Latitude
 

Ok, the big selling point of this camera to me is supposed to be it's wide latitude as compared to other 1/3 or 1/4 CCD cameras. Anyone who has ever used a 2/3 CCD camera knows that 1/3 camera's blow out highlights, and have deep shadows all over the place. Bring on the V1...but just how much extra latitude does it have, and how much of it is processor trickery?

In all the reviews I've read, Nigel Cooper, DSE and Adam Wilt, none have mentioned anything about superior latitude. Steve Mullen has written an article about the V1's apparent superior latitude without giving any comparison to other cameras. He also suggests its latitude could be down to processing. This is intersting because the HVX has a cinegamma mode which also gives better latitude, although other parts of the picture are lost. Is this what the V1 is doing?

Just how much better latitude does the V1 have, if any?. Is it as good for example as a 1/2 CCD, or maybe even a 2/3 CCD camera?

Re-read my posts and my quick looks at the camera. I mention latitude in a couple of para's, and have several shots where I captured a fast-moving dark image flying directly into the sun, and held it on the sun to see where it would wash out.

I thought Steve Mullen's article compared the V1's latitude to the Z1 and perhaps also the HC1 ... can't recall where the article is though.

Yes I think I read this. But my question is just how much latitude?? - as compared to 1/2 or 2/3 cameras for example? I am sorry but cant' think of another way to measure this, I am not a techie. And just how much of this is due to internal processing like the HVX's cinegamma modes which also appear to increase the latitude?

Check out wolfgang's site and decide for yourself:

http://www.fxsupport.de/12.html

Over there are a lot of grabs fx-1 vs fx-7, and since the fx-1 is supposed to be the lower-latitude camera, you can see if it holds its claim...

Wolfgang did a fantastic job comparing these two cameras. My guess is that the FX7 has at least one f-stop, but less than two, of latitude greater than the FX1. It seems that Wofgang exposed for the hightlights and kept the sky evenly exposed between the two cameras (as it should be done). It isn't easy to know exactly the difference. You must look at the shadow areas and the foreground instead of the sky. In the FX7, the grass and other things in the foreground is clearly brighter than the FX1. In those areas, the color is much more pronounced.

What I noticed the most about the comparison is the colors sometimes seem richer in the FX7 and there is also a clear resolution advantage to the new camera. There is also less noise in daylight shots with no gain on either camera. This was noticed several feet away from the screen by my roommate.

The FX1 is definitely brighter in low light, but the FX7 still looks decent. There is less noise at 18db of gain, but the picture is darker.

I still like both cameras, but the FX7 is better in most comparisons.

All this changes a bit for the V1. It has black stretch and knee control, so a bit more apparent latitude may be possible.

Is it me or does the FX7 show noticeably less noise than the HVX200 and the Canon A1, in general? I wonder how the V1 will compare when it's all said and done.

since latitude is just the range available between the darkest and the lightest part of a picture, i do not really see how the same technology (HDV) could produce a different latitude.
if you got 8 bits to express a range, there is nothing you can do increase the numbers of steps you got (255 values).
There are some trick to crush black or white and give illusion of greater latitude, but nothing you can not do by setting a proper exposure (using histogram from DVRack for example) or correcting in post.
Obviously this is content dependent, and you will need to trade black or white (or both) regarding what is on the picture and what is your target.
My experience shows that most of time , at least 25% of the exposure range is lost by either underexposure or overexposure, so there is room for improvement just by shooting correctly.

All good points, but both the FX7 and V1U have CMOS chips. CMOS provides better latitude than CCD. I know that first hand. For years I have shot with CCD camcorders until I purchased my very first CMOS camcorder. The differences can be easily detected, even by untrained eyes.

Giroud,

If a CCD is capable of differentiating more lattitude, then it will get downconverted into more in HDV. HDV is a format. but not all media that uses it will have equal sharpness, dynamica range and color. It is a codec that is pretty flexible. It essentially is just Mpeg-2. All of the DVD's you watch in SD are Mpeg-2 and have the ability to retain a fine degree of latitude.

From a casual observation, I would not say that the FX-7 has a higher latitude than the Canon A1. The FX-7 has a tendency to lose details in the shadows, for example.

Of course, comparing one to the other is difficult because the contrast is different.

"since latitude is just the range available between the darkest and the lightest part of a picture, i do not really see how the same technology (HDV) could produce a different latitude."

Look at a DVD professionally produced from a big-budget motion picture. It shows much greater latitude than video from a camcorder. DVD mpeg2 doesn't have greater latitude than DV or HDV, but the big-budget movie looks better. It has nothing to do with the delivery format. It is due to expert color correction that brought the highlights and shadows into the MPEG2 range and the originating media's ability to discern detail in both the shadow and highlights. If the imager is physically more resistant to overexposure in the highlights, it will look better. If it can pick up detail in the shadows before the highlights are clipped, it will look better. Different technologies have different physical properties and this is evident in the reviews done by Wofgang. CCD and CMOS gather photons and relay their signal a bit differently.

quote:
"Look at a DVD professionally produced from a big-budget motion picture..."

You are right, professional DVD made out of professional movie look better.
As i said, it is just the proof of clever use of the latitude given by mpeg2 at 8 bit RGB. It does not change the fact that the range is given and fixed to the one defined by the standard used.
That is why i maintain that good shooting , will give you an increase in quality of several magnitude than than the thin difference you could get between a CCD and a CMOS.

You bring up a good point. Even though the V1 appears to have significantly greater latitude than other HDV cameras, that won't make any difference in the hands of someone who doesn't know how to shoot decently.
Even output from a FW950 will look awful from someone who doesn't know what they are doing, and a monster with a camera can take a PD150 and do wonderful things.

Latitude is not something that is going to be lightyears ahead in terms of quality. It is better but it isn't the sort of thing that is going to knock somebody on the seat of their pants. It is one of those subtle enhancments that give the camera a slight different look compared to other HDV cameras but thats about it.

Good cinematography and knowing the limits of a CCD based camera will give you better images then just a decent shooter with a camera with a little bit more latitude.

Greater latitude is not the perfect cure for bad and sloppy shooting and exposure. It can help but there are a lot of other things that can help as well.

Lets face it we are still in the age of tradeoffs for these small HD cameras. If latitude is the most important aspect of the image to you and you do not mind sacrificing other elements of the image then the V1 should be a great camera. The V1 is still tied with the Canon A1/G1 with me at this stage for my next HD camera. I will not decide however until I see more footage and get a chance to rent the camera and try it out. I already know I love the look of the Canon due to the countless dozens of sample clips I have seen in various shooting conditions. There are however some features I love from the V1 such as dedicated hard drive recorder, HDMI, CMOS. Latitude while nice is not a huge concern for me because I tend to take great care in exposing my shots and in fact I usually prefer more contrasty scenes. Any extra latitude would probably be killed by the time I was done color correcting the footage.

"that won't make any difference in the hands of someone who doesn't know how to shoot decently."

"Greater latitude is not the perfect cure for bad and sloppy shooting and exposure. "

Come on, guys. I always assume someone knows how to use a camera on these forums unless they state otherwise. Yes, every single one of us can improve our own skills, but I think we have established that the baseline on this forum is above "tourist".

In the hands of an expert or even a knowledgeable enthusiast, a single f-stop of exposure can make a big difference in many situations. It means areas that would have washed out may now have enough color to grab hold of when doing color correction. It means that poorly-lit areas of a shot will have enough color to look lively. Anything that improves a camera's abilitly to retain proper exposure across the image is a very good thing. Couple an extra f-stop of latitude along with a good shooter and color correction and the results will definitely be far and above the average snapshot from a tourist.

Another way that greater latitude will be helpful is in reducing the cost of lighting. I mean this in the entire sense that particularly includes the time to set up and transport equipment. It can take a lot of equipment to bring a scene one f-stop narrower. There is no substitute for good lighting, but a better camera could make an otherwise difficult shot feasible.

Personally, I am looking forward to my first bright-blue sky background with the subjects lit by a single white reflector as fill. I'm sure the talent will also be happier. I did an acting part that required a silver reflector in my face. My eyes watered and I barely could get through a 20-second take.

The first reason is that the sensors (CCD or CMOS) are digitized at 10 to 12 bits, then the camera maps the sensor values to the 8 bits of the HDV format. This mapping is not linear and basically, depending on the curve used, one can trade more shadow or highlight detail for more or less contrast.

The second reason is that, even if the sensor is digitized at, say, 12 bits, the sensor may not have a linear response itself, some values may be all the same because of noise, etc... Noise is usually the biggest problem: if you expose so that highlights are not saturated, all shadows may look the same because they are all below the noise floor.

I found the proof! :)
 

Go to Wofgang's comparison of the FX1 and FX7:

http://www.fxsupport.de/12.html

Open up both the FX1 and FX7 copies of Image 13 (13fx1.jpg and 13fx7.jpg). Make sure that they are full resolution in your browser windows or viewing program. In Firefox, I have to click on the pictures with the magnifying glass. I clicked back-and-forth several times comparing the image. I noticed that the sky in the FX7 image has a richer, darker blue at the top of the image. At the same time, the green foliage and the rust-colored seawall have brighter color.

Here's the real proof: Look at the metal smokestack (not the red and white one) to the right of the middle of the pictures. It is the tallest object in the right side of the frame. First look at the FX1 image. The stack appears as charcoal grey on the shadow side and is pure white and probably clipping/smearing on the highlight side. It almost looks like a black strip with a white stripe going down the middle. Now, click on the FX7 image13. You should immediately realize that the shadow side is about an f-stop brighter and the highlight is about an f-stop darker than the FX1 shot. There is also a transitional mid-grey color between the highlight and shadow sides. This could mean that the FX7 approaches two f-stops of greater latitude since both the highs and lows are noticeably darker and brighter in the favorable directions.

There are different angles of this same shot that seem to correlate my observations.

It is also very obvious that the FX7 has better resolution than the FX1. Look how much "fuzzier" the trees look in the FX1 images compared to the FX7 shots. I was astonished when I first compared the FX1 to my VX2000 and to think that the V1 will have even more image controls than the FX7 really has me excited.

I'm not criticizing the FX1. I have used it many times and I really like that camera. The fact that the FX7 images are clearly better (although a bit of apples and oranges comparison) is not an insult to the FX1, but rather a testament to Sony's continuing success with CMOS imagers.

I am also not against Canon. I think the A1/G1 look like a great overall package that delivers a very nice image. Assuming they are both reliable machines, it really seems that you can't go wrong with the V1 or it's Canon rival the A1. Choose the one with the features that suit your shooting style. I just really think this is a great time to be in the market for an HD camera!

Aloha and Mahalo (thanks) again to Wolfgang for doing such great work.

The most obvious factors between the FX1 and the FX7 for me are;

Richer, more neutral colours on FX7
More contrasty images that retain an excellent dynamic range.
Even less noise in well lit conditions on FX7
Much less edge sharpening on FX7
Much more resolution on FX7.

Not to mention the "certain something" that makes the 2d images of video look more 3d.

Check out Wolfgang's video of the supermarket (FX7_02.m2t), the older gentleman with the brown hat seems to jump off the screen compared to the FX1 footage.

I'm sold on a V1.

TT

I see no big DR improvement
 

Sorry Marcus, I see the noticeable increase in resolved detail and higher saturation w/the new camera, but basically every highlight blown in the FX1 shots are still blown in the FX7.

The shot you described, plus shot #8 (the lock and dam, which looks like a better test of light-dark range) both show only Very slight differences in range, and exposure appears set to the same on both cameras. I'd say maybe a third stop difference (maybe) in highlights retained, not 2 stops or anywhere near that. Perhaps it's just that in the jpgs, the fx7s higher detail becomes more evident in those parts of the image that are over -or under- exposed.

Still, it does make for a better FX7 picture overall, but I certainly don't see any giant step forward in DR here. Would have been interesting if he had an A1/G1 shooting at the same time. I suspect those two could be set up to match pretty closely in the daylight shots.

Also its absence of smear is really nice to Not see in the night shots w/the light fixture, but again in those shots, no real big improvement of highlight handling, other than that due to 1 stop lower exposure of the FX7 vs the FX1 in some of them.

ps- I'm not trying to cut down the Sony, I like the smaller size, and its long lens reach w/such close focusing (and f/2.8 all the way out). But was hoping that a real advance in DR would make up -it could more than make up- for its lower sensitivity and less-wide lens.

Okay, Vlad, I'm going to partly conceed your point. Then, I'm going to disagree so my ego isn't bruised. Thanks to you, I looked at shot #8 in detail. I'll agree that the same highlights are blown to a fair extent, but there is often just a shred of color left in the FX7 shots. I kept switching back and forth and still see the crushed almost-black shadows and colorless highlights in the FX1 images.

Unfortunately for my ego, part of this is due to sharpening in the FX1 and a lack of smear in the FX7 instead of greater latitude. I kept looking at all the poles and handrails in the lock photos (shot #8) and realized that the FX1 is sharpening the heck out of small high-contrast areas. There is that sort of "black glow" around the objects in the FX1 that I really hated in my VX2000. This is a typical Sony look that enhances resolution. It also screams "This was shot on video!" in my visual cortex.

I still think the better color while retaining overall proper exposure is due to higher latitude. This could allow a slightly reduced exposure to preserve some highlight detail. I don't mind if shadows crush a bit as long as color doesn't wash out in the entire image. I also think the highlights look much better whether it is due to sharpening and smear or latitude.

I really think the FX7 looks significantly better than the FX1 and I've said before that I like the FX1. The FX7 just seems more transparent in it's image than the FX1. I don't expect the camera to get the shots for me, I just don't want it's flaws to get in my way and this seems like the first video camera that may be able to get the job done (as long as there is enough light).

Oh, and Sony has done a great job at getting rid of that "cool" tone to their images. The FX1 started the trend and the FX7 is infinitely "warmer" than the FX1 without losing the blues and greens. The FX7 actually looks like video shot during "golden hour". They have also enhanced saturation without ruining detail. Don't forget that these two things are often mutually exclusive.

I think it is important to understand that from the purely technical point of view, increased latitude is impossible between FX1 and FX7 since they use same video HDV specification
the exemple you gave are just showing that each camera handle high, middle and low light (and eventaully saturation) a different way.
basically if you see more values into the middle range, since there is only 255 values to distribute, there will be less value to give elsewhere.
it is exactly the same problem you see on LCD screen.
Some are good in the blacks but burns the picture in the whites.
some others , it is the reverse.
That means that some cameras are smarter at handling light in automatic mode than others, and can produce better looking pictures (like the FX7 vs FX1). But as you stated earlier, we can expect that the skill level of readers on this forum let us expect that they do not rely exclusively on the automatic settings stored in their camera to shoot properly.
If we really want to compare latitude, lets publish pictures and histograms so we can really compare.

I think it is important to understand that from the purely technical point of view, increased latitude is impossible between FX1 and FX7 since they use same video HDV specification
the exemple you gave are just showing that each camera handle high, middle and low light (and eventaully saturation) a different way.
basically if you see more values into the middle range, since there is only 255 values to distribute, there will be less value to give elsewhere.
it is exactly the same problem you see on LCD screen.
Some are good in the blacks but burns the picture in the whites.
some others , it is the reverse.
That means that some cameras are smarter at handling light in automatic mode than others, and can produce better looking pictures (like the FX7 vs FX1). But as you stated earlier, we can expect that the skill level of readers on this forum let us expect that they do not rely exclusively on the automatic settings stored in their camera to shoot properly.
If we really want to compare latitude, lets publish pictures and histograms so we can really compare.

I agree with you, the fx-7 grabs look definitely nicer (sharper and color-wise), but I'm not seeing the increased DR either. But maybe under more controlled circumstances and/or manual tweaks on the camera it can be achieved.

And what to you guys mean with the '3d effect'? I don't have a HD-tv so I don't have proper means to view the footage other than on my computer monitor, but I was curious what you guys are seeing exactly...

the 3d effect is easy to explain.
Usually, SD video shows nice subject, but background is fuzzy, because the lack of resolution.
with HD, everything can be shap and detailed, so your eyes and your brain can see the leaf in the trees in the background (instead a green mud on SD).
so your brain can focus on the different layers foreground-background and that gives a real impress of 3d or reality (your mind can travel in the depth of the picture, because the eye can)
on hdv, it disappear almost instantly when the camera is moving , because high compression. you can do the experience by yourself, by shooting somebody on grass and do a pano. the grass goes sharp, blurred, sharp as the camera start to pan and stop (you can even see the gop, since the blur is slightly delayed).

Marcus, could you elaborate on the above. As an owner of the VX2000 and seriously contemplating replacing it with the FX7, I'd love to hear your thoughts. I know we'll give up some low light capability of which the VX2000/2100 is in a class by itself (which I hope the new Sony LED video light will help recapture on the FX7 when and if needed), but what were your comparisons?

actually, bit-depth is not the same thing as latitude. the range between the lightest and darkest values is independent of the number of steps between them.

on the encoding end, latitude is kinda like how high you can reach with a staircase. and bit depth is the number of stairs on the staircase.

i can build a 3-story staircase with 253 steps, or make 3-foot model staircase with the same. a "3-foot model staircase"-type sensor system will capture much less of the real world than the "3-story staircase." but the number of steps, the gradations are the same between both. film converted to mpeg2 looks fantastic because this is a system in which the stairs reach pretty high. a handicam converted to mpeg2 looks contrasty and video-ish, because these stairs are low. both have the same number of steps.

on the decoding end, you still see the "height" of the stairs as always being the same because your display hardware doesn't "know" how high the original stairs went. so for example, a given pixel doesn't get any brighter when given an intensity of 253, whether it was produced from a film scanner (a very high staircase), or from a cheap handicam (a very low staircase). in both cases, you get the same number of "stairs."

i'm thinking there are probably a few hundred posts on this subject already (and 50+ years of textbooks on dsp theory), so maybe i'm missing the point of the previous post. forgive me and ignore this if that is the case!

ah i DID misunderstand the post! my apologies. you're saying the distribution of values is different between the two cameras, but that the high and low points are the same?

yes that should be the case with video.
values are relative to a standard video signal that range from IRE (0) to max level (usually 100%) or you could be in serious trouble playing video from different origin.
so whatever you do, the latitude is fixed in the luminance (Y) encoding and this luminence is on 8 bit in HDV (255 steps) while lower and higher value should be the same for everybody . and if not the same, the distance between low and high (latitude) should be the same at least, so you can slide whole range to adjust it to a known standard.(usually we call that the brightness button)
Now , if your camera is clever, it can read an histogram of the shoot, and distribute this value in a non-linear way to allow a dark picture to assign more level to dark.(called Knee on professional camera).
By analogy the contrast button, would be the space between each step, but knowing that increasing space between each step would clip the value going out of range at both end of the scale.
The only way to get a better latitude would be to correct the 8 bit signal with a 10 or 12 bit codec (like cineform is able to do) so overflow value would not be clipped.
it is good to think at luminance like an histogram.
slide the range to left, it goes darker. to right, it goes lighter.
stretch the range to left, you got better definition in blacks, stretch to the right, you got better whites.

"Marcus, could you elaborate on the above. As an owner of the VX2000 and seriously contemplating replacing it with the FX7"

This is getting away from "latitude", but here goes:

The FX1 has so many more control options better control methods that it is nicer to use. The ergonomics and options alone make it worth the upgrade. Of course, for event videos where no supplemental light is possible the VX/PD cameras are king. Otherwise, the HDV image quality and ease of getting a good image are much better than the VX2000. The FX1 really seems to have better color and a bit less of the "video look". It also looks "warmer" than the VX without losing other color frequencies. The previous Sony "look" was beyond just white balance differences. The overall image did not seem to retain red frequencies without losing blue. The reflective LCD is also very nice and I did prefer the higher and more forward mount on the FX1. At least the FX7 seems to have the reflective screen.

The FX7 shots really look bette to my eyes as they both increase color saturation and resolution. Increasing saturation on other cameras I have used simply made the colors look bad and the detail diminish. An improvement in both of these qualities simultaneously is a real achievement in my mind. I really think the V1 will be worth the extra money (if you can afford it) as it has more image control options and progressive scan. Sony put more significant differentiation between the V1 and FX7 as there is with the FX1 and Z1.

I can't wait to get the V1. I'll determine what seems to improve the image myself and report back in a few weeks. I can't wait to play with the color adjustment program.

AFAIK, all digital video cameras do exactly that: they digitize the sensor using 10 to 12 bits and map the values to 8 bits using a non-linear gamma transform. On the more expensive ones the user can adjust the gamma, Knee level and master pedestal according to his/her tastes. This is possible on the Canon H1, A1/G1 and I suppose also on the Sony V1. Typically, the user will adjust the Knee point lower to avoid oversaturating highlights (e.g. in concert show), press the blacks to hide the noise in the darker parts of the picture or, on the contrary, extend Knee point for a high-key effect, increase the blacks fr a low-key effect or lower the gamma if a softer picture is wanted.

Of course, each of those settings has an effect on the dynamic range as perceived after reduction to 8 bits for recording on the tape.