Survey: What min/max levels does your cam shoot?

Nick Hope wrote on 6/17/2011, 4:24 AM

I'm keen to build up a list of what black and white levels various cameras are shooting by default, in order to reduce the guesswork when giving guidance about manipulating levels in post production.

I'm hoping to get results for a range of video cameras, dSLRs, Go Pro, phones, etc. etc.. The more devices we get results for, the better.

I am aware that minimum and maximum luminance levels can be adjusted on some cameras, so please set them to default settings as far as possible.

Black footage method: Cover the lens with the lens cap or whatever, close the iris, set gain to zero, set the shutter as fast as possible.

White footage method: Open the iris as wide as possible, set gain to max, set the shutter as slow as possible, remove any ND filters, point camera at sky.

Get the footage on the Vegas timeline, match the project settings to the footage, set pixel format to 8-bit and set preview quality to "Best (full)". Note the levels in the histogram video scope. If you put the cursor over the histogram, you can see the luminance value in yellow at the top right of the video scopes pane. Let's limit the results to the overall luminance (the white histogram) unless there is any wild difference between the R/G/B channels.

Here's the black histogram for my Sony Z1P HDV camera. This shows the minimum black at 17 and the black "peak" at 20:

White histogram for Z1P. White peak at 254. Maximum white at 255. (I have no idea what the secondary peaks are):

Black histogram for my Sony VX2000 DV camera. Minimum black at 13 and the black peak at 17. (Not sure why it's giving me the secondary peak):

White histogram for VX2000. White peak at 253. Maximum white at 254:

Please post the camera make and model, approx year of purchase, video format, minimum black level, black peak level, white peak level, and maximum white level in the following format. Feel free to post screenshots of the histogram too if you have the time and inclination.

Sony VX2000, 2003, DV, min black 17, black peak 20, white peak 254, max white 255.

Sony Z1P, 2006, HDV, min black 13, black peak 17, white peak 253, max white 254.

EDIT: I will maintain the results below as we go along.
EDIT: Note that many hybrid stills/video cameras shoot 0-255 for stills, but a more limited range for video. Please test the video mode.
EDIT: Please add what format your footage was shot in. This is especially important if, for example, it shoots both AVCHD and .mov. If possible, please test both formats since they may well be handled differently in Vegas (see Laurence's posts of 4/28/2014).

Cameras that shoot nominally 16-255

DV

  • Sony VX2000 (Nick Hope)


HDV

  • Sony A1 (LoTN)
  • Sony FX1 (johnmeyer, riredale)
  • Sony Z1 (Nick Hope)
  • Sony HC3 (riredale)


AVCHD

  • Canon XA10 (MikeLV)
  • Canon Legria HF S100 (Pete Ferg)
  • Canon Vixia HG21 (amendegw)
  • Canon Vixia HF G20 (aka Legria HF G25 in Europe)(Lou van Wijhe)
  • Canon Vixia HF G30 (OldSmoke)
  • Canon Vixia HF M40 (DaveT2)
  • Canon Vixia HF M52 (_Lenny_)
  • Panasonic TM700 (amendegw)
  • Sony NX70 (Stephen Crye)
  • Panasonic AG-AC130 (amendegw)
  • Sony HDR-XR520 (PeterDuke)
  • Sony NEX-FS100 (paul_w)
  • Sony SLT-A77 (Udi) - (actually a little darker - see post below)
  • Sony TX55 (ritsmer)
  • Sony NEX-EA50 (malowz)
  • Sony NEX-FS700 (essami)


XAVC S

  • Sony AX100 (Rich Parry)


XDCAM EX / MOV / AVI

  • JVC GY-HM750 (Tech Diver) -(adjustable pedestal (min black level))


Cameras that shoot nominally 0-255

  • Apple iPhone 4 (Grazie)
  • Canon 5D MkII (farss)
  • Canon 70D** (malowz)
  • Canon 80D (john_dennis)
  • Canon Powershot G9 (john_dennis)
  • Canon Powershot G15 (john_dennis)
  • Dell Streak tablet/phone (Marco.)
  • GoPro HERO5 Black (Nick Hope)
  • Kodak PlayTouch (musicvid)
  • LG Optimus One P500 Android phone (Nick Hope)
  • Nikon D5100 .mov (Laurence)
  • Panasonic Lumix GH3 .mov (musicvid)
  • Samsung Galaxy II (megabit)
  • SN9C01 webcam (farss)


Cameras that shoot nominally 16-235

  • DJI OSMO (Nick Hope)
  • Huawei Mate 9 phone (Nick Hope)
  • Panasonic Lumix GH2 AVCHD (.mts) (Marco.)
  • Panasonic Lumix GH3 AVCHD (Laurence)
  • Panasonic Lumix GH4* (wwjd, Nick Hope)(can be set to shoot 16-235, 16-255 or 0-255)
  • Sony EX1* (farss)
  • Sony RX100 (ritsmer) (Nick Hope's shoots 16-240)


* - depends on how the camera is set up
** - other settings can be created with Canon Picture Style Editor

Save

Comments

farss wrote on 6/17/2011, 4:38 AM
Well with my EX1 correctly setup Black = 16 and white = 235 at 100% zebra with factory gamma. Of course different gamma curves will set 400% white to 255, thanks to Serena for the considerable effort for ploting the curves that show that.


Nick,
a horrible thought has entered my head. I fear many are falling into a BIG trap with Vegas. Vegas out of the box with an out of the box computer, Vegas's preview monitor, does NOT DISPLAY VIDEO CORRECTLY.
To see your video correctly you must apply the Levels FX with the Studio RGB to Computer RGB preset to the preview monitor. You must ensure you turn it off for rendering.

If you do this, render and Upload to YoutTube what you see on the average PC will match what you saw in the Vegas monitor with the correction applied. If your preview monitor is not correctly setup in the first place you'll think YT is screwing with your levels.

If you want to check this it's pretty easy. Set the preview monitors background to black and then white. Your video's black and white should match those backgrounds.

Bob.
amendegw wrote on 6/17/2011, 5:13 AM
I love these tests!

Panasonic TM700, AVCHD:
...Black min:13 peak:18 max:22
...White min:243 peak:249 max:254



...Jerry

PS: If I get ambitious, I'll try to get the same test with my Canon HG21
amendegw wrote on 6/17/2011, 5:29 AM
Side question, "What's the cause of the lower level peaks on Nick's all-white Histogram? Some type of harmonics? And why don't I see them on my TM700?"


...Jerry
farss wrote on 6/17/2011, 6:08 AM
""What's the cause of the lower level peaks on Nick's all-white Histogram?"

I cannot say for certain, could be from quantization error, slight variations in the sensor pixel sensitivity etc, etc.
There's a whole lot of calibration required in cameras, most cameras calibrate on power up,. They can do this because the camera can control the iris i.e. close it, calibrate out the black errors. When you get to the high end cameras you have to do this manually which is a PIA if you forget or do it wrong. I did it wrong once and was in a panic for hours thinking the camera was stuffed :(

Bob.
megabit wrote on 6/17/2011, 6:31 AM
Bob, do you mean the Black Calibration (or Balance, don't remember what it says exactly in the maintenance menu)?

How did you use it - it's always greyed-out on my EX1?

Piotr

AMD TR 2990WX CPU | MSI X399 CARBON AC | 64GB RAM@XMP2933  | 2x RTX 2080Ti GPU | 4x 3TB WD Black RAID0 media drive | 3x 1TB NVMe RAID0 cache drive | SSD SATA system drive | AX1600i PSU | Decklink 12G Extreme | Samsung UHD reference monitor (calibrated)

amendegw wrote on 6/17/2011, 7:24 AM
Okay, Here's my Canon HG21, AVCHD:
...black: min: 14 peak: 15 max: 17
...white: min: 251 peak: 253 max: 253



This appears to be the first camera to show variations between the luminance & R,G,B. Is this because it's the only single sensor camera in the test so far?

...Jerry

Edit: All my histograms were captured at preview = best/full.
johnmeyer wrote on 6/17/2011, 9:20 AM
I am probably being really thick, but doesn't the Video Scopes Settings make a difference, and shouldn't you specify what it should be set to before anyone does this test? My assumption is that only Studio RGB should be checked, as follows:



My Sony FX1 HDV camcorder:

Black: Min: 15; Max: 21





White: Min 251; Max 253



P.S. I originally typed the word "s-t-u-p-i-d" instead of "thick" in my post above, but the Sony forum replaced that word with all asterisks!!! Did they install some sort of obscenity filter, and if so, is this word considered vulgar?
Nick Hope wrote on 6/17/2011, 9:20 AM
Bob, I'm sure a large proportion of users don't realise that the preview window is not showing them the same levels they'll get in most delivery scenarios, or they just don't really care about it. For years (in SD world) I noticed that it never showed showed deep blacks but I didn't really care because the image on my TV, hooked up over IEEE1394, looked fine. As an underwater videographer I was always more obsessed with fixing redness or greeness than in tending to the overall luminance. Even in HD world I trundled along with it for a good while before digging deeper and learning what was going on. As long as the DVDs were looking OK on TV, I was happy.

Jerry, thanks for the great replies! I have no idea what those secondary peaks are. They wander around during playback. There is no sign of any grey bands in the footage.

Hopefully what all this lead to is being able to give better guidance on what to do to video levels before rendering, for situations where one doesn't want to individually grade every clip. For example, apply levels preset A to footage from cameras of type A, preset B to footage from cameras of type B etc..

Keep 'em coming folks. Any devices at all. Would love to see what dSLRs do, Go Pro, pocket cams etc. etc..
johnmeyer wrote on 6/17/2011, 9:25 AM
On my IRE/Studio RGB question, a little more testing reveals that it only affects the Waveform display and not the Histogram. So, its settings don't need to be specified for this test.

Sorry to be s-t-u-p-i-d.

Nick Hope wrote on 6/17/2011, 9:26 AM
John, that setting has no effect on the histogram (stupid Vegas), which is why I chose that for these tests instead of the waveform scope.
johnmeyer wrote on 6/17/2011, 9:29 AM
Did you type the forbidden word above, or did you type "****"?
musicvid10 wrote on 6/17/2011, 10:02 AM
My Kodak PlayTouch shoots 0-255.

This is typical of Pocket HD, Point-and-Shoot, Phones, DSLRs and Hybrids, but not universally so. It's a shame (or maybe a blessing now that I think about it?) that non-Vegas-Pro owners can't quantify and report their results in the same way, since footage from those cameras constitute the vast majority of what is being shot and shared on the internet for the past few years, IMO.
musicvid10 wrote on 6/17/2011, 10:11 AM
NOTE TO EVERYONE:
The Histogram in Vegas is most accurate and gives the fewest superfluous results when the video Preview is set for Best/Full.

This may account for some (but not all) of the noise seen in some of the graphs.
johnmeyer wrote on 6/17/2011, 10:16 AM
The Histogram in Vegas is most accurate and gives the fewest superfluous results when the video Preview is set for Best/Full.Mine was definitely set to Best/Full. I tried Best/Auto, and it definitely spreads out the Histogram.

In fact, these preview settings can produce some very strange results, not only with levels, but also with how adjacent frames are blended. The only setting that produces correct results when doing speed changes is Best/Full. Any other setting -- in particular Best/Auto -- can give you unbelievably strange results, showing blends and ghosts where none in fact will be in the final render.

Slightly OT, I know, but it is an important aspect of using Vegas that not everyone may be aware of.
farss wrote on 6/17/2011, 12:26 PM
"Bob, I'm sure a large proportion of users don't realise that the preview window is not showing them the same levels they'll get in most delivery scenarios, or they just don't really care about it"

I'm quite certain that you're right there although there's been a huge number of threads about this over the years and Glenn Chan has done a considerable amount of work in explaining how to get it right.

The problem is if you start with an incorrect image judgement you can end up with incorrect assumptions down the track.

Vegas's Secondary Preview provides a switch that sets it correctly i.e. expands levels so you'll see your video correctly on a cumputer monitor.

Bob.
farss wrote on 6/17/2011, 12:29 PM
Megabit said:

"How did you use it - it's always greyed-out on my EX1?"

The EX1 does a Black Balance on power up. It can do that with 100% reliability because the camera always has control of the iris. The EX3 gives you the option to do it manually because you can fit lenses that don't let the camera close the iris on the EX3.

Bob.

risce1 wrote on 6/17/2011, 4:28 PM
Nick,, did you use the black stretch with your z1 ?
Nick Hope wrote on 6/18/2011, 3:00 AM
No I didn't. I've never used black stretch actually. But reading about it now, perhaps I should have done for some darker scenes.
amendegw wrote on 6/18/2011, 3:05 AM
Hey, where are all the testers here? If you follow Nick's instructions, it takes all of 15 min / camera start-to-finish (including Vegas time).

...Jerry
LoTN wrote on 6/19/2011, 10:29 AM
I haven't took the time to perform specific tests but according to some footage samples HVR-A1 range is 15 to 253.
Marco. wrote on 6/21/2011, 12:53 AM
Panasonic Lumix GH2, 2011
AVCHD, 1080p24

min black: 15
black peak: 16
min white: 234
white peak: 235

But this is the result only when used the way you described in your first post (color range in GH2 is set to cRGB, don't know if a Adobe RGB setting would affect video recording).

Actually the GH2 behaves exactly according to BT.709 (and BT.601). In real-life recording you will often have peaks below RGB 16 and above RGB 235 and in one or two shots out of hundred shots it may reach from 1-254.
Not illegal though in terms of BT.709 and BT.601 - only RGB 0 and 255 are reserved as control bits. 1-15 and 236-254 are meant as under- and overexposure buffer and it is legal to be used by the camera.
malowz wrote on 6/21/2011, 5:02 PM
small tip:

some cameras have a "border" around image, that may cause the extra "spikes".

try a little zoom or using the border filter to add a fixed color around the image.