Videotape Program Production at CBS Studio Center By WILLIAM G. CONNOLLY The one-inch helical videotape recorder has been utilized by CBS to configure a film- plitude modulation and phasc modulation style (one VTR per camera) videotape production system at CBS Studio Center in do correlate well, particularly with PM Hollywood. Operating experience with these machines is discussed, with emphasis on noise, which is a measurc of color distor- observations of video, audio, and transport performance. I n addition, this paper tions. Subjectivcly, the results are better describes a unique system for synchronization of one-inch VTRs in post production that than that indicated by a simplc review of has allowed CBS to offer an alternative to traditional multi-camera videotape production signal-to-noise curves. Coherent picture methods. disturbances such as banding and first line error are usually more detrimcntal in Introduction that othcr recorders built to the SMPTE dubbing operations than signal-to-noise Type C standard will be capablc of equiv- loss. A nonsegmcntcd format is, of course, In February of 1975, CBS issued target frcc of those errors. specifications for the design of a videotape alent performance. The pcrformancc data were originally recorder with improved editing capabilities. Audio Performance Thc thrust of these specifications was for measured in mid-I976 on three pre-pro- duction prototypes. Machincs of later As can be secn in Tablc 11, frequency a recordcr with vastly improved transport response and signal-to-noise ratio approach characteristics, coupled with a tape format production runs have been compared and thc data proved to be typical of the 23 units that of studio-quality audio recorders. which would permit the display of viewable Harmonic distortion of 0.7% and wow and pictures at a varicty of speeds in both for- in service or currently being installed at CBS. flutter mcasured at 0.03% promise excel- ward and reverse. The intended application lent results. These results arc in large was for off-line cditing systems from which Video Performance measure due to a unique push-pull dual- a work copy would bc produced. The capstan design. camera originals would later bc conformed Typical video performancc is shown in Tablc I. All thc pcrformance data shown Thcrc arc three audio tracks in this for- through automatic assembly in an on-line mat, with Audio I and Audio 2 separated editing systcm. compare favorably with quadruplcx rc- corders, particularly in differential gain, by 0.8 mil (0.2 mm). These tracks can bc I t had bccome clear at that time that uscd to rccord sterco or independent audio computcr-assisted videotape cditing facil- phasc and moirt, but notc especially the improved interchange performancc. A signals. I n the latter case, crosstalk of -38 ities, such as thosc pioneered by CMX, had dB at 50 Hz and -50 dB at 100 Hz may become limited, not by software, but typical signal-to-noise ratio of 49.5 dB, coupled with an interchange loss in S N R seem unacceptable to the audio purist. through the inability of videotape recorders Howcvcr, the ASA “A” weighting curve to respond efficiently to commands without of only 0.2 dB, gives rise to excellent mul- tiple-generation performance. First-gen- comes to our aid at these frcqucncics. Lis- continual breakdown. Moreover, the cost tening tests confirmed the acceptability of of intcrfacing various VTR types and cration quadruplex performance is typi- cally 47 dB and interchange loss is 2.0 dB. the 0.8-mil separation; tcsting included models to computcrs had become an im- hard rock music passages which evidcntly pediment to economical dcsign. Clearly, This means that a tape recorded on one VTR and played back on another (the most have bccomc the acid test currently in favor VTRs with TTL-compatible command in the audio world. systcms were needed. usual kind of operation) would, typically, have an S N R of 45 dB for quad and 49.3 VTR manufacturers had also identified Transport Performance this problem. Less than six months later the dB for hclical. In other words, thc 1-in he- lical product is “up” more than one gen- Tape transport performance meets a first of scvcral laboratory prototypes was very high standard. Variable-speed search shown. The following year brought sevcral eration, compared with a quadruplex re- cording. is controlled by a dial with still framc at a more prototypes to light. More than centcrcd dctcnt. Clockwise rotation of the mccting many of thc criteria for an cditing Figurc 1 shows plots of noise perfor- mance versus number of generations for a dial moves the tape forward in increments systcm machine, howcvcr. they had an from still framc through 64 times play unforeseen advantage: the audio and video 1 -in hclical VTR. The uppermost curve shows wide-band speed, with the speed proportional to the signal performance challcnged that of ex- amount the dial is moved off ccntcr. isting 2-in quadruplex broadcast record- noisc measurements out to the twelfth generation. These tests wcrc made using Countcrclockwise motion performs the ers. same function in reverse. CBS chosc a nonsegmented or onc field three machines with a procedure which per scan video format from among those ensured that no recording was played back 1-in helicals offercd at that time. This on the machine on which it was made. format, later to be standardized as SMPTE (Time-base correctors were used for each Table I. BVH-1000 video performance. (Mea- Type C, has the inherent capability of of the 12 dubs.) Each generation, therefore, surements include BVT-1000 time-base correc- producing viewable pictures at many tape included an interchange loss and a dubbing tor.) speeds. Sincc that time a digital store has loss. Thc middle and lowcr curves show am- Frequency f0.3 d B from 30 Hz to 4.1 5 been designed to provide this capability responsc MHz; -3 d B at 4.5 MHz also with thc scgmented formats. plitude- and phase-modulation losses as measured on a Shibasoku (of Tokyo) 925C Signal-to-noise 49.5 dB unweighted CBS currently uscs the Sony BVH-1000 ratio in editing systcm designs and, necessarily, meter. The significance of these curves can Differential 3.0% thc data which follows describes that ma- only be adequately explained in a paper on gain chine. However, there is every indication that subjcct alone. Suffice it to say that Difl-erential 2.5O both the AM and PM noise curves are as phasc Prcsentcd on 3 February 1978 a t the 12th Annual smooth as the wide-band curve and that K-rating (2T 0.8% SMPTE Television Conference in Atlanta by William they also track that curve rather well. Our sin* pulsc) G . C onnolly. CBS Television Network. 51 W. 52 St., experiencc has shown that wide-band noise Moire (75% -43.0 d B New York, NY 10019. Copyright (0 1978 by theso- color bars) cicty of Motion Picturc and Television Engineers, Inc. measurements do not correlate well with the subjcctive evaluation of videotape sig- Interchange 0.2 dB Previously published in unedited form in the SMPTE loss (SNR) book One-Inch ffcliiwl Videotape Recording. nal degradation. On the other hand, am- Volume 87 November 1978 SMPTE Journul 161 - 5 0 dB Table 11. BVH-1000 audio performance. Frequency *0.6 dB from 50 Hz to 15 response kHz Signal-to-noisc 60 dB unweighted (ref.: ratio 3% THD at 400 Hz) 0 Harmonic 0.716 (ref.: 8 dB below I- distortion 3% THD at 1 kHz) 4 M Wow and 0.03% rms, NAB Y -40dB flutter unweighted Y Audio I / -38 dB at 50 Hz 0 Audio 2 -50dBat 100Hz z crosstalk -65 dB at 1 kHz 0 -57 dB at 15 c; kHz d< MODULATED NOISE Z 2 facilities, one a t CBS Television City in In Hollywood and the other at CBS Studio 51 PHASE MODULATED NOISE in New York. The Television City facility -30 dB is described in a separate paper by William C. Nicholls entitled “A New Edit Room Using One-Inch Helical VTRs” - see p. 764 in this issue of the Journal. 0 d B The Studio 5 1 installation is unique in 1 2 3 4 5 6 7 8 9 1 0 1 1 ’ one important sense. The system is dedi- GENERATION cated to a studio 30 blocks distant from Fig. I . BVH-1000m ultiple-generation performance. CBS Broadcast Center. Prior to the advent of lower-cost machines, the recorders would have been integrated into the Proportional control is actuated by the modc, vidco lock is achieved typically in 2 Broadcast Center complex for maximum samc dial through operation of a mode s. Three seconds is a safe preroll time. utilization, with local telephone-company button. Continuous clockwise rotation of Video lock-up time from the stop mode loops providing service to the off-premise the dial moves the tape forward in speeds requires about 13 s - most of which is studio. Now, however, the lower capital from still frame to one-fifth normal play consumed in bringing the scanner up to cost of the I-in VTRs has allowed four speed. The faster the dial is turned, the speed. Also, due in part to the use of ac machines to be dedicated to that studio faster the speed. Counterclockwise rota- servos throughout, the transition between alone. tion, of coursc, performs the samc function modes requiring high motor torque seems The third facility is different in many in rcvcrse. The best way to describe the sluggish. This is no doubt due, in large part, ways, principally because it is located a t a action of proportional control is that it is to conccrn on the part of the manufacturer film production facility - CBS Studio akin to grabbing the reels by hand and ro- to avoid tape damage.