Matt Sandstrom

if you polarize the eye light and then filter on the camera with the same polarizing axis, that should let the specular highlight through while cutting a stop from the diffused face reflection? i haven't tried it but it sounds like fun.

great info, thanks. so will it extend the range thus lowering the contrast or will it let my whites go even hotter, if you know what i'm saying? i'm asking because "video gamma" is the only color space i can monitor properly so i was hoping to grade for video and just take advantage of some hot headroom.

yeah this is why i'm not fully confident with dpx and so on, but then again neither with tiff. after effects has a cineon export option which does dpx and seems to be taking care of the lin-log conversion, see setting below. guess i really have to get hold of some expert at the post house. :-)

thanks, that makes sense. i'm not transferring film nor shooting raw though, i'm working with regular video footage. the thing is that everything was (wisely) shot to protect the highlights, and when i bring the images up in post it would be nice to be able to keep the highlights bright and contrasty letting them go as hot as they want rather than compressing everything between 95 and 100 ire, which both gamma correction and broadcast legal filters do. if that means they clip in the cinevator that's obviously not what i want, but if they are preserved it would be great. i guess i could always render to pdx or something instead, but i'm more confident with video space and there's still quite a bit of headroom over 100 ire.

hi,

i've done a few tests with the cinevator and while the results have been great i just realized that i always used "broadcast safe" hd masters. does anyone know if i might get more contrast and more punch in the highlights if i let them go all the way up? maybe i can even retain more saturation in the extreme highlights this way? i mean i should but how does the cinevator handle this? my contact at the post house doesn't know, but then again that's not really his field either.

(the next problem is that i don't really have monitoring to properly check superwhites but that's what testing is for, right?)

thanks,

/matt

wanna trade?

/m

Kodak 7250 400T/ 250D color reversal stock is quite grainy and low contrast, but I am sure it was discontinued a couple of years ago. There may be some leftovers somewhere.

for the record we've already shot the footage on video for various reasons, otherwise i agree that shooting film is always the best way of getting the film look. :-) i mean i'm not trying to make this look like film, it is what it is, but it's supposed to look a certain way that i think it's easiest to describe that way.

/matt

If you want a 16mm reversal look, I don't know if grain is the answer. All of the 16mm reversal I've ever shot was very slow, very saturated with nice inky blacks, and was quite smooth.

sounds like you've shot mostly kodachrome? i'm not looking for huge grain, but the old vnf stock was definitely grainer than any any modern negative, especially if you're trying to create the 16mm look on 35. yes, the most prominent feature is the contrast and the weird saturation where it sometimes looks almost b&w and sometimes super saturated, but that i'm handling digitally. i've added grain that i'm pleased with too digitally, but i don't think it will transfer well to film.

/matt

Thanks guys, much appreciated. Yeah i've thought about super 8, a medium i've worked a lot in as some of you may know, or 16mm but it complicates things too much in this case, plus it probably gets more expensive, plus quality control becomes a problem, after all this is for a theatrical release and not an "art piece". /matt

hi,

i'm recording a 720p video project to 35mm and i want it to be quite grainy, like old 16mm reversal. for the digital master i've added "film grain" using software and i love how it turned out, but i guess when going to film it would be better to create the grain optically/chemically? my idea would be to use the grainiest film stock there is, fuji reala 500d or f-400 perhaps, then underexpose it a stop in the recorder, then bleach bypass the print? i'm not sure if that's gonna give me enough grain though, but i can't think of any way of adding more. what about adding grain/gaussian noise digitally, will it look fake on film? i can imagine it becoming a "layer" rather than an organic part of the image. i know it's done all the time but subtly, heavy grain sounds like it would be a different story.

thanks,

/matt

some blackmagic configs are fully compatible with apple's own uncompressed codec, i know 10-bit yuv is, probably 8-bit too, but maybe only in quicktime containers, and blackmagic has a bunch of configs that are not compatible. since it's uncompressed you can convert to whatever without loss though. the blackmagic codec isn't gamma corrected for the mac, so conversions to and from rgb as well as gamma corrected codecs like apple uncompressed, dv and mpeg-2 will cause color and density shifts.

/matt

jose, i think you're wrong. cineframe25 works quite well and unlike cineframe24/30 doesn't ruin anything whatsoever. you will only get a very slight increase in sharpness by doing it in post, and in my tests i noticed that the lower amount of data actually helped the compression causing fewer artifacts. this said i've also shot interlaced for all the projects meant for 35mm, but i'd test it first, and i've used cf25 for hd projection and it looks gorgeous. you can get test blowups in slide frames from many labs that you can project and compare.

/matt

i've done it plenty of times. the light is beautiful and quite easy to capture, the only real problem is that the brightness of the sky changes dramatically every minute. you probably won't see the difference until it's on film so use your meter constantly. using nd's and working at a pretty constant aperture is probably a good idea.

where are you? while magic hour can last several hours here in sweden it's usually not even an hour long further south.

/matt

nobody mentioned cts yet? [EDIT: ok, one] i almost always prefer it over cto. i've lit two entire features for two different dp's with white diffusion (216?) as my only diffusion so the others can't be *that* necessary, if you can only get a few. as for theater/party gels i carry a small bag with every little piece i've ever found on the floor on the set/theater/party. they are very useful but there's no need to carry rolls or even sheets.

/matt

hdv is *always* uncompressed before color correcting. it's then recompressed if you're rendering back to hdv, but that's abother story. i like to master both my dv and hdv projects to uncompressed but converting to uncompressed before is just a waste of disk space.

/matt

alex, what exactly is not true? i'm sure you know about s/n ratio and bandwidth, two entities that limit beta sp in *exactly* the same way as digital quantization and compression limits the digital formats. the artifacts when you push them to the limit look different, but the limit is still there, no cheating unfortunately.

what is this thread doing on the lighting board anyway. :-)

/matt

Forgive me, I should have expanded what I meant. Tweaking the camera's internal menu color controls too much is where the damage to the image can be "bad". Using filtration on the lens won't necessarily degrade the image to an unacceptable level (you judge what's acceptable) so long as the chips aren't deprived of light to the point where noise creeps in.

i understood you just fine, but i still don't agree. tweaking the "internal menu color controls" does much less harm than doing the same in post, and compared to an optical filter there's not much of a difference in degradation either. the camera converts raw to yuv anyway, whatever setting you chose and even no setting is also a setting, just one selected by sony. of course you can't bring back blown out pixels, which an optical filter can, for example. which brings us to the wise suggestion above: test, test, test! ;-)

/matt

The falling apart is just as true when you tweak the image too much in camera.

not really, since both optical filters, white balance and other image settings are applied before conversion to video gamma and hdv compression. but of course *too* much tweaking is bad, that's why it's called *too* much.

/matt

If HDV falls apart quickly in post, could it be related to using firewire in rather than the more robust component pathway with a higher quality codec?

no, it can't. firewire is the most "robust" transfer you can get since it's an error corrected digital protocol ensuring a bit for bit identical clone of what's on your tape. however if you're not shooting tape but outputting component live to your computer, you do get better quality since you bypass the hdv compression. some people do this on studio shoots and even on location, but the hdv codec isn't *that* bad. the whole idea of using a small hdv camera is to move fast, right?

/matt

My math says 1.85 is 'exactly' between the two. (2.35+1.33 divided by 2 = 1.84)

that's assuming the height is constant, meaning the area increases as you widen the aspect, but if the width is constant thus reducing the area this is the calculation: 1/((1/2.35+1/1.33)/2)=1.70. inbetween 1.70 and 1.84 is 1.77.

i also like this way of counting:

4/3=4/3

4/3*4/3=16/9

4/3*4/3*4/3=64/27~2.35:1

/matt

gain reduces the dynamic range, not the definition, the result is more noise. you can use 9db without problem if it's for tv. in fact i like to use it even for things going to hd or film since i'm not very fond of the smooth texture this camera produces, especially with detail turned down.

/matt

i recommend doing as much as you can in camera, since hdv falls apart quickly in post. the z1 supports color phase and level corrections as well as wb shift in your picture profile, play around with that, and the white balance of course. i like to white balance on the blue sky if i want the tobacco filter look. as for filter vs digital you probably lose about the same number of stops. if you shift the white balance towards orange you're underexposing the blue channel the same way the filter cuts the blue light.

/matt

For shorts I always throw in a 2.5k HMI. In the UK you can plug into the wall

well, that's really pushing it since i'm pretty sure many breakers are 10 amps there just like here, which at 250 volts means you're at the very limit. i never dared. maybe i should because a 1.2k is a bit limited in daylight, even indoors. of course there's usually a 15 amp circuit or two in most places.

/matt

yeah, projected you will get less sharpness with anamorphic since it's magnified more, while on tv you will get more sharpness since it's magnified less, since the width is fixed. 16:9 is exactly betwen 4:3 and 2.35 though so in that case it will be exactly the same. although the anamorphic lens probably isn't as sharp as a spherical one, as others have mentioned.

/matt

we usually bring one 1.2k hmi, one 4x4 kino, two single tube fixtures, and a couple of redheads. plus stands, flags, gels, frames and so on of course. it all depends on what you're doing but this kit has served us well for shorts, features, music videos and commercials, indoor and outdoor, winter and summer. it's very limited for daylight exteriors though, for which you need bigger frames and more powerful hmi's. and forget lighting very creatively. it's all about working with the available light, extend it, fill it, block it. ok, that's creative too but you know what i mean. but hey, it fits in a station wagon.

/matt