DVX.
I love the idea of the HVX, I really want it to work for me, but seriously, how can we be excited about a camera or format as it is delivered, on a specific size CCD if it doesn't have better range or shallower DOF.
I must admit that until your statement about expecting lower range from these various HD cameras, I had no idea it could be less than my DVX. I had assumed they would figure out a way to advance, (looks like they at least got the HVX on par with the DVX, thank god)
Narrower latitude is the last thing we can live with! I am going on a bit here, but this stuns me. Barry, where you really ready to live with the possible narrower range? Am I missing something? is this the cost of getting HD at this price point? does the JVC have narrower range than the DVX?
Holy crap!
I have always said that the irony of HD is that most of our projects will go to SD. But I love the idea of better colour depth, tapeless recording and 60p.
I guess I should be happy that it appears the HVX may have range equal to my DVX.(can't believe I'm saying that :) )
Cheers,
Chris
Barry_Green
10-11-2005, 07:24 PM
Seems like your expectations may have been a bit high, yes.
The JVC HD1 delivers maybe four stops of total dynamic range. The Sony Z1 is two stops slower than a PD170, and has maybe 5.5 stops of total latitude.
There is no free lunch; giving with one hand (more resolution) has traditionally meant taking away with the other hand (less latitude and sensitivity).
Expecting there to be a difference in DOF is, frankly, naivete. The DOF is established by the lens, not whether something is high-def or not, and it's not anything anyone can do anything about. The longer the lens, the more open the iris, the shallower the DOF. The HVX will be capable of shallower DOF than the DVX was, but only because the HVX's lens is 10mm longer than the DVX was. The HD100 is capable of significantly shallower DOF because its lens goes to 88mm (although the chromatic aberrations become extreme and the edge sharpness drops off perilously at that length, the DOF is definitely nice and shallow). And the XL H1 is capable of significantly shallower DOF still, as its lens goes to 110mm. Whether you'll find it practical to try to frame a headshot at 110mm is another matter entirely.
Here's a reality check: high-def progressive-scan costs $70,000 to $100,000. That's what it costs. Here we're being offered high-def progressive scan for less than 1/10th the price -- did you think it would come with no tradeoffs?
Of course there are compromises involved. JVC seems to be taking the approach that low color sampling, a split-screen effect and substantial chromatic aberration and compressed audio are a reasonable compromise, the price you have to pay for getting a high-def camera for only $5500. Lots of us would disagree with that, but there it is, it's available. If you can make it work for you, you can be shooting high-def for less than 1/10 of what it used to cost. If you don't like those tradeoffs, there's the forthcoming GY-HD7000U at around $28,000 (plus lens).
Sony seems to be taking the approach that MPEG-2 artifacts, low sensitivity, low color sampling, compressed audio, narrow latitude and interlace-only acquisition is the price you have to pay in order to get their high-def cameras. But, they're asking as little as $3300 for a 3-chip, and $1600 for a one-CMOS cam! Are the tradeoffs worth it? You, the customer, get to decide -- but you have to be very aware that the tradeoffs are there. If you don't feel like making those kinds of tradeoffs, they offer XDCAM-HD at around $25,000 (plus lens).
If the HVX were able to match the sensitivity and latitude of the DVX, but do so at high-def resolution, it would be an unprecedented breakthrough. I don't expect it to happen. I don't think anyone with reasonable expectations expects that to happen. We may want it, sure, but hopes & wishes vs. reasonable expectations are not the same thing.
thisiswells
10-11-2005, 07:36 PM
I am not as bothered about lab style latitude as I am about highlight handling. I hope it is good at that.
mmm- maybe this is interesting ?
--------------------------------------------------------------------
The AVERAGE scene brightness range is 160:1 (2.2 Log Exposure units), or a little over 7 stops. Modern color negative films have a "straight line" sensitometric curve well over this range, and are able to capture information over an even wider range.
Kodak published each film's characteristic curve in the technical data for each product.
Here's a link to a good tutorial:
http://www.kodak.com/US/en/motion/support/h1/structure.shtml
http://www.cinematography.com/forum2004/index.php?showtopic=9364&st=0&p=70497&#entry70497
--------------------------------------------------------------------
Jan_Crittenden
10-11-2005, 08:10 PM
The article from your trip to LA says the HVX appears to have at least 7 stops. Do you think that basically it will be the same as the DVX in range, versus a marked improvement (range that is)?
Chris
The range or latitude will be similar, 7 stops, but the mid point changes. F8 at 2000 lux for the HVX, F11 at 2000Lux for the DVX, F5.6 for the Z1. So faster than the Z1, slower than the DVX but 7 stops.
Does that help?
Jan
sukram
10-11-2005, 08:50 PM
The DOF is established by the lens, not whether something is high-def or not, and it's not anything anyone can do anything about. The longer the lens, the more open the iris, the shallower the DOF. The HVX will be capable of shallower DOF than the DVX was, but only because the HVX's lens is 10mm longer than the DVX was......
To add to Barry's explanation, DOF is really a combination of 2 things.
(1) the lens, as described.
(2) the size of the negative/chip (which ends up being a function of the lens, or vice versa)
The reason behind both has to do with the "circle of confusion." Read about it in any camera assistant book. The smaller the negative/chip, the greater the depth of field. Therefore, in decreasing DOF assuming constant aperatue, format (i.e not anamorphic), and focal length:
1/3" chip, 8mm negative, 2/3" chip, 16mm negative, 35mm negative.
(Of course, a 50mm lens on a 16mm camera is equivalent to a 75mm lens on 35mm film)
The point is this: using similar lenses, you stick the best 2/3" chip into a $100,000 camera, you still ain't going to get the depth of field as on a $10,000 50-year old eyemo camera - although your image might be a bit shaky. :grin:
:beer: I shouldn't write and drink at the same time. This cursory explanation really belongs on a cinematography board. Get a good camera assistant book, or better yet a camera assistant with some patience (nevermind), and learn about how cameras function. For what this camera delivers, 4:2:2 and DVCPRO-HD (i.e. not HDV with motion artifacts), it still to me seems like an amazing deal.
markus
mmm- maybe this is interesting ?
Thanks for the link :thumbsup:
Landon D. Parks
10-12-2005, 03:55 AM
Throw a Mini35 on it, and you know what, 35mm DOF! I hear Mini35 for this should be out around the first of next year, so we can have our cake and get shallower DOF too.
Knock Out Films
10-12-2005, 05:16 AM
Here's a reality check: high-def progressive-scan costs $70,000 to $100,000. That's what it costs. Here we're being offered high-def progressive scan for less than 1/10th the price -- did you think it would come with no tradeoffs?
If the HVX were able to match the sensitivity and latitude of the DVX, but do so at high-def resolution, it would be an unprecedented breakthrough. I don't expect it to happen. I don't think anyone with reasonable expectations expects that to happen. We may want it, sure, but hopes & wishes vs. reasonable expectations are not the same thing.
Barry, I didn't expect the HVX camera to be the same as a varicam, I didn't say that. I am fully aware of what a varicam is worth. No reality check needed.
I just expected that the HVX would perform as well as the DVX we have all been shooting on already! I am willing to except tradeoffs. I didn't think the tradeoffs could have been that it is worse than a DVX in range. I am just a little surprised.
Cheers,
Chris
taubkin
10-12-2005, 08:26 AM
If the HVX is cleaner in the darks, it can actually have more useable latitude than the DVX!
When people say 7 stops, they ususally mean, from black to blowout, but without realizing that you can't have much stuff in your image one or two stops over black. That is too grainy and ugly looking. You can't have stuff one stop or two under the highlight, because that looks bad too. So the 7 stops quickly become 4-3. Maybe I'm exaggerating, but it does happen. With a cleaner camera, you can push that useable latitude one or two stops and that can make a lot of difference!
What I don't get is that I thought the DVX lost latitude in the DV compression. I mean, that's what the andromeda claims! But if the HVX will compress straught to DVCPRo50 or HD, shouldn't it have a wider dynamc range?
If the HVX is cleaner in the darks, it can actually have more useable latitude than the DVX!
When people say 7 stops, they ususally mean, from black to blowout, but without realizing that you can't have much stuff in your image one or two stops over black. That is too grainy and ugly looking. You can't have stuff one stop or two under the highlight, because that looks bad too. So the 7 stops quickly become 4-3. Maybe I'm exaggerating, but it does happen. With a cleaner camera, you can push that useable latitude one or two stops and that can make a lot of difference!
Same thing with sensitivity too! The HVX may well give better low light performance than the DVX because using gain will not cause unacceptable noise so easily.... well that is if it is as noiseless as everyone says!
taubkin
10-12-2005, 01:16 PM
Yeah, but KHMuse already gave us a class about to really be better than the DVX in low light, the HVX would need to have a 18dB gain that was unnoticeable, so maybe that would be not the case. But truth, they could be closer than they seem...