Home | Who We Are | The Sound Manifesto | Agreements | Projects | Links | Contact Us
|
Canon HV30 Exposure Calibration
by Eddy Grabczewski
INTRODUCTION
This article describes how to calibrate the exposure of a Canon HV30 camcorder. The result is a useful exposure chart for cinematographers and videographers who need to achieve a film look to their video by manually controlling HV30 exposure.
I performed all tests with a Canon HV30 PAL model, so the exposure chart is limited to 1/50 sec. exposures.
NOTE FOR NON-TECHNICAL READERS
To avoid the technical detail, scan the article quickly to pick-out any useful points and read the Conclusion. At the end of the day, you only need the exposure chart and instructions.
EXPOSURE CHART INSTRUCTIONS FOR A HV30 (PAL)
- Set your light meter to 1/50 sec. (or 25 fps at 180 shutter angle) and ISO 80 (Tv 50) or ISO 100 (CINE MODE).
- Using the light meter, take an exposure reading of your subject.
- Set the camcorder mode to either 25p Cinema Mode (PAL), Tv 50 (PAL).
- Point the camcorder at your standard light source and take an exposure reading using the joystick and joystick menu.
- Semi-press the PHOTO to find the current aperture and shutter speed.
- Using the exposure chart, find the current exposure parameters.
- Move the joystick left or right to choose the required exposure.
A Director of Photography does not take light readings with a light meter and simply transfer these to the camera. There's a whole world of artistic choices that need to be made first. If you want to know more, take a look at some good books on photography and cinematography. I've mentioned a few in the Bibliography.
EQUIPMENT REQUIRED FOR CALIBRATION
As well as the Canon HV30 camcorder (firmware version 1.0.1.0. I), you'll need a light box, light meter, Kodak grey card and white card (which is on the reverse of my grey card). You'll also need to download some software called HDV Data Monitor and install it on a Microsoft Windows computer with Firewire interface.
There's much more detail in the body of this article and References..
NTSC AND PAL MODELS OF THE HV30
The Canon HV30 PAL model nominally runs at a frame rate of 25 fps in 25p Cinema Mode and was tested at a shutter speed of 1/50 sec in Tv 50 mode. The DV video tape records at 50i regardless of whether you've chosen an interlaced or progressive mode.
The Canon HV30 NTSC model nominally runs at a frame rate of 24 fps in 24p Cinema Mode and should be tested at a shutter speed of 1/48 sec in Tv 48 mode. The DV video tape records at 60i regardless of whether you've chosen an interlaced or progressive mode.
EXPOSURE
Film Speed
Photographic film has a sensitivity to light, the film speed, that's expressed as an ISO rating. If you want a more sensitive film, you'll need to buy another roll with a higher ISO rating. The trade-off with increasing film speed is more grain in the final picture.
Chip Gain
Unlike film, a camcorder can vary its sensitivity to light by amplifying the electronic signal output from the CCD or CMOS chip - no need to switch chips for a faster one! The trade-off with increasing the chip gain is more noise in the final video.
The HV30 uses up to +27 dB of gain to vary the sensitivity of its CMOS chip but there's no way the user can control or display the gain.
The HV30 applies gain only at the following f-stops: f/2.8, f/2.6, f/2.4, f/2.2, f/2.0 and f/1.8.
As an aside, the human retina is more adaptable than either film or chip. Every "pixel" can adjust its own "gain" - one reason why the human eye has an apparently huge dynamic range.
Aperture, Shutter Speed
Both film and video cameras need to control the amount of light falling on the recording medium. Both camera types have an aperture and shutter, whose size and speed control exposure to light.
ND Filters
It's common practise with film cameras to manually place ND (Neutral Density) filters in front of the lens to deliberately reduce the amount of light falling on the medium. Video cameras have some ND filters built-in, but whereas professional camcorders allow the operator to manually switch ND filters, consumer models are designed to do it automatically.
The HV30 does have inbuilt ND filters which it controls automatically. A Canon patent describes exactly how it works, particularly the "third embodiment". I guess the HV30 uses a circular ND filter with ten gradations of optical density ranging from 0.15 to 1.50 in increments of 0.15 and operating only at f/5.6 (Movie Camera mode). But, as with gain, there's no way to control these values or display them.
To summarize, there are four exposure parameters we'd like to control manually on any movie camera in order to manipulate shots both technically and artistically. These are:
The Canon HV30 is marketed as a consumer HDV camcorder. As such, it doesn't provide independent control of all four parameters (a professional camera would). Instead, the HV30 allows the user to manually set either the aperture or the shutter speed, but not both at the same time. What's more, it doesn't allow the user to set chip gain or ND filtering. As I said earlier, the camcorder won't even display their values!
Given these hopeless limitations, it's taken the wit of the worldwide community to tame the HV30 into manual submission. Barry Green first proposed a method for setting these four parameters manually on a HV20. His article inspired me to write this one.
MEASURING EXPOSURE IN THE HV30
In this section, I describe a method for finding the four exposure parameters of the HV30 and to derive an exposure chart showing all four parameter settings for a range of exposure values. To do this, I make the following three assumptions:
- You're trying to give your video a film look.
- You want to eliminate chip gain to prevent video noise spoiling your image.
- You want to retain the motion charactersitics of film by keeping the shutter speed at 1/50 sec. (European film) or 1/48 sec. (American film)..
But why choose a shutter speed of 1/50 sec? European film runs at 25 fps, which means the shutter speed should be set to 1/25 sec. shouldn't it? The reason is because a movie camera uses a half-moon shutter (shutter angle is 180 degrees) to expose the frame to light for only half the frame interval; during the unexposed period the next film frame is positioned by a pull-down claw. So, instead of the film being exposed for 1/25 sec. it is actually exposed for half that time, namely 1/2 x 1/25 = 1/50 sec. An analogous argument applies to American film and 1/48 sec.
As an aside, to convert Shutter Speed (in seconds) to the equivalent Shutter Angle (in degrees), use the following equation:
Shutter Angle = Shutter Speed x Frame Rate x 360
Example: 180 = 1/48 x 24 x 360
You can transform this equation to find the equivalent Shutter Speed:
Shutter Speed = Shutter Angle / (Frame Rate x 360)
Example: 1/48 = 180 / (24 x 360)
To measure the four exposure parameters, I needed a suitable light source - which turned out to be a small 4 x 5 inch light box.
To measure the quantity of light emitted from the light source, I used a light meter. This particular model is a favourite of cinematographers, allowing them to take exposure readings by directly setting the frame-rate and shutter-angle. If you have a standard photographers light meter then use the equation above to find an equivalent shutter speed. However, in this case I used the lumisphere in a lowered position to take readings of illumination, measured in Lux.
Measuring the quantity of light emitted from the light box was tricker than I'd expected, because the light was not very evenly distributed across the white viewing screen. In the end I chose three areas - the first emitted 2100 Lux (190 fc) near the centre of the screen, the second emitted 1300 Lux (120 fc) slightly off-centre and the third emmited 600 Lux (55 fc) near a corner of the screen. I chose these values because I found they gave practically ideal exposure charts.
I also needed to measure the aperture (f) and shutter speed (s) of the camera at any given setting. This was done by the well-documented method of buying a miniSD card for the HV30 and semi-pressing the PHOTO button to display the aperture and shutter values on the camcorder's LCD screen. The HV30 was in TAPE MODE with the Camera Mode switched to P (Programs).
Finally, to measure chip gain (in dB) it was necessary to download and install HDV Data Monitor onto a Windows XP computer.
Unfortunately, I know of no way to measure the internal ND filtering of the HV30, so I had to make a calculated guess. There's more on ND filtering in References.
CALIBRATING THE HV30
There are two important menu settings to understand when working the HV30 manually: Recording Programs and Video Standards. To be honest, I'm still not clear where the dividing line is drawn between these two areas, particularly with regard to shutter speed. Sometimes the Recording Program determines shutter speed (as in Tv 50); at other times it seems that shutter speeds are determined by the Video Standards (as in HDV25).
Recording Programs
Recording programs work at the camera end of the camcorder. They control the four exposure parameters (chip gain, aperture, shutter speed and ND filtering) in different ways, depending on the subject and surrounding light. A subject photographed in snow requires different exposure parameter values to one taken at night.
The recording programs we're interested in are: Program AE (P), Shutter Priority (Tv) and CINE MODE.
Program AE (P) recording program
This program allows the camcorder to choose the aperture, shutter speed, chip gain and ND filtering automatically.
Shutter Priority (Tv) recording program
This program allows the user to set the shutter speed; the camcorder automatically chooses the aperture, chip gain and ND filtering.
CINE MODE recording program
This program varies the exposure parameters when necessary. It also changes the image colour and contrast, to create more of a film look, by applying a Gamma Curve that mirrors the Charactersitic Curve of film, with the addition of knee and black stretch segments to mimic the mountain and valley of the Charactersitic Curve, thereby approximating the highlight and shadow details of film.
It's interesting comparing CINE MODE and Program AE (P) programs using a Kodak colour chart and grayscale. CINE MODE turns violet colours to blue (meaning a loss of red, considered a typical characteristic of film) and there's an obvious loss of contrast and brightness too, but highlights and shadows have a bit more detail. Several users have noticed a loss of sharpness in the image however, this may depend on the aperture. Not all of these are bad. For example, you don't want high contrast or saturated colours if you're recording video for film. Grading is the right place to modify colour and contrast - at the end of film processing.
Video Standards
Video standards work at the recorder end of the camcorder. They allow you to select different video output formats, which change at least the following parameters: vertical resolution (1080, 625 or 525 lines), aspect ratio (16:9 or 4:3), field rate (50 or 60 fields per second) and recording system (interlaced or progressive).
The video standards we're interested in are primarily HD standards: HDV, HDV25, HDV24 and HDV30. Before we launch into a description of these standards, let's look at some chip scanning concepts:
Progressive Frame
To record an image, the camcorder chip is scanned once. A single complete scan of the chip is called a progressive frame and represents a complete image recorded at a single point in time. Progressive scanning starts at the top left-hand corner of the chip image and proceeds from top to bottom, left to right, until you reach the bottom right-hand corner.
The chip scan is so fast, it could record several thousand images per second but the human eye only needs about 50 for smooth, flicker-free motion.
The HV30 CMOS chip gives a HD image comprising 1080 lines, where each line contains 1920 pixels. The HDMI interface on the HV30 outputs a high-level HDV resolution (1920 x 1080), however the Firewire inteface and DV tape recorder reduce this to a lower resolution (1440 x 1080).
Non-progressive Frame
A non-progressive frame is a sequence of two progressive frames. It comprises a composite image scanned at two successive points in time.
If the first progressive frame is modified to contain only odd lines and the second one only even lines, interlacing both progressive frames results in a non-progressive frame.
It's important to note, the two progressive frames making up a non-progressive frame represent different images at different points in time.
Segment
A progressive frame may be broken-down into two segments. These two segments are called the odd segment and even segment; they're the result of a single scan of the camera chip. If the odd segment contains only odd lines and the even segment only even lines then by interlacing both segments we can recreate the original progressive frame.
Each segment carries only half the information of its parent progressive frame. The two segments are transmitted consecutively and then interlaced to re-constitute the original progressive frame.
Note, the two segments comprising a progressive frame represent exactly the same image at exactly the same point in time. This is a crucial difference between a segment and a field. Figure 1 illustrates the concept of a segment.
Since each segment contains half the information of the progressive frame, it needs only half the bandwidth of the the full frame but two such segments require twice as long to transmit.
Field
A non-progressive frame may be broken-down into two fields. These two fields are called the odd field and even field; they're the result of two consecutive scans of the camera chip. If the odd field contains only odd lines and the even field only even lines then by interlacing both fields we can recreate the original non-progressive frame.
Note, the two fields making up a frame represent different images at different points in time. This is a crucial difference between a field and a segment. Figure 2 illustrates how the concept of a field relates to that of a segment.
Each field carries only half the information of its parent non-progressive frame. The two fields are transmitted consecutively and then interlaced to re-constitute the original non-progressive frame. Since each field contains half the information of the non-progressive frame, it needs only half the bandwidth of the the full frame but two such fields require twice as long to transmit.
Example: TV engineering created the definitive field. The odd and even fields of a TV broadcast are progressive frames.These fields are scanned at two different points in time and each field represents a slightly different image. In a standard TV, these fields are never really interlaced into a single image but are transmitted rapidly in sequence to give the impression of smooth motion. If these fields are interlaced into a non-progressive frame (as in video editing) then you must expect to see motion artefacts.
Interlaced Frame
An Interlaced Frame is a non-progressive frame created by interlacing odd and even fields. Remember that non-progressive frames comprise fields that were scanned at different points in time. As a result, when these fields are interlaced, it's common to see motion artefacts.
The importance of this technique lies in conserving video transmission bandwidth. Interlaced frames can also be stored on a DV video tape recorder (DVTR) in a manner illustrated by Figure 3.
Now you can see what motivated the distinction between progressive frames and non-progressive frames, and why the term "interlaced frame" isn't adequate to describe a non-progressive frame - because progressive frames and non-progressive frames are both interlaced.
Progressive Segmented Frame (PsF)
A Progressive segmented Frame is a progressive frame created by interlacing odd and even segments. Remember that progressive frames comprise segments that were scanned at the same point in time. As a result, there are no motion artefacts.
The importance of this technique lies in the utilization of interlaced video recording technology to store progressive frame video. PsF frames can be stored on a DV video tape recorder (DVTR) in the same manner as for Interlaced frames. This is illustrated by Figure 4.
Digital Video Tape Recorder (DVTR)
Given the similarities between interlacing fields and segments, the HV30 uses the same hardware to do both. The DVTR encodes 50i PAL or 60i NTSC frames and stores then on DV tape. The data stored for a field and a segment are the same; the only difference lies in the temporal interpretation of that data by the hardware and software.
HDV is a video standard with vertical resolution=1080, aspect ratio =16:9, field rate=50 (PAL) or 60 (NTSC) and recording system=interlaced. Each camera image is recorded as a non-progressive frame, comprising an odd and even field. The two fields are then interlaced in 50i (PAL) or 60i (NTSC).
HDV25 is a video standard with vertical resolution=1080, aspect ratio=16:9, field rate=50 and recording system=progressive. HDV25 emulates. European film recorded at 25 fps (film is always projected at 24 fps). Each camera image is recorded as a 25p progressive frame, comprising an odd and even segment. The two segments are then interlaced at 50i and stored in the 25PsF video format.
HDV24 is a video standard with vertical resolution=1080, aspect ratio=16:9, field rate=60 and recording system=progressive. HDV24 emulates American film recorded at 24 fps (film is always projected at 24 fps). Each camera image is recorded as a 24p progressive frame, comprising an odd and even segment. The two segments are then interlaced at 60i with pulldown applied and stored in the 29PsF video format. Before film editing or film recording, a reverse pulldown is applied to retrieve the original 24p camera frames.
HDV30 is a video standard with vertical resolution=1080, aspect ratio=16:9, field rate=60 and recording system=progressive. HDV30 does not emulate film since it records at 30 fps. Each camera image is recorded as a 30p progressive frame, comprising an odd and even segment. The two segments are then interlaced at 60i and stored in the 29PsF video format. The resulting video is easily edited to create DVDs or Web content since no pulldown or reverse pulldown is necessary as in HDV24.
Cinema Mode
The Canon HV30 Instruction Manual defines Cinema Mode to be a particular combination of recording program and a video standard:
25p Cinema Mode = CINE MODE + HDV25 PAL 24p Cinema Mode = CINE MODE + HDV24 NTSCThe HV30 also allows other interesting combinations that have no name: CINE MODE + HDV, Tv 50 + HDV25, Tv 48 + HDV24 and Tv 60 + HDV30.
Calibrating the HV30
Test 1 - Tv 50 + HDV
In the first test, I configured the HV30 with the recording program set to Tv 50 and the HD standard set to HDV. I placed the camera lens onto the light box at the 600 Lux position and locked the exposure. As I passed from -11 to +11, I recorded the aperture, shutter speed and gain. The result is the following chart (the first four columns):
The procedure was repeated using the 2100 Lux position. This resulted in the following chart (the first four columns):
Finally, I pointed the camera at the brightest light source I could find, which turned out to be a 20 W low-energy light bulb at a distance of 1 inch, which gave 10000 Lux (950 fc). This resulted in the following chart (first four columns again):
I then combined all three results into the following exposure chart:
In the Gain (dB) column, I've highlighted in red where gain may be added by the camcorder. Similarly, the ND Filter (f) column is highlighted red for calculated ND values at a fixed aperture of f/5.60. Note how the values in the Total Aperfure f column rise steadily - as you might expect.
The extra columns in these charts can be ignored unless you're interested in examining the formulae in the spreadsheet from which all the exposure charts were derived:
The internal ND filter value expressed as an f-stop The relationship between aperture (f) and density (D) is f = 0.3 x D.
Test 2 - CINE MODE + HDV25
In the second test, I configured the HV30 with the recording program set to CINE MODE and the HD standard set to HDV25. I placed the camera lens onto the light box at the 600 Lux position and locked the exposure. As I passed from -11 to +11, I recorded the aperture, shutter speed and gain. The result was exactly the same as for my first test at 600 Lux.
I then repeated this procedure but this time using the 1300 Lux (120 fc) position. This gave exactly the same result as for my first test at 2100 Lux.
When it came to using a 10000 Lux (950 fc) source, the HV30 deviated from the results of Test 1. At all apertures smaller than f/8.0 at 1/50 sec. the camcorder began using smaller shutter speeds, starting with f/8.0 at 1/60 sec. and ending with f/11 at 1/500 sec. Clearly, given our previously stated assumptions, CINE MODE has it's limitations for digital filmmakers.
Film Speed
The next investigation was to find an equivalent film speed for the HV30 camcorder when shooting in Shutter Priority and Cinema Modes.
The HV30 was white-balanced to a Kodak white card (90% reflectance). After setting the camcorder mode, I exposed the shot to a Kodak grey card (18% reflectance).
The light meter was set to incident mode, meaning it uses a white-domed diffuser to collect light as if it were reflecting off a Kodak grey card (18% reflectance). Since the meter does not support the same f-stops as on a HV30, I had to choose the nearest values. The results were as follows:
Shutter Priority Mode (Tv 50 + HDV): Camera: f/4.0 Light Meter: f/4.08 at 25 fps (180 degree shutter angle) at ISO 80 25p Cinema Mode: Camera: f/5.2 Light Meter: f/4.09 at 25 fps (180 degree shutter angle) at ISO 100The combined exposure chart applies to the Canon HV30 in Shutter Priority Mode and in Cinema Mode up to f/8 - beyond that, the chart only applies to Shutter Priorty Mode.
The equivalent film speed of the chip, without chip gain (0 dB), lies between ISO 80 and ISO 100. At such a low film speed (even by film standards), the HV30 needs at least 600 Lux (55 fc) to open-up to f/1.8. Moviemakers will probably need artificial lighting to shoot indoors - a single 1 kW tungsten lamp radiates 600 Lux at 12 feet.
As a final tip, should you find yourself without a standard light source to calibrate your exposure settings, simply switch-on the LIGHT button and point the lens approximately 6 inches away from a white sheet of paper. This should give you enough light to set the aperture to f/1.8 at the +11 exposure setting. If you need a smaller aperture, point the lens at a brighter light source.
- Canon HV30 Instruction Manual PAL NTSC
- Canon HV30 Exposure Calibration PDF file
- Canon HV30 Exposure Calibration Excel 2007 spreadsheet
- Controlling Exposure on the Canon HV20 by Barry Green
- Exposure Control, Canon Style by Michael Jouravlev
- Progressive Modes Explained by Michael Jouravlev
- HDV Data Monitor software by elCutty of Videotreffpunkt
- Controlling Aperture and Gain video by Patrick Jennings
- Canon HV30 ND Gradation System on Canon's web site
- Canon HV30, HV20, HV10 forums on DVInfo
- HV10 - ND Gradation Feature forum on DVInfo
- Canon's Light Control and ND Filtering patent
- Pixel Size, Aperture and Diffraction by Sean T. McHugh
- Gamma, Knee, Black Stretch and Colour Temperature article in DV magazine
- Canon HV20 24p Pulldown
- Dynamic Range in Digital Photography by Sean T. McHugh
- Canon HV30 Exposure Calibration post on DVInfo forum
- Canon HV30 Exposure Calibration post on HV20/30 forum
- The Camera by Ansel Adams
- The Negative by Ansel Adams
- The Print by Ansel Adams
- Digital Photography Tutorials by Sean T. McHugh
- Film Production Technique by Bruce Mamer
- Lighting for Film & Electronic Cinematography by Dave Viera
- Practical Cinematography by Paul Wheeler
- Digital Cinematography by Paul Wheeler
- Sony HVR-Z1E and FX1E paper by Alan Roberts, BBC
