Chris Bowman

Oh, and the Canon D5 Mark II, which Peter mentioned, is a digital still camera that is also capable of 1080p @ 24FPS. Since it uses a sensor that is the same size as a full frame 35mm and uses full frame 35mm lenses, it has a depth of field very similar (identical?) to working with a 35mm still camera, and slightly shallower than a 35mm motion picture camera.

Correct, you cannot run uncompressed from the XH-A1, since all of the output signals are generated after compression. Is getting an uncompressed signal worth the extra $3,000 for the G1, plus the extra equipment you need to capture and edit it, plus the hassle of carrying that equipment with you as you shoot? I don't know. If you intend to finish this to 35mm, than probably yes. But if you can afford a film-out, you probably wouldn't be buying a $3,000 video camera. To be honest, I've never had much problem with panning my A1. In 1080i it's smooth as silk, in 30F and 24F the limitation is more the frame rate than compression artifacting, and you would have that same limitation on any camera shooting at 30 or 24 frame progressive. The thing that is most difficult for the codec is lots and lots of fine detail in a shot, especially if it's not all moving in the same direction. An example might be a shot of a tree with the leaves all shaking in a strong wind.

SDI is Serial Digital Interface. It's major advantage is that it carries a signal generated before in camera compression. This can be very advantageous because compression is one of the major limiting factors on image quality and what you can do to the image in post. You can use SDI to record to a stand alone recording deck with much lower compression and better color space, or even capture uncompressed video directly to a properly equipped computer, rather than recording a compressed signal to DV tape in HDV format. Gen-Lock is used to synchronize multiple cameras for live switching and mixing. Both of SDI and Gen-Lock use standard coaxial cable with BNC connectors. I did make a mistake in my previous reply to Saul. The format that has been used in wide release films is HDCAM SR, not DVCPro (which is used extensively in TV production). HDCAM SR has a maximum resolution of 1920x1080. Let us know about the kind of filming you are thinking about doing with your camera, your needs should determine the tools you buy. Are you shooting high action? Dark scenes? Do you need long record times? The more we know, the better advice we can give you.

Saul, I've never even heard of a Canon 1920i HDV codec. Canon advertises their XL and XH series cameras as 1080i, with an actual chip resolution of 1440x1080 on each chip. This is the maximum resolution of the HDV codec, and anything higher would not be considered HDV, as it would be 100% incompatible with the HDV codec standard. This is also the maximum resolution of the DVCPro HD format, which has been used to shoot ultra-wide release feature films, so it's not a terribly limiting factor. The limited color space and relatively high compression ratios of HDV are far more important concerns. In the $3000 market range, however, HDV is tough to beat. Most of the AVC variants in this price range use too much compression to really match HDV, and have their own slew of compatibility issues, as well as requiring much more power (or time) for editing and post. Most of the other options are simply two expensive to consider at this price point.

Well Mark, the short answer is that it depends on what you want to do with it. Since you haven't said, I can't even begin to tell you what camera is the best for your budget. I can tell you a little about the Canon cameras, though. As you already know, the XH-A1 is a 3CCD 1080i HDV camera. It can also shoot in 30F and 24F modes. The most probable explanation of how the F modes works that I have found is that the green chip is run with inverse phase, so that when the red and blue chips are reading the A field, the green chip is reading the B field. This essentially means that the camera is shooting a Bayer pattern in F modes and 3CCD in 60i mode. Bayer patterning does reduce resolution and perceived sharpness slightly, so if maximum sharpness progressive scans are desired you may want to look elsewhere. Another downside is that XH-A1 cannot shoot progressive at rates above 30 FPS, which limits you greatly if you are planning to shoot slow motion. The HDV codec is a further limitation, it introduces compression artifacts and limits the color space to 4:2:0. Canon's 24F mode uses a proprietary version of the HDV codec that uses 20% lower compression, but requires you to capture to your editing system using a compatible Canon deck (usually the camera you shot it on). Unless I am mistaken, Avid still hasn't gotten around to fully supporting Canon 24F HDV, although there are workarounds and they say full support is coming. Other editor platforms seem to handle it just fine though. All that said, the XH-A1 is a very impressive camera, especially since it can be acquired for ~ $3,000. Since you already have the Letus adapter for the XH-A1 (yes it's a 72mm), you should definitely consider it carefully. The first difference between the XH-A1 and the XH-G1 is that the G1 has a "professional jack pack" which includes uncompressed SDI out (interlace only), and Genlock in/out. The other difference is that the G1 costs about twice as much. Both cameras are also very similar to the XL-H1, except that they are fixed lens and have slightly more customizable settings. If you absolutely need the SDI and Genlock the XL-H1 is probably a better buy than the XH-G1. Hope this helps

Those pulsating colors are from the rotating color wheel on projectors that are not 3-chip designs. The wheel spins to filter the single chip image through each of the primary colors, and usually a clear section to enhance brightness, for every frame of video. You may or may not be able to adjust the shutter speed to match the color wheel strobe, but if your camera can do clearscan that may help. The other option is to shoot 30p or 24p (assuming you can) rather than 60i. The longer exposure should give enough image persistence to overcome the color switches. Another important consideration is that many LCDs and EFVs do not refresh fast enough to reveal if the color pulses are occurring. Make sure you use a color CRT monitor to check the projector for color pulses, as the LCD and EFV may show a perfectly good picture, but you could still wind up with rainbow-vision in post.

Meant to say: "The cleanest images are captured with the gain set to -3dB."

Well, I wouldn't call myself an expert, but this might help. The cleanest images are captured with the noise set to -3dB. I have found that noise is usually not a problem at the 0dB and +3dB as well. 6dB is sometimes acceptable, depending on the scene, but after that noise just becomes unacceptable for most purposes. Be sure that the auto gain switch is set to off, and don't shoot in auto or "easy" mode (green box setting) as these will ignore your gain setting. If you are really getting a lot of noise with the gain at 0dB, it might be the way that your preset is configured, try changing to a different custom preset. You can download them (Here are a few to get you started) or tweak one of your own. You can store up to 30 on a SD card (the camera won't recognize more than 30) If you do this and still can't get an acceptable image, there might be something wrong with the camera.

From what I understand, CRI has a direct correlation with green spikes, and any spike in general. There is a fairly good explanation of CRI and fluorescent lights here

I did a no budget project once where we used a few 500 and 1000 watt halogen shop lights from the hardware store. They were frustrating to work with because even distribution was not even an afterthought in the design of the reflectors. They produce a very uneven light that is extremely hard to get satisfactory results with. I would not use them again if I have a choice. They were just something that was on-hand in a production that could barely afford the DV tapes it was shot on. DIY lighting is really not worth it. To do it right, it costs almost as much as just buying the finished product, while consuming untold hours of your time. If you are looking to keep it cheap, your best bet is probably ebay, auctions, and estate sales (it's really remarkable what you can find at an auction or estate sale, often for a tenth of what it's worth).

Of course you are right, making sure everyone is on the same page before the tape starts rolling is the best way to go about it. I just realized that my profile info said that I was sound dept. instead of student. This was unintentional, and I apologize for any misdirection that it may have caused. I'm not trying to give my own comments any more weight than they deserve, and on this topic especially I may have appeared to be speaking with authority on a topic where I am not.

As I recall, Adobe released several post-release patches for both CS3 and CS4 to improve compatibility with AVCHD/H.264. Go to the Adobe site and make sure you install all of the patches. I have also heard that installing the CoreAVC decoder helps greatly in improving h.264 compatibility for both Premiere and Windows Media Player. I believe it costs ~$15 USD

Honestly, I'm a little confused about your whole question on sound frame rates. Sound recordings aren't directly linked to image frame rates in digital video. For example, HDV specifies that audio be sampled at 48kHz regardless of the video frame rate (44.1 kHz is also permitted). If your sound mixer is overlaying the sound on to the video, it should always be done with the video its native frame rate. If you shot in 24F, you should be editing in a 24p timeline, and sound should be layered onto the footage in a 24p timeline as well, otherwise it will not sync exactly because the 3:2 pull down of the video will be slightly off cadence with the final 24p footage. As for Final Cut Express not having 24P, any footage captured from a firewire port should still be an exact duplicate of the original data recording on the tape. If the software can capture your footage from your deck, it has imported 24p footage, and any pull down is added by the software. The file captured to the hard drive should be 24p regardless of what modes Final Cut Express supports, but it should only be edited and exported in a 24p timeline, which it sounds like it you will need another program to do.

To clarify, the Canon XH/XL series cameras clock the CCDs to 24Hz, or a multiple thereof, in 24F mode (according to the manual). The HDV recording deck then takes this 24fps signal and records it as 24fps HDV without pull down. All signals off of the camera except the DV/HDV port (FireWire) are 60i, and the 3:2 pull down from them gives a somewhat disturbing off cadence jitter (which is what I suspect most people complain about), but the actual digital recording has a smooth 24 frame cadence. Canon 24F HDV recorded tapes do NOT play nice with other decks, so a Canon HDV deck must be used for capturing to the editor, but once the computer has ingested it, there doesn't seem to be any difference from 24P.

As I understand it, Canon used CCDs that were designed to be interlaced. Since the chips can only read out every other line to a frame buffer in a single pass, Canon's solution was to attach two frame buffers, with the A field reading to one and the B field reading to the other simultaneously. The fields are then combined in the processing chip. Since this is not truly "progressive" (i.e. reading every line concurrently to a single frame buffer) Canon could not call it 1080P*. Theoretically, if this is true, there should be no visible difference between 1080F and 1080P at any frame rate. In reality, some people swear that 1080F sucks, while others attest that there is no visible difference. All I know for sure is that 1080F on my XH-A1 gives visibly better resolution than 1080i, and that the frame rate goes all the way down to 24fps. The exposure levels indicate that 30F is not really 60i, nor is 24F really 48i. If they were really the same, the exposure levels would be the same. Instead, there is significantly higher exposure in the F mode exposures, just as you would expect in a "progressive" scan at that frame rate. I know for a fact that 60F and 60i give the same exposure level on my Canon, while 30F is much brighter. *This information was something I read on another forum months ago. I can't find the original thread, and I don't know what evidence the original poster had to back up this info.

Thanks for the info, Peter. I'll try some test footage in HDV and see what happens.

Someone I was talking to recently claimed that my XH-A1 would output pre-compression 1080i in 4:2:2 from the component video out. I haven't been able to locate any documentation that either confirms or refutes this. I have a friend who wants to shoot some chroma key scenes, and I was wondering if I could get better results from the component outs to a Blackmagic Intensity Pro card and recording uncompressed or with a low compression format? For that matter, would I be better off shooting most of my footage that way? (Assuming I can haul around a PC and all)

The answer is that the camera can adjust the volume much faster than you can. Any time the volume exceeds the 0dB level, the volume is in excess of a digital recorder's ability to reproduce fully, and clipping occurs. Auto volume controls are designed with compressors which decrease the gain on the signal very rapidly as volumes approach the 0dB level so that the sound gracefully reaches the the loudest sound that can be faithfully reproduced but does not exceed it. If the sound manages to build quickly enough that it outruns the compressor and exceeds max levels, the electronics continue reducing the gain, to produce as little clipping as possible. As soon as the signal level has dropped down again, the compressor begins to raise the gain up so that even quiet signals are recorded at high volume, this creates a desensitizing effect that makes minor clipping less noticeable, but has a negative effect if you are trying for broad dynamics in your sound. Manually adjusted volumes that hit that 0dB level tend to blow right through it, and the sound clips and distorts. If the gains were set so that the volume maxed out exactly at 0dB there would be no problem, but you can't respond fast enough to turn down the gains as the signal approaches max , while the electronics can and do respond in milliseconds. Remember, the volume may be reading 0db, but if it is less than the next level (say +3dB) it can still be above 0dB it will not show the excess, but will clip noticeably.