TrueColor V3.0 available on-line
 

Hi all.

This is not news since I talked about V3.0 for a couple of weeks but I finally managed to have a short page writtent about it with the complete "recipe" written down. This is, IMHO, the most accurate configuration that I can define. Given the real life situation in which we used it, shooting 3 commercials with lots of vivid colors, skin tones and dark/bright spots, I find this configuration to be one of the most flexible available for our camera. Give it a try and let me know.

The description is "Part 3" at the top of http://www.paolociccone.com

Oh, and take a look at the "Stilt World" pictures if you have a couple of minutes :)

Thanks for sharing! :)

Very good Paolo. Thanks for doing the work.

PS The settings are very close to PanaMatch now with a couple of extra tweaks.

You're very welcome

Thanks...for TC3

Great minds think alike ;)

Thank You!
 

Hi Paolo,

Thank you very much for all of your efforts!

I will send you a private message later with a link to some of the stuff we have been shooting with your settings.

Your hard work is very much appreciated.

Tom Chaney

www.tomchaney.com

PS When you say Black Stretch off, do you mean "Normal"?

Great work!

1) Your goal of providing the "most bits into post" is exactly the same approach I describe in my HD100 Handbook. We both use histograms to indicate how well one is doing this.

However, unlike Sony's A1, the HD100 has no histogram display. So how are you setting exposure?

I think that even if folks get convinced they should use an "engineering approach" -- actually setting exposure in the real world is not easy.

I present the "Zone System" in my Handbook, but that doesn't mean much to young videographers.

2) Panasonic talks of the dangers of a knee causing color shifts. I'm wondering if you checked this. Did this lead to your use of 90IRE rather than the 80IRE that Tim and I have suggested?

3) In V3 you switched BS from 1 to OFF. It's clear, by looking at a histogram, that 2 (and especially 3) really alter the mid-levels way too much. But, I'm wondering why you switched BS off. I trust your lab tests -- so I'm wondering why "on location" you made the change.

By the way, your DSC colors are spot on. There appears to be no true yellow in any of the other settings! (Again, we both agree that we want to go to post with an exact recording of what the real world looks like.)

Folks should look at the shade of the DSC grayscale in Part 1. Your DSC grayscale goes from true white to real black.

Thank you Steve. I followed your posts in the past months and I'm impressed by your knowledge of HDV. Recognition coming from someone like you is greatly appreciated.

I took the hint about histos from Scott Billups. He mentioned that in his book and I thought it was a great idea to have a more objective analysis of the picture.

Mostly via the Waveform monitor. I look at the zebras too.

Agree if we are talking about outdoor, short time for setup, kind of situation. Indoor, controlled environment is a lot easier. For the commeriacials that I helped shoot with Image Line (Ben Jehoshua) the DP used an on-camera LCD display by Astro with built-in WFM. He kept the level generally just above 80%.

Ansel Adams, I looked into it but I didn't grasp it at first. Thank you for the reminder, I wanted to go back and understand it, it's gonna be my proposition for the next week :)

That's correct. If I remember well, I pretty much spent a day switching between knee levels and examining the WFM and VScope. I don't know much about color shift but the WFM shift was modest but more visible at more aggressive knee settings. I just believe that the camera can hold the highlights nicely with that amount of knee, without risking to alter the digital signal more than that. Ben and I were actually examining some nice bright reflections on plastic toys during the commercial that we shot. The highlights, with knee 90% rendered very nicely. Here are a couple of pictures of the lighting on the set. The second picture is about the light at the far right of the scene. As you can see we were using tons of light :)

Main scene
Right side

Well, when I did that I was also looking at the WFM and I noticed that MB normal with BS off leads to the same rendering of blacks while affecting the midtones less. My first approach, when I started V1.0, was to drop the MB until I reached the 0IRE level. It seems to me that V3.0 achieves the same but without stretch. I still map the User1-3 buttons to BS 1-3, just in case.

Cool! Thanks for the verification. That's what this forum is all about. I love that we can validate this kind of test from multiple sources. It makes it much more useful for everybody.

Yeah, it's kinda strange that the JVC settings are not better. To me the inspiration to calibrate the camera was created by the disappointment I had with the colors from the default settings. I didn't know much about HDV but what I got out of the box was less exciting than my $1000 Optura DV camcorder, colorwise. With TC I'm finally getting the exciting colors that I expected.
I believe tha this is truly a confirmation of the value of this camera, that we can tweak and personalize so much at this price point. I believe that, in a few years, we will see the HD100 mentioned in books as a milestone in the evolution of digital film making.

Thank very much for the excellent feedback!

Thank you for your very kind comments!

All my explanations, here and in my book, break down at the point of actually setting exposure. It seems one needs to do three things at once:

1) Get the real blacks under 5IRE.

2) Get the brightest white above 105IRE.

3) Confirm that faces -- which must be exposed correctly -- are correctly exposed.

And one must be able to do this in three different types of situations:

A) High contrast -- 8 or more stops of light.

B) Medium contrast -- 4 to 7 stops of light.

C) Low contrast -- under 4 stops of light.

To add to this problem, LCD monitors (even an external monitor) are not good at displaying true blacks. So, we really can trust our minitors to set blacks to black.

Let's assume we use zebra to monitor face levels. Now we depend on LCD monitors to judge how bright whites are. Once again, LCD monitors tend to be very bad at displaying bright whites.

For those who aren't using a WFM or have no way to display a histogram -- we only have zerba to set exposure. Worse, we only have one zebra.

I realize with your rig these are not your problems, but they are problems for those of us who shoot alone.

I do have one idea -- a tiny box fed by the composite video cable. Logic would light two Green LEDs when "some" of the signal was under 5 IRE and "some" of the signal was above 105IRE.

Let's see if we can figure this out.

This can be solved by using DVRack. It has zebras for both highlights and dark tones.
I hope you don't refer to the shoot with Ben & Co. as "my rig" :) I can't even touch that. I decided for DVRack exactly because I can run it on a laptop, using the battery, and so it becomes a very portable solution for one-two people shoot.
That would be cool if we can make something like that at a very affordable price point (~$200). I don't know the cost of the Astro LCD, on-camera display but it was really useful, with it's WFM superimposed over the image. It has component input.

HDVRack really isn't practical for gun&run or even solo shoots.

IF the HD100 had a constant ASA sensitivity, it would be really easy to use a lightmeter.

But, one might be able to use Zebra -- since it indicates what the camera is registering -- and from it -- know what the low and high foot-candles should measure to ensure all the bits are used.

For example, if the Zebra indicated 70IRE on a face, the brightest area on the light meter should be almost 2 stops hotter (110IRE). The darkest area, should be about 3.5 stops lower. (This isn't exactly right because a stop isn't always 20IRE.)

Hi Paolo,

Do you mean Black Stretch "normal"?

I can't seem to find "off."

Thanks,

Tom

Given that HDV is not great for brightening in post due to compression artifacts, I try to get fairly bright without clipping (depending on the subject). But I was wondering what are some good numbers for exposing different subjects at. For instance would you say that the upper limit for skin tones is around 70 IRE? I really wish that there was a waveform monitor or histogram available in the camera, but given the zebra function that we have, what's a good strategy for shooting with them in "the real world" without external equipment?

Ciao Paolo.

And Grazie!

I wish I had found your tests before doing my own. Sort of like coming up with the Theory of Relativity anew, only to find out someone already did it :)

I also calibrate to the DSC chart (DSC makes a lot of handy charts, I use them all the time), and my HD-100 "broadcast settings" just about mirror your true color settings.

Where they differ is: detail (-5), black stretch (stretch1), color gain (-1), and red gain (2) - these are my adjustments based on "after editing" ATSC broadcast ready segs. I also have the skin adjust set to on, and the range set to -2.

Turning on the skin adjust feature changes the overall color matrix of the cam somewhat which may also account for the differences.

While my "standard" setting is as above, I do adjust the settings as the situation calls for. No one setting is correct for every shooting situation, but your settings are a excellent starting point, and will hold a solid picture in most situations. Thanks for making them available.

I'm taking two NTSC HD-100 cams to Cannes tomorrow, we are shooting PAL Digibeta for the most part, but I am shooting some festival atmosphere with the HD-100s for the US domestic broadcast market. We are also taking two Pana NTSC HVX200 cams and a Steadicam Flyer for one of the HD-100 units. I'm curious to see how the Mediterranean and added range of festival colors mix with my standard color settings.

Full report when I return.

Yes. It's the same, meaning that there is no stretch so the black is normal. But you're right, that's the correct term.

Hey Enzo, e tu lavori a Hollywood e Cinecitta' ? ;)

Thank for the compliments about TC. So, you're keeping the detail level quite high, eh? I'm interested in this. I found the level detail too "aggressive" and I'm constantly shifting between MIN and OFF.
If you do broadcast work normally, TC might be reproducing "illegal" levels. Of course you can reduce them in post but you're settings require less manipulation. I'm aiming at making the HD100 as close as possible to film so I didn't check the values against broadcast levels.
I never touched the skin tone circuit as I don't want to have other factors changing the parameters out of my control. I'm glad that you mentioned, it confirms my fears :)

You're very welcome. I'm glad they have been of help.
Sounds like you're gonna have fun in the French riviera. We are all eager to see some clips or frames, if you have the chance.

In bocca al lupo!

When JVC released the HDV GR-HD1, folks complained about its extreme edge enhancement. When they released the JY-HD10, JVC set EE lower. However, folks claimed it was still too much for film. JVC responded that a 30p camcorder wasn't really intended for filmaking (true) and that, for broadcast HDTV, "some EE was necessary." So assuming JVC was correct, it makes sense that for ATSC Enzo would use more detail than you would.

You may be right that his shifting both Color Gain and Red Gain down one notch may guard against exceeding broadcast levels. (Although one would expect JVC would have set default Color Gain correctly.) Alternately, DSC makes some comments, I don't understand, about THEIR 75% color bars being too hot.

If this is the case, there may be two TCs: TC-film and TC-video.

By the way, these two links seem broken:

Main scene
Right side

Makes sense, thanks.

Yeah, I was surprised about the color gain set lower but since I don't have all the other settings in Enzo's config, I'm just guessing.
The DSC colors, if I read their comments correctly, will cause the NTSC VScope to reach 75% of the level. They suggest using gain for positioning the colors in the boxes.

Thanks for link check, I fixed them.

posting of true color v2 in my documentary
 

http://img150.imageshack.us/my.php?i...vetsb404fb.jpg

http://img249.imageshack.us/my.php?i...liteer01yz.jpg


hi, these are the captures from JVC 101E HDV capture using truecolor v.2.
I set all color at zero. as i prefer the warmer tone.

the lighter image was after adjusted the middle bar at levels
overall brighten up and not losing an highlight details

Jy

Thanks Paolo for your greatly appreciated work
 

Does you TC3 also increase latitude? We're using the HD101's live at the moment, and did a large fund raising concert on Friday night, but we found on standard settings, we had to run the cameras upto 12db gain. Just wondered if you settings would save using so much gain. Thanks

Hi John.

Hard to rate the colors on that kind of image since there are very few colors, they are generally desaturated and you are shooting against light.
Myself I would go for a darker image but I'm glad it works for you.

Hi Adam.

V3 doesn't change the latitude of the settings. It gives a better balance and contrast but it's a small variation compared to V2.
Depending on the type of concert you can use black stretch at 3, longer shutter speed (1/30 if shooting 30fps or 1/24 if 24fps, for example) and gain, and a combination of all 3 :)

darker i also prefer
 

Paolo,
I agree too, darker seems better in color saturation.
but i just show what can be done in shadow and dark detail
note that the man 's ear shown in shadow.

I m not think in video so i only think in film ( not film look!)
i have not to test a 35mm blow up but i think ur TC has a very sound and linear exposure latitude for blow up.
as this is always confusing to many people that they mis-used film look or cine like gamma.

i test with sony and found cine gamma is just a 3 stop cut.

in plain exposure term -- how many stops ur TC v3 can preserved?

i still prefer to use a spot meter with ASA/ISO setting for good zone system exposure allocation ( pardon my english)

have u test that the JVC has a constant ASA/ ISO?

color wise i rather do it in post, i actually prefer t o do it as im age sequence or use cineform for output

JY

I see now what you mean, thanks.

In terms of going from real bright to very dark, I measured the camera di be able of ~6 f-stops with TC. I measured this with the grayscale of the DSC chart, stopping when the last 2-3 blacks where barely visible. This means that, in real life situations you would be barely able to see the subject but, on the other hand, you are still able to recognize details. This procedure was suggested to me by Scott Billups.

I haven't done that test but, if I remember well, Tim D and others reported a variable ASA value depending on the gamma settings. Adam Wilt rates the camera at "little less than 320" in his review on DV Magazine.

I wanted to follow-up on this topic.

Have you -- or could you -- post a vectorscope of your V3 settings and a vectorscope with the Color Gain set to -1 rather than NORMAL and Red set to +2 rather than +3.

If it turns out your V3 is too hot for Broadcast video, then you may want to post two versions: TC-film and TC-video.

Lastly, could you find the time to post one second of .m2t V3 (film) and perhaps V3 (video).

Hi Steve.

I'll shoot a clip of the DSC chart with TC V3 and I'll get a frame grab of the VScoper + histogram. I'll also post some outdoor footage probably tomorrow.

Cheers

Steve, I just remember that I have posted several VScope images for V1, and V3 can't be "hotter" than V1. V1 looks perfectly in the safe area, regarding saturation levels. V3, having a lower gain in the blue, not the red, is even more in the safe area for video.

Looks like we have to wait for Enzo to return and explain why he slightly reduced color and slightly increased red. I wonder if that was his taste in color -- or if there was a technical reason.

For example, we know we can have very saturated red using HD. But what happens when we broadcast an NTSC or PAL version of an HD production?

How about when we make an NTSC or PAL DVD?

Does the conversion from HD to SD colorspace automatically reduce chroma saturation IF it is necessary to reduce it?

Fasinating stuff guys!
 

Paolo and Steve - great work on this thread. Very educational.

Thanks again Paolo for all the hard work you've put in - and we'll for sure mention you in our credits. Keep up the good work guys!

john
evilgeniusentertainment.com

You're very welcome John.
I'm glad this has been of help.

Sì, Paolo, e gli Studi di Eclair (FR) anche!


I set the detail depending on the frame, and no, the scene file(s) I use
are well within the ATSC broadcast legal envelope. Yours will be just about
the same. Generally, the Skin Detail setting (on) will desaturate reds somewhat and
spike a very narrow range in the yellow spectrum down.

You are using the HD100 to match film, I only use it for broadcast. I'm
fortunate enough to use film cameras for film. But, I will post more later
on several on set comparison tests ("on the fly") I have done with the HD100
vs. a S16 Arri, A Pana Genesis, and several Arri and Pana film cams, plus
the VariCam and CineAlta.

I think JVC is just starting to scratch the surface as far as making the HD
series a "real" film camera alternative. And I think the split eyepiece is a
good indication of that. It makes a perfect jumping off spot to add an
eyepiece extender so the cam can be used on a gear head.

For a straight to DVD 16:9 movie, you are not going to find a better cam
under 25K, period. I have tested all the other HD cams, about 40 hours worth
on each (about 80 on the HVX200), and they are all great for the price point
they are selling at, but the JVC is better than any of them (in my
opinion -- more on this later also).

Can't post clips or even frames from my Cannes footage due to copyright and
embargo issues, but if you saw the CNN or any of the US entertainment
based shows broadcast footage of "The Da Vinci Code" opening day, the
Steadicam coverage of all the principals walking up the stairs and then into
the press crowd is all my operating (using a HD100 and Steadicam F-24). We
were on a pool feed, and ours was the only crew with a Steadicam.
The F-24 is a perfect fit for the HD100 (used in both high and low configs),
and even after a few hours of continual use, I still had legs :)

I see that someone has a thread entitled "Turn Power Off Turn Back On
Later". My HD100 must have been made in Italy, it says "Turn power off,
have a cup of espresso, turn back on later".

L'azione non significa pizzicare di fermata, significa appena il pizzicotto
più veloce.

Steve,

No, it's not my taste in color that determines my color settings, it's what
I know the output will look like in the broadcast envelope. I am constantly
changing settings (on all the cams I use, not just the HD100) to match my
frame. Paolo's TC settings are a good starting point, but the settings
always have to be optimized for the frame you're shooting.

I reduce both color AND _red_. Lucky for us (and JVC), the ATSC standard is
pretty loose right now, but as you already mentioned, it has a pretty strong
magenta bias.

When you add broadcast engineers that have tons of NTSC broadcast experience
and are just starting to work within the ATSC output frame to the mix, you get a strong
red bias and oversaturated colors along with a lower gamma curve. The HD100
loves to shot more on the toe of the curve than most video cams, and I think
this was a good approach for JVC to take. The downside is that failing to
have enough HD experience, the broadcast engineer will set the mid point IRE
lower than he/she would for NTSC output.

As I do with film labs (I set the MY mid exposure to THEIR mid point light
for the entire production), I am starting to set my color and detail
settings according to the broadcast venue that is airing the footage -- when
I can.

Yes and no, the color space changes on a HD to SD NTSC broadcast convert
(higher gamma curve, more desaturated colors), but very little on a HD to SD
PAL convert (mostly a little higher gamma curve). It makes a fantastic
convert to DVD in a 16:9 frame. Really a strong but little mentioned aspect
of the HD100. Of course we all know NOT to shoot the HD100 in 4:3 SD mode,
it's not what the camera was designed for

The HD100 is not really a true broadcast cam, but it's the best HD cam that
can be used for HDTV 16:9 broadcast output available for under 25K.

>>Can't post clips or even frames from my Cannes footage due to copyright and
embargo issues<<

Embargo issues?

The French never forgave us for the "Freedom fries" deal, this is their retailiation: no footage for you!

;)

Ciao Joe,

I know, it's a very confusing term :)

I work in commercial broadcasting, and on studio funded films, so basically my output is on a "work for hire" basis. I don't own what I shoot, and most of the time, I don't even see the tape I have shot. It goes right to the broadcast outlet I am shooting for (but I do check my return video a lot :).

An "embargo" in commercial broadcasting is a condition placed on the coverage you are shooting in exchange for a studio (or a celebrity's rep or agent) giving you access to an event or personality.

In the case of Cannes, I was shooting for French commercial TV, Italian commercial TV, and doing a pool feed for several broadcast wire-services (domestic and international).

Besides the common sense copyright issues (I don't have any rights to what I shoot on work-for-hire gigs), there was a Sony sanctioned embargo (condition for Sony arranging access) on all Da Vinci Code coverage to the US domestic market because it was an Access Hollywood broadcast exclusive for the North American market.

Embargos are usually time, venue, or broadcast overlay specific (can't be shown before a specific date, or by a particular outlet, or on the Web, or a specific geographical area). In addition to the above, almost all the Cannes events (except for the table interviews which are pretty useless) had an embargo against any Web use of any kind on their granted access coverage (except for a few pre-approved outlets).

Hahaha, so true, but I think it's the "Le Big Mac" that done us in :)

It basically depends on what side of the pond you're on.

In America we say "We saved France's a** during the 2nd World War".

In France they say "We saved America's a** during the Revolutionary War".

Like all things, your point of view is based on which side of the camera you're on.

Miseria!

That will be very interesting. Just to be clear, I know well that the HD100 cannot touch the quality and resolution of film. Any Arri camera will beat the crap out of the HD100. No argument here :)
My interest i the HD100 is based on my "digital background" (20 year of software development) and the fact that, for what I want to do, it's good enough. I might add the M2 along the way.

Good to know!

What about the Royale with Cheese?

INHD has been showing films from the `70s and `80s. I've been feeling a bit guilty watching these in HD and feeling "video ain't ever gonna look like this." The wonderful texture of grain just feels right. The colors are so great.

I watched "Umbellas of Cherburg" last night. Would love to re-see Godard's "La Chinoise" with each room painted a super saturated color.

Or, maybe I just like the way the French and Italians used color. "Juliet of the Spirits," for example. I supose one could try setting an HD100 to a high level of saturation, designing sets, and adding grain in post.