Information About The

Digital Video Enhancement System

What is the Asiva^® system?

Asiva is a digital image enhancement system designed to perform up to 22 proprietary functions on still or motion picture video material. The range of enhancements (detailed later) include adding desirable qualities such as grain and the rich color response found in film, scene-to-scene color correction, selective-area image sharpening or softening and a true 24 frame-per-second  emulation without introducing artifacts or motion judder.

Asiva can enhance a variety of source video formats whether they are in the NTSC or PAL standard, single image file formats such as PICT, or computer-based digital video formats like QuickTime®. The input, translation, and output of different image media types are independent from the enhancements performed on them. For example, a QuickTime movie can be opened in Asiva, and after specific enhancements are performed, the enhanced source (which is translated internally to D1) can be output to a variety of video tape formats.

The types of enhancements that are available within the system go beyond giving video the aesthetic of film. Asiva's film-style enhancements are in reality a subset of a much broader capability, and we are careful NOT to put ourselves in such a narrowly defined niche such as occupied other products that claim to give a cinematic look to video material. As an example, it is possible to pull 16mm telecined film into the system and perform certain cleaning functionality to reduce the heavy grain content. It is also possible to bring in 35mm telecined film and perform basic color correction tasks, albeit not as sophisticated as a daVinci system, but also not nearly as expensive. In the future the Asiva system will be able to perform pure NTSC to PAL and PAL to NTSC conversions, and HDTV capability is also targeted for late 2000.

Describe how a videotape is processed with Asiva®

Asiva as an end-to-end process could be thought of as encompassing both the production phase and  the post-production phase of a particular show or project.

Any video may be enhanced, but the overall quality of the finished product may be adversely  impacted if the source video is not clean. While in the production phase, it is best to set-up and shoot  with the Asiva post processing in mind.

What does this mean? In the most simplistic terms, the DP and camera operator(s) should try to minimize the amount of change imposed on the recorded video signal. The use of nets, special lens filters, and on-camera adjustments should be avoided. 

Professional video cameras have many on-camera adjustments that the camera companies have developed to prop up the inherent weaknesses of CCD image devices. Some of these adjustments filter (and remove image information) before the image is recorded. Most of these functions are much more effectively accomplished by our powerful multi-computer/multi-DSP processing system, so we recommend minimal use of these controls. Specifically, color corrections should be limited to basic white balance, and the use of "detailing" should be minimal.

Camera "Detailing" creates artificial edges to give the illusion of a sharper image. This actually removes background information that Asiva could have effectively used to recalculate a true sharp edge, so it is preferable that the detailing be set so that the image is slightly soft rather than highly "detailed." The right amount is so that the image isn't mushy, but not enough so that there are noticeable black edges.

Prior to Asiva processing, it is quite important NOT to modify the source material during the post-production phase. This is the case regardless of whether or not we will receive a rough-cut or a finished piece to work with. For example, an editor performing a basic chroma gain or "gamma adjustment", or a frame synchronizer improperly set up can wreak havoc on our ability to enhance the source properly. Once received, the source material is treated in four distinct steps:

   1. The first step is primarily a mechanical one, where the source material is logically partitioned such that distinct shots may be enhanced separately. This is important for not only maintaining time-code consistency, but for other enhancement capabilities such as shot-to-shot color matching. Using either a customer-provided EDL, or a scene-detector and manual examination of the material, each shot in the material is logged and stored on a digital disk recorder, in 10-bit D1 format.

    2. The enhancement step consists of the Asiva colorist choosing the enhancement tools that he or she needs to perform particular functions, and then carefully adjusting the parameters of each tool  such that a desired result will be achieved. Note that the Asiva colorist may "turn off" any particular  tool or tools at his or her discretion.

       Since the Asiva system is capable of so many different "looks", it is at this stage that the DP or others may sit in with the  Asiva colorist to express their desires. Note that the system has the capability of imparting distinct characteristics on every different shot with respect to the source video, if desired.

    3. Once a project has been set-up, the Asiva system will automatically render the source material to produce the desired characteristics, at which point the system can be left unattended. Rendering takes about 5 minutes per each minute of video.

    4. The final step is to produce some output from the Asiva enhanced material, stored internally in the system. The system currently outputs 10 bit D1 (NTSC or PAL), and the combination of recording decks, D/A, A/D converters, and encoders/decoders that are installed as part of the video interface package defines what video formats can be delivered to the client. This is the back-end of the Asiva system. In addition to supplying multiple video tape formats, Asiva could output  QuickTime, PICS or PICT files that could be delivered on DAT or DLT.

Does the Asiva^® process require the video source material to be of a certain quality or format?

The Asiva system stores and processes everything internally in 10-bit D1 (NTSC or PAL). Analog to Digital converters and other equipment allow virtually any video tape source format (e.g. HI-8, MiniDV, DVCPro, BetaSP, Digital BETACAM, etc.) to be converted to D1 for input to Asiva. We have successfully worked with HI-8 source material on up the format ladder to D1 source.

Some broader-based comments are in order:

       Insofar as a COMPOSITE VS. COMPONENT video format, it is best to avoid composite. Overall, a HI-8 source is preferable to D2 simply by virtue of the fact that HI-8 is component. Avoid scenarios like shooting in BetaSP analog component and running the RGB signal through a frame synchronizer, which outputs a composite video feed to a BetaSP recorder. The reason is simple: crosstalk. Crosstalk can be further delineated by the terms cross-color and cross-luminance. Cross-color results in luminance information being interpreted as color information, and cross-luminance is the result of mistaking color to be luminance information. This crosstalk is visible "garbage" in the video frames, and results from encoding from a component signal or decoding a composite signal to a component signal. Crosstalk can limit the range of enhancements possible with Asiva. The bottom line is to shoot component and stay in component video all along the production and post-production line.

       With regard to the source video capture RESOLUTION, essentially you get out what you give into Asiva. Higher resolution cameras are preferable, of course, for capturing fine detail and NOT being as susceptible to aliasing artifacts (like you get with MiniDV). Asiva does not fill-in or make-up data; it works with what it has. For excellent results, a 750 line (or greater) camera is suggested. (Note that even MiniDV looks great processed, with the exception of the inherent image aliasing due to low camera resolution.)

       The NTSC and PAL VIDEO STANDARDS are not an issue for the Asiva enhancement system, since Asiva can work in either mode!

       Asiva enhancement of desktop-edited material may be successfully accomplished as long as the material is not too compressed. For example, Avid AVR 77 video can be input into Asiva; just make sure that it has been recorded from the Avid in component form.

       Provide test samples early on during the production phase such that any problems can be addressed before it is too late to correct them.

       Don't "mess with" the video signal on-camera, or otherwise before we get the source!

Can the final "look" be tailored to suit MY eye?

YES! Our colorists specialize in the look and feel of film, but as that is a very subjective matter, our clients are invited to supervise setup sessions and build a working and trusting relationship with us!

Can Asiva^® process final-cut material? Transitions? Animations?

YES! Processing the final-cut material saves you money. Most transition effects are handled quite easily, as our process treats  each clip separately. Animations are very effectively enhanced without changing the original artist's colors!

Should I Use Special Camera Techniques?

The best results will be seen when Film-style lighting with attached shadows is used rather than the flat, "toppy" lighting characteristic of most video taping styles. The separation of foreground objects from the background by reducing the depth of field is desirable. Camera setup should include setting white balance and color temperature, low detailing and masking off.

And it goes without saying, try to avoid blown-out over-lit video!

Does the Asiva^® process meet the requirements for HDTV? Will productions processed through Asiva^® be compatible with HDTV?

NTSC or PAL material that has been processed with Asiva software can be upconverted to HDTV as well as can any normal 4:3 broadcast material. At present, material that is originated in HDTV formats cannot yet be processed by the system; however this is a planned future capability. The Asiva software algorithms are format-independent, but accommodations for user interface are needed, and the incorporation of the various HDTV aspect ratio, resolution and frame rate combinations has to be done. With these modifications and the addition of appropriate front-end hardware and monitors, Asiva should be ready for HDTV in late 2000.

Can the Asiva^® technology work with the new DVD technology and will Asiva^® hold up under MPEG2 compression?

Absolutely. We have successfully tested MPEG2 on all forms of Asiva output and can state that the results were highly satisfactory. Our frame rate emulation does not throw out frames or duplicate fields or use other techniques that interfere with the MPEG2 compression algorithms. Our demo "road show" is being produced on DVD.

What ELSE can Asiva® do?

The Asiva system can do selective sharpening and or softening of images, color-timing and matching of telecined film to intercut video, and a number of color-based special effects. In addition, a side-benefit of the Asiva system is that it automatically corrects out-of-gamut color problems in source video, as well as insuring that the output IRE is perfectly within spec. The process can often bring inferior video up to broadcast standards.