Skip to main content

Full text of ""Tune In, Turn On, Drop Out: Section 108(c) and Evaluating Deterioration in Commercially Produced VHS Collections""

See other formats


13 
Tune In, Turn On, Drop Out 


Section 108(c) and Evaluating Deterioration 
in Commercially Produced VHS Collections 


Walter Forsberg and Erik Piil 


ABSTRACT 


Analog videotape, an imperfect moving image technology format since its introduc- 
tion, is reaching the end of its life cycle. However, large quantities of out-of-print 
and irreplaceable VHS titles still comprise significant portions of library and archival 
collections and circulations. Given the need to preserve this content, this study in- 
vestigates the use of the “dropout” metric (counts of disruptions in the video signal) 
for determining whether libraries and archives can invoke their rights of reproduc- 
tion under the United States Copyright Act. Videotape technology and deterioration 
problems are explained and prior deterioration studies are reviewed. Dropout tests of 
four pairs of commercially produced VHS titles are conducted and relationships be- 
tween videotape deterioration as measured by dropout counts, circulation statistics, 
and manufacturing quality control standards are evaluated. Offering noninvasive 
evidence of videotape deterioration, quantified dropout counts appear to provide 
libraries and archives with an objective measure to meet the vague “deterioration” 


standard of the Copyright Act. 


INTRODUCTION 


Particularly among research-level university institutions, circulating library collec- 
tions still consist of large amounts of VHS videotape. With the advent of “peak- 
VHS” sales in 1998, and the 2008 announcement of commercial discontinuation 
by its “final supplier” (wholesaler Distribution Video and Audio), university librar- 
ies defaulted to become some of the most significant North American entities still 


213 


214 Walter Forsberg and Erik Piil 


annually dealing with substantive quantities of the videotape format (Simon & 
Kugler, 2008). Faculty preferences, budgetary limitations, and the unavailability on 
subsequent formats of content originally held and acquired on VHS, all contribute 
to these institutions’ continued reliance on this practically obsolete format. As one 
example, the Avery Fisher Center for Music and Media at New York University 
(NYU) Libraries recorded 4,371 checkouts of VHS tapes from 2011 to 2012, repre- 
senting nearly 15% of its total circulation for video materials. Four years prior, VHS 
circulation stood at half of that of titles on DVD. 

During this decade-long period of decline for VHS as a moving image user format 
in consumer spheres, the landscape of media use in libraries radically changed. Infor- 
mation management and delivery mechanisms met a period of rapid technological 
development and deployment. Impacted directly by the Internet, library patron expec- 
tations expanded across diversifying technological strata where speed of delivery, facil- 
ity of discovery, and digital access became the new normal. The trending obsolescence 
of VHS “fast-forwarded” alongside these shifts, as content became available on DVD 
and BluRay, and as digital files through internal and external streaming services. Citing 
the growing unavailability of VCRs, a wave of quiet notifications by institutional in- 
formation technology (IT) departments now inform media librarians and faculty that 
support of VHS playback will soon draw to a close. As one recent contributor to the 
VIDEOLIB listserv characterized it, "VHS death went from a lingering, gradual one 
(to which we seem to be slowly adapting) to a quick bullet to the head.”' 

This "VHS death" in libraries is troubling for a host of reasons. Immense resources 
have been expended on growing libraries’ research and teaching collections in a for- 
mat they are being told will soon be unusable. Furthermore, a significant portion of 
content held on VHS is out-of-print or unavailable on other formats (though, few 
libraries have undertaken the required research to systematically determine what the 
scope of this unavailability is). Many librarians are ill-equipped with limited insti- 
tutional support, financial budgets, specialized hardware, or technical knowledge 
needed to perform preservation of their videotape collections in-house, or through 
a third-party vendor. For these reasons, and others, circulating videotape collections 
in libraries are critically threatened. 

What is frustrating, in the face of these circumstances, is the fact that fair use 
rights afforded under Section 107 of the United States Copyright Act actually ex- 
ist to make copies for the sake of preservation. Congress, even, explicitly mentions 
that preservation of deteriorating moving images on aging formats is an activity that 
“certainly” qualifies as a fair use.^ Furthermore, Section 108 of the United States 
Copyright Act explicitly provides an additional special exemption for libraries and 
archives to make copies of material held on obsolete formats, and for material that is 
“damaged, or deteriorated.” Yet, many libraries have remained timid and inert about 
preservation reformatting for commercially produced circulating VHS collections. 

This chapter is an attempt to buoy the confidence of libraries in their present 
(or, yet unrealized) efforts to digitally preserve their at-risk analog circulating VHS 





Tune In, Turn On, Drop Out 215 


collections. With specific attention to deciphering the vague and format-agnostic 
legal qualification of “deteriorating” required under Section 108(c), this paper 
presents a survey of magnetic media deterioration studies. Considering dropout as 
a key practical metric for deterioration determination, we propose and employ a 
methodology, and present results from technical tests undertaken at DuArt Film 
and Video with VHS videotapes from New York University Libraries collections 
at the Avery Fisher Center for Media (AFC), conducted as part of the Andrew W. 
Mellon Foundation-funded project, Video At Risk: Strategies for Preserving Com- 
mercial Video Collections in Libraries. 


SECTION 108(c) AND MAKING COPIES 


Aside from fair use rights of reproduction for preservation purposes set forth in Sec- 
tion 107 of the Copyright Act, under specific circumstances libraries and archives 
possess additional legal rights for making copies of material for which they do not 
own copyrights. Section 108(c) of the United States Copyright Act provides an ex- 
emption for libraries and archives to make copies of copyrighted material if a copy 
of a work belonging to a library or archive is “damaged, deteriorating, lost or stolen,” 
provided no replacement copy can be found in the marketplace, and resultant (no 
more than three) new copies are not circulated to the public outside the premises. 
These copies can be digital copies. As the statute states: 


(c) the right of reproduction under this section [i.e., § 108] applies to three copies or 
phonorecords of a published work duplicated solely for the purpose of replacement of 
a copy or phonorecord that is damaged, deteriorating, lost or stolen, or if the existing 
format in which the work is stored has become obsolete, if— 

(1) the library or archives has, after a reasonable effort, determined that an unused 
replacement cannot be obtained at a fair price; and 

(2) any such copy or phonorecord that is reproduced in digital format is not made 
available to the public in that format outside the premises of the library or archives in 
lawful possession of such copy. 

For purposes of this subsection, a format shall be considered obsolete if the machine 
or device necessary to render perceptible a work stored in that format is no longer 
manufactured or is no longer reasonably available in the commercial marketplace. (17 


U.S.C. § 108(c)) 


The existence of manufacturers still producing combination VHS/DVD playback 
machinery likely disqualifies VHS from being safely considered an “obsolete” format, 
according to a conservative interpretation of legal language. Thus, while the terms 
“lost” and "stolen" conditions are self-evident (if paradoxical, in practical terms of the 
ability to make a copy of an absent item), and the term “damaged” somewhat less so, 
the statute's failure to define the term “deteriorating” prompts the current inquiry. 


216 Walter Forsberg and Erik Piil 


VIDEOTAPE TECHNOLOGY 


Articulating the aspects of deterioration in analog VHS tapes first requires a brief 
recap of what videotape technology is, and how it basically works. Analog videotape 
stores magnetically encoded electrical information to represent recorded sound and 
moving images. Through a complex technological language of electromagnetic en- 
gineering, a magnetizable layer of (most often) ferric oxide in polyurethane binder 
is evenly applied to a thin film polyester base and serves as a carrier for every man- 
ner of content in broadcast, educational, and home theater markets. Other binder 
materials have been employed; however, those described above have been the most 
common and widespread methods with regards to the VHS format considered in the 
experimental study, below. 

For a half-century, across dozens of video format platforms, magnetic tape 
manufacturers sought to improve videotape stability and performance by experi- 
menting with various “recipes” for this magnetic binder. From several metal par- 
ticle formulations of ferric oxide, cobalt ferrite, and chromium dioxide, to other 
metal-evaporated tape formulations, these variations in binder makeup largely 
remain undisclosed trade secrets. While manufacturers made advances in binder 
density capacity, and introduced many subsequent analog and digital tape formats, 
the imperfections inherent in magnetic videotape recording were never completely 
eliminated; attempts at alleviating them continue to this day in the field of mag- 
netic storage for data. As an organic carrier, videotape naturally experiences some 
degree of physical change of state with use and over time. The stability of infor- 
mation stored using this technological memory system—a kind of semi-hard “rust 
soup" on plastic strips—is subject to a wide array of liabilities and potential points 
of failure that can contribute to deterioration. 


WHAT CONSTITUTES DETERIORATION 
IN MAGNETIC VIDEOTAPE? 


Relevant Previous Deterioration Testing 


A modest body of scientific scholarship exists relating to experimentation employ- 
ing chemical analyses of videotape binders, documenting their instability and sus- 
ceptibility to breakdown caused by moisture absorption, known as “hydrolysis.” One 
1993 publication saw British scientists and the Agfa Gevaert company collaborating 
on measuring naturally and accelerated-aged tapes using Fourier Transform infrared 
(FTIR) spectroscopy to determine the resilience of polymer chains in the magnetic 
binder, with specific emphasis on the impact of hydrolysis. Their work scientifically 
reiterated the central role binder hydrolysis plays in videotape deterioration (in lieu 
of organic breakdown in the polyester film base), yet did not specifically consider 
the factor of repeated playback (Edge, Allen, Hayes, Jewitt, Brems, & Horrie, 1993, 
pp. 207-214). A recent Institute for Museum and Library Services grant awarded 


Tune In, Turn On, Drop Out 217 


to researchers at the University of South Carolina in collaboration with the Library 
of Congress supports the development of a rapid, nondestructive, degradation- 
identification tool using infrafed (IR) spectroscopy and an IR spectral database for 
determining deterioration. Other studies often also document hydrolysis break- 
down of tape binder via optical magnification, using scanning electron microscopes 
(SEMs) (Gilmour, 2000). Such SEM-based approaches can provide visual evidence 
of changes in the surface of the videotape binder, including impurities on the tape 
surface caused by binder breakdown and deterioration. 

Other studies consider the role of repeated playback in precipitating physical 
videotape binder deterioration. A 1992 peer-reviewed published test by several engi- 
neers at Sony considered the effects of repeated video head wear on an immobilized 
videotape track in “still mode.” Their test methodology revealed physical portions of 
the tape binder to be removed by rotary heads of a three-head VTR spinning over the 
same surface area at ninety cycles per second, citing several physical stress responses 
of the inherently uneven surface of magnetic tape (Osaki, Oyanagi, Aonuma, Kanou, 
8c Kurihara, 1992, pp. 76-83). The repeated physical contact between head and tape 
in this test parallels some use concepts for the current study we describe; however, 
the authors’ detail in describing their metric for documenting performance loss in 
the videotape is vague. 

In a 1999 paper, a United Kingdom-based police unit considered the effect 
multiple recording passes had on diminishing ability to remagnetize tape binder 
with different recorded content. Using visual test pattern charts, the Police Scien- 
tific Development Branch sought to determine the number of times closed-circuit 
television (CCTV) systems might reuse a single videotape to record new (and, not 
merely repeatedly playback the same) security camera footage, determining that 
after an average maximum of twelve recording passes deterioration of the videotape 
binder resulted in unacceptable visual quality. Again, it is worth noting that their 
methodology was dependent on subjective human visual assessments of the test 
charts they employed (Mather & Neil, 1998, pp. 220—224). While such rerecording 
and magnetic remanence investigations are less pertinent to investigating playback 
performance of material recorded only once, they serve to reiterate the complexity of 
the chemical and organic processes with regard to videotape performance. 

One playback-based study by the British Broadcasting Corporation in 2000 
found that over 45% of the 2,800 34-inch U-matic videotapes digitized required 
that “technical comments” be made during transfer (Lee, Prytherich, & King, 2000, 
pp. 177-186). This study, however, did not speculate as to whether problematic 
playback issues (such as incorrect audio levels, “low RE" and “noisy/low-quality 
pictures") were by-products of the original recording process, or the age of the tapes 
at the time of transfer (all of which were between eleven and eighteen years old). 

As Jean-Louis Bigourdan, James M. Reilly, Karen Santoro, and Gene Salesin of 
the Image Permanence Institute observed in a 2006 National Endowment for the 
Humanities final report, articulating and documenting videotape degradation with 
any degree of specificity is a challenging and under-sophisticated reality (Bigourdan, 


218 Walter Forsberg and Erik Piil 


Reilly, Santoro, & Salesin, 2006, p. 15). Across the extant scholarship, at some point 
along the spectrum of testing, analysis, and comparison, a lack of condition assess. 
ment (most often, at the moment of creation or acquisition) renders any subsequent 
condition without a comparative counterpoint. 

One of the most in-depth studies on videotape deterioration to date is the 2005 
PrestoSpace report, Report on Video and Audio Lape Deterioration Mechanisms and 
Considerations about Implementation of a Collection Condition Assessment Method. Its 
authors define “deterioration” as resulting from “an alteration process” in the mag- 
netic tape (Thiebaut, Vilmont, & Lavedrine, 2006, p. 15). Such an alteration process 
can be identified by four symptoms that lead to loss of performance during playback, 
including: tape-transport instability, a decrease in signal strength, loss of signal, and/ 
or dropout. The PrestoSpace report echoes comments by Bigourdan, et al. on the 
complexity of deterioration evaluation processes, alluded to above, concluding that 
deterioration of videotape: 


depends not only on the intrinsic material stability but also on the player specifications 
and tolerance to media deterioration as well as tape handling. In addition, the com- 
plexity of mass-manufactured tapes with numerous formulations and manufacturing 
practices result in a variety of deterioration mechanisms. As a consequence, finding a 
unique deterioration marker is highly challenging and would probably involve much 
more consequent research efforts. (Thiebaut, Vilmont, & Lavedrine, 2006, p. 42) 


The complexity of the magnetic recording processes with regards to deterioration 
is also found in, arguably, the most authoritative publication on magnetic media tri- 
bology—the study of interacting surfaces in relative motion: Bharat Bhushan's 1990, 
Tribology and Mechanics of Magnetic Storage Devices. In it, Bhushan categorizes six 
distinct kinds of wear mechanisms resulting from playback: adhesive wear; abrasive 
wear; fatigue; impact by erosion, or percussion; corrosive wear; and electrical-arc- 
induced wear (Bhushan, 1990, p. 412). Bhushan's analysis of wear is extensive, and 
he categorizes tape wear as having functional problems rooted in: high friction, loss 
of reproduced signal amplitude, and excessive dropouts caused by debris adhered to 
the tape surface (p. 462). 


Establishing a Practical Deterioration Metric 


The experiment and results presented in this paper engage the complex issues sur- 
rounding videotape deterioration as a starting point to establish a practical method 
and potential justification for libraries and archives to reformat their deteriorating 
VHS collections. As Legal Counsel Robert Clarida commented in a white paper for 
the Video At Risk project, 


Decause the process of deterioration is very context-dependent, and viewing tapes for 
visible deterioration is extremely time-consuming, it could be very helpful for the Li- 
brary to prepare and publish a formal study of the degree to which various factors about 


Tune In, Turn On, Drop Out 219 


a VHS tape correlate to the degree of deterioration, even of unplayed VHS tapes. Fac- 
tors such as age, tape stock used, storage conditions (temperature, humidity), number 
of plays, and type of playback equipment might all contribute to the speed at which 
tapes deteriorate, and if a sufficiently large sample of tapes could be analyzed as to these 
variables the Library might be able to develop a set of reliable criteria for knowing, in 
advance, how seriously compromised a given tape’s condition would be at a given point 
in the future. This could help the Library plan its digitization efforts more effectively. 


Drawing from Claridas recommendation, PrestoSpace's testing methodology, the 
facility of machine-based dropout quantification, and its status as both a videotape 
manufacturing quality-assurance metric and a visually evident on-screen phenomena, 
the authors chose dropout counts as a point of inquiry and deterioration metric for 
playback testing of several used, circulating, commercially produced VHS videotapes, 
as well as in unused, shrink-wrapped, new duplicate VHS copies of the same titles.? 
Importantly, the imperative that deterioration testing be noninvasive and nondestruc- 
tive—a requirement not normally held to by other scientific inquiries—was another 
reason for selecting dropout as a testing metric. Nondestructive chemical analyses of 
circulating VHS tapes in libraries may prove viable and feasible in the future, but were 
not so at the time of this testing. Unlike analyses of chemical changes in tape binders 
(for which no information related to their original chemical state is available), dropout 
affords a means of physical condition assessment of tapes based on the assumption that 
they were originally in a state of acceptable quality and performance. Other metrics, 
specifically signal-to-noise ratios, may also prove useful but were not employed here. 


Measuring Dropout 


Videotape manufacturer literature and specification documentation indicates that 
dropout count measurements were one of several key quality-assurance metrics by 
which manufacturers assessed blank VHS videotape. Dropouts result from disrup- 
tions in the video signal, caused by an interruption of contact between the videotape 
recorder (VTR) playback head and the videotape (such as debris clog), or by missing 
portions of the tape binder that should hold signal information but fail to because of 
a manufacturing or recording defect. Physical damage such as creasing or crinkling 
can cause dropout, as well. 

As interruptions of the video signal, dropouts are measured by the degree to which 
they cause the video signal to drop below its nominal decibels (dB) value (Braith- 
waite, 1989, pp. 3-14). Most tape manufacturers considered a decrease of 20 dB as 
constituting a dropout, and categorized dropouts by the amount of time required 
for the signal to return to its nominal value, as measured in microseconds (us). Here, 
Video Magazines editor-in-chief Lancelot Braithwaite explains his methodology's 
measurement rate of 15 us: 


In our NTSC video system it takes 63.5 microseconds to make each line of a picture 
of which 52.5 microseconds actually contain picture information. So a dropout of 15 


220 Walter Forsberg and Erik Piil 


microseconds causes a bit more than a quarter of a line to lose information. If the results 
of each dropout were visible, we would have a very patchy picture, but all VCRs have 
circuits called dropout compensators. They reduce the effects of dropouts, but they dont 
eliminate them. (Braithwaite, 1989, pp. 3-14) 


Dropout compensator circuitry can replace missing information with stored 
information about nearby preceding lines of video, which is temporarily stored by 
the compensator circuitry, often built into a VCR or VIR. While dropout com- 
pensators offer replacement chrominance (color) information—but not luminance 
(achromatic) information—the effect on the viewing experience is less disrupting 
than a “raw” dropout that appears as a brief horizontal white flash in one of the lines 
of the onscreen video image. In some cases, even with the presence of compensator 
circuitry, significant enough amounts of dropout experienced during playback can 
exceed the compensator's error concealment ability and visually persist as horizontal 
white flashes. As Dave Rice and Stefan Elnabli point out in their discussion of similar 
compensation techniques in digital videotape, assessment of error concealment is a 
“meaningful aspect of the preservation process, as it reveals the extent to which dif- 
ficulties in reading magnetically encoded signal information can compromise picture 
integrity (Rice & Elnabli, 2010, p. 5). Again, dropout compensators do not elimi- 
nate the existence of dropouts, they merely mask them from the viewer. A VIRs 
dropout compensator circuitry can be bypassed when the radio frequency (RF) signal 
of videotape playback is measured through an RF output. This is the method by 
which accurate dropout counts can be electronically quantified using a specialized 
piece of dropout counter machinery. 


DROPOUT TESTING METHODOLOGY 


Investigation Rationale 


By employing professional-grade tape signal processing and evaluation equip- 
ment, this investigation sought to discover whether a moving image work held on 
the VHS format might be demonstrated to be "deteriorating" based on quantitative 
measurement. In the case of still-playable VHS tapes in circulating media collec- 
tion, what was the correlation between the number of circulations counts and such 
damage or deterioration, if any? Did tapes with higher amounts of playback register 
higher dropout counts? Did older tapes perform worse than newer ones? Could such 
characteristics of age and playback counts be established as contributing to greater 
deterioration? And: if there were such clear thresholds after which point the tape 
could be characterized as deteriorating, what were those outer limits? 


Technical Approach 


Four pairs of commercially produced VHS titles were employed as test tapes 
in the course of these experiments: Disney's Bambi,‘ the educational title Child- 





Tune In, Turn On, Drop Out 221 


hood: Great Expectations,’ the live-action feature film Kids,° and Disney's animated 
feature Snow White and the Seven Dwarfs.’ To measure dropout counts on each of 
the eight tapes, a calibrated ShibaSoku VH01BZ Dropout Counter at DuArt Film 
and Videos Restoration Department was employed to measure the RF output 
of the VIR. All tape playback occurred on a professional-grade JVC BR-S610U 
S-VHS VTR. Once received by DuArt, the tapes under analysis were acclimatized 
to the facility’s controlled environmental conditions for a period of thirty days, at 
an average temperature of 69°F. Tapes of matching titles were played sequentially. 
During playback, each tape was simultaneously monitored for dropout counts per 
minute using the ShibaSoku Dropout Counter. Two dropout measurements (—20 
dB at 5 us; and —20 dB at 15 ps) were taken at minute intervals for all tapes under 
testing. All tape playback was also digitized to a Quicktime-wrapped 10-bit un- 
compressed v210 file for future visual reference. No cleaning of the tapes occurred 
prior to playback (to ensure condition authenticity and to mimic the circulating 
library use environment), but video and audio heads on the VTR were cleaned 
before each playback pass with trichlorotrifluoroethane (CAS 76-13-1), isopropyl 
alcohol (CAS 67-63-01), and nitromethane (CAS 75-52-5). 

To address the lack of an initial condition assessment, as described above by Big- 
ourdan et al., this experiment tested four used and four unused pairs of the same 
titles: one taken from the circulating VHS collection at the Avery Fisher Center for 
Media at NYU Libraries; and, the second a new, unused copy, from the same era, in 
original shrink-wrap, purchased from after-market vendors via Amazon.com. Great 
care was taken to ensure that new duplicates were of the same edition, produced 
by the same distribution company, and with identical packaging. It was impossible 
to verify if duplicates were made at identical commercial tape duplicator locations; 
however, similar machine identification and barcode markings on the physical VHS 
carriers suggest that they were, at least, made by the same duplication company. 
Combined with the expense of such testing, the difficulty in finding matching pairs 
of videotape titles (used and unused) in the marketplace significantly limited the 
experiments test set of tapes. 

Comparing used, circulating copies against new, unused copies, was thought to 
hold promise in isolating the effects of playback (documented by library cataloging 
systems as circulations) on tape deterioration. A third set of metrics for condition 
assessment comparison were measurements taken from published manufacturer 
quality-control standards and literature regarding dropout counts of new, blank 
videotape stock. 


Acknowledged Test Liabilities 
Environmental Storage 


Temperature and humidity conditions for storage of each tape may well have been 
different, and this difference may sway test result accuracy. 





222 Walter Forsberg and Erik Piil 


Differences in Packaging 
While both used and unused Kids tapes were 1996 Vidmark releases, and both 


contain similar white machine-inscribed markings on their tape spines (suggesting 
they were duplicated by the same manufacturer), these markings are also not identi- 
cal. The etched pattern on each tape’s physical carrier shell differs slightly. 


Limited Manufacturer Information 


As with the guarded details of manufacturer recipes for tape binder composi- 
tions, there is relatively limited manufacturer literature on the technical details of 
videotape. The manufacturer dropout count measurements cited in this study come 
from only two manufacturers (Sony and 3M) from 1991 to 1992, and it cannot be 
determined which manufacturer’s tape stock was employed in the creation of the 
titles under testing. 


Back-Coating 


In higher-quality tape stock, the back surface of the tape (opposite the emulsion 
side that faces the video playback and recording heads) is usually coated with carbon 
to reduce static generated by friction when travelling through the cassette past the 
metal tape guides. A preliminary inspection of each test tape for carbon-backings 
yielded interesting results: both the new, unused tapes for Childhood: Great Expecta- 
tions and Snow White and the Seven Dwarfs featured carbon-backed videotape, while 
their circulated counterparts did not. This finding revealed important realities about 
the consistency of tape stock during large duplication runs of commercially cre- 
ated titles in processing plants. As with the preceding acknowledged test liabilities, 
however, these are unavoidable realities that contravene ideal experiment control 
environments. 


DROPOUT TESTING RESULTS AND DISCUSSION 


The following charts depict dropout counts for each of the four pairs of titles at 
both measured dropout sensitivities (-20 dB at 5 us; and —20 dB at 15 us). Readers 
will note that each title’s pair of graphs represent measurements performed simul- 
taneously, at different sensitivities. Used, circulated tapes appear as light gray bars. 
Unused, new tapes appear as black bars. Minute-by-minute dropout counts can be 
found in the appendices. 








mn — ma 





223 


Tune In, Turn On, Drop Out 








AUT. aint au a gd. tah AA Sete o vvv swell hu vehi fyb enn de deve Bebb ne vanes 
were g CE ES epe NAG of TAN EAM Geet 
rad 3 = ; OR 0 


i à AMA 
ISS 
HUMANAS ped 
m nte us 
> ' " Q 


pod 


Y vs : T. ue 
gus Soars 
ee ae 





i 






Circulated 

















A USUS NS 


— ANN 
initiiert 
LL. Em 


New 


E 


"asang - ^ te ^ 
a A s aaa a na ‘oes a ve NAN) 
tn t 
4 sass) 






BAMBI - Normal Dropout (-20 dB at 15 Lis) per Minute 


Pi "nsa 
x (ANSA, 
TU UPON PERRA. Saree, RES ERPRNS " RERO AA 3990929959 ILEIN vete HARANGAN mog pox 
eene clem metn entm emang m n s rendered vt 





9 8 7 8 99 8 8 8n a yn g n” o 





Figure 13.1. 


R 8 S € 8 NAN 8 3 8 vn o 





New 


13 57 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 


BAMBI - Small Dropout (-20 dB at 5 us) per Minute 


Figure 13.2. 


Walter Forsberg and Erik Piil 


224 


ons 


Great Expectati 


II. Childhood. 


So OPER ISALI eee 


m POOE Se E EE swe d SAR a ae 2 : vost 





“a &*5* ECR a 53. A AA ^ ^ 





123456789 10111213 14 15 16 17 18 19 20 24 22 23 24 25 2627 28 29 3031 32 39 34 95 96 37 3839 4041 42 43 4445 46 47 4849 50515253 54555657 
CHILDHOOD - Normal Dropout (-20 dB at 15 us} per Minute 


Figure 13.3. 


il 


Circulated 


New 





Circulated 


New 


CHILDHOOD - Small Dropout (-20 dB at 5 is) per Minute 


Figure 13.4. 


229 


Tune In, Turn On, Drop Out 


Ki 


III. 


ae uaa 





€: 





ES be 
IAS 





MAMMA P 


T EUIS PEER? 


EIER PIPES 





AA CEET Cer CERTIS 


ATA! ES ens sane 
MIN EM reet 


NAA 
ITTEN S 


EEANN 


Steers ST 


MEANS 


PER SRL S A vs RS 


HC OUNTAPTASAARARRASRERISVSSARRRRSSSSSRRRAELRSSBSESARRRSE 


EDR DASS V VENISSE ERIS YY 
TERETE T ORE NN 29 MAMMA 


RRMA ERAN SGA ARV M 





Circulated 


New 


KIDS - Normal Dropout (-20 dB at 15 ps} per Minute 


Figure 13.5. 


IAEI warn ¥ 
yo dt E 


ISSUES SR SISIKAT 


son nes eases Wa 
saat nut OOPERIS, 
ma 5 


SRS eens BISAN OD SOLARIS ANALOG SORA RADII 


paaa OERI ttt REIR PERL ISES 


8 8 RESRREERRGSRERNARAR”O 








LPEPPEPEEPEEEEEEEEEEEEEEETETETETTTETEREPEFETEPEEREY: 


KIDS - Small Dropout (-20 dB at 5 us) per Minute 


Figure 13.6. 


i 


Circulated 


New 





Walter Forsberg and Erik Piil 


226 


IV. Snow White and the Seven Dwarfs 


g 9$ 8 8 8 FK R 8 9 S 8 9 9 8 RRRA” 





Circulated 


New 


SNOW WHITE - Normal Dropout (-20 dB at 15 us) per Minute 


Figure 13.7. 


EA ien 
RE 303532922000 1033009200000 93999 000:29€ 133200200229: 390022002999 109002009015 GON 


RECS ER bun eii See nine a 
1 " 
T PIERDE SEES SESE EN ORS E secs 
LE LLL OBIE SILLS SE 
FEI 200993999944 I 100099199070 220999990005 1000395720006 
eoo 
x opos. 





eis 
Ae 





SIE MENSURA SPORE SEI IIRL PAA: SETS 


TE IRIE OSES. LEGS KASING 
PERE BBRA KARAPATAN. EIERE IIT 


ASS 
A 
KSS 
x 
ics 


IRE IIIS BAKBAKAN KARRI DIINO BELEE SIRES NILES SRI SRS 
re KANG KAPAG. PAKPAK ISNAN, GARAPATA, PANTALAN 7721 
RSE: RETINA VAILE BABABA PARA: ANE 
m 23 2 POPS PANA EIEE o 


seeeuaeerreeseanegeseaenanaaee 


1 : 5 2 BS tS E mY Ne i 
ERE BANA PLELEELELE TANAUAN BANG 
1 3 5 7 9 1113 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 


SNOW WHITE - Small Dropout (-20 dB at 5 us) per Minute 


Figure 13.8. 


Circulated if 


New 





Tune In, Turn On, Drop Out 227 


Comparing Test Results with Blank Tape Stock Manufacturer Literature 


Taken from original manufacturer tape stock specification pamphlets, below are 
some examples of dropout count acceptability standards for new tape stock: 


Video Performance for Sony’s Professional Grade VHS Video Cassettes,” dropouts 
(—20dB, 15 us): 10 per minute. 

Video Performance for Sony MQST-120 S-VHS,” dropouts (-15dB, 20 ps): less 
than 5 per minute (average). 

Video Properties for 3M Broadcast VHS videocassettes,!° dropouts (-20 dB, 15 
us): maximum average: 6 per minute. Typical average: 3 per minute. 

Video Properties for 3M Master Broadcast S-VHS Videocassettes,'' dropouts (-12 
dB, 30 us): 8 per minute. 


By the early 1990s, the era from which these standards are available, it appears 
that fewer than ten dropouts per minute was an acceptable threshold of quality 
that dropout counter sensitivity metrics (in dBs and us) may have been adjusted 
to achieve. Acceptability thresholds for videotape also appear to have changed over 
time. According to Video Magazine, "in 1982, fewer than 15 blemishes per minute 
was considered excellent. [In 1990] a tape must have fewer than five blemishes per 
minute to be rated excellent" (Woodcock, 1989, p. 16). 

Likely due to a combination of deterioration contributing factors—age, possible 
hydrolysis, suboptimal environmental conditions, and (for some) amount of play- 
back wear—all tapes tested in our experiment failed manufacturer quality standards 
for acceptable dropout counts. If we are to assume that these tape copies were origi- 
nally created during the duplication/recording process on manufacturer tape stock 
that met quality standards quoted above, the tapes under testing can be understood 
to have experienced some manner of “alteration process" and should be seen to 
qualify as deteriorating. 


Comparing Test Results with Third-Party Blank Tape Stock Tests 


As a service to its readership, an exhaustive 1989 quality survey test of tape 
manufacturers blank stock by Video Magazine (1989, p. 12) established the follow- 
ing ratings hierarchy for dropout count testing of blank videotape stock (measured 


at -204B, 15 ps): 


0—5 dropouts per min. - Excellent 
6—10 dropouts per min. - Very Good 
11-20 dropouts per min. == Good 
21-30 dropouts per min. = Average 
31-40 dropouts per min. = Fair 

41-50 dropouts per min. = Weak 


over 50 dropouts per min. = Poor 


228 Walter Forsberg and Erik Piil 


Of the sixty-five stocks from twenty-one manufacturers Video Magazine tested, 
most manufacturers’ blank stocks fell in the upper half of this hierarchy, supporting 
the validity of manufacturer-stated measurement literature. Only two stocks satisfied 
the “Poor” rating for dropout counts, none satisfied the next category of “Weak,” and 
only one qualified as “Average” (Video Magazine, 1989, pp. 18-23). 

Using this hierarchy for comparison, the tapes under testing in this experiment 
averaged the following dropouts per minute (measured at —20dB, 15 ps): 


Bambi (used, circulating) 27.0 dropouts per min. [Average] 
Bambi (new, unused) 13.3 dropouts per min. [Good] 
Childhood: Great Expectations (used, 

circulating) 38.3 dropouts per min. [Fair] 
Childhood: Great Expectations (new, unused) 13.7 dropouts per min. [Good] 
Kids (used, circulating) 17.2 dropouts per min. [Good 
Kids (new, unused) 41.1 dropouts per min. [Weak] 
Snow White and the Seven Dwarfs (used, 

circulating) 8.3 dropouts per min. [Good] 
Snow White and the Seven Dwarfs (new, 

unused) 13.4 dropouts per min. [Good] 


Dropout measurements for all of the tapes under testing in this experiment, with 
the exception of one, fail to surpass the “Good” categorization and two of them rate 
in the lower rungs of this hierarchy. 


Comparing Test Results with Third-Party Tests 
of Commercially Produced Tapes 


While instructive, charting dropout counts in commercially produced tapes in 
this experiment against dropout measurements of blank, new videotape stock is 
somewhat of an imperfect comparison. Ideally, commercially produced tapes would 
be measured against commercially produced tapes, and a 1990 investigation by the 
trade publication Video Review makes such a comparison possible. 

Titled, "Videos Dirty Secret,” this study by author Ron Goldberg (1990) uses 
quantitative and qualitative testing to reveal defects in over 60% of the commercially 
produced tapes tested, finding that even tapes of titles duplicated for larger Holly- 
wood studio releases evidenced excessive dropout counts—the principal evaluation 
metric employed by Goldberg (pp. 36-39). 

None of the Video At Risk (VAR) test tape titles overlap with those tested by 
Video Review, but Goldberg' findings are especially relevant, as the VAR test stra- 
tegically selected two Disney animated feature titles that carried the THX label 
of quality—a quality certification system bestowed on only the highest-quality 
duplicated material. Among the new, unused copies of those titles, Bambi failed to 


Tune In, Turn On, Drop Out 229 


perform significantly better than its used, circulated counterpart, and Snow White 
and the Seven Dwarfs actually performed worse than its used, circulated counterpart. 

Goldberg’s conclusion is akin to an indictment of videotape duplicators looking 
to save money by using inferior-grade blank tape stock in the duplication process. 
The notion that the tape duplication industry could utilize substandard tape stocks is 
not surprising. As one former tape duplication industry employee described it, “tape 
duplication was a penny-pinching business.”'* Goldberg's study found that of the 
thirty-six tapes tested (pairs of eighteen titles) more than one-third of them exceeded 
dropout rates of fifty per minute (measured at —20dB, 15 ys). ^ 

The suggested inconsistency among duplicators quality-assurance processes that 
Goldberg uncovered may be echoed by our experiment. The two THX-certified 
Disney titles had discrepancies in each tape’s physical characteristics: only one of the 
Disney titles (new, unused Snow White and the Seven Dwarfs) contained anti-static 
back-coating on the tape—a quality measure intended to extend the life of the tape 
and protect against damage. Moreover, if inferior tape stock was, in fact, so commonly 
employed by tape duplicators, the authors believe that the case for deterioration quali- 
fication is strengthened: tape stock of inferior manufacturing quality is likely to result 
in decreased stability and performance, over time, when compared with tape stock that 
meets the high manufacturer quality-control standards cited above. — 


Impact of Playback and Circulation Counts on Deterioration 


The circulation statistics for tapes tested in this experiment appear below: 


Bambi (used, circulating) 45 circulations 
Bambi (new, unused) 0 circulations 
Childhood: Great Expectations (used, circulating) 248 circulations 
Childhood: Great Expectations (new, unused) 0 circulations 
Kids (used, circulating) 203 circulations 
Kids (new, unused) 0 circulations 
Snow White and the Seven Dwarfs (used, circulating) 17 circulations 
Snow White and the Seven Dwarfs (new, unused) 0 circulations 


Looking at the two highest-circulated titles, Childhood: Great Expectations and 
Kids, a pattern may be said to corroborate claims that tapes experiencing high levels 
of playback experience a burnishing or “calendaring” effect whereby the videotape 
binder becomes so well-worn that the rate of subsequent physical deterioration 
diminishes, or plateaus. The specifics of calendaring resulting from VTR head-to- 
tape physical contact are reflected in the research of Osaki et al., and Bhushan, cited 
above. But, for the purposes of this inquiry it can be stated that highly circulated 
tapes are likely to have already experienced such a degree of physical wear and 
deterioration that by their 200th-odd pass the microscopic roughness of the tape 


230 Walter Forsberg and Erik Piil 


binder has been significantly worn down—in essence, polished. This fact may be 
reflected by the dropout counts in Childhood: Great Expectations, yet inconsistency 
in tape backing between the used and the unused copies makes such a determination 
difficult to definitively assert.'* What this calendaring effect does directly suggest is 
that libraries wishing to detect deterioration in their circulating VHS tapes begin 
such evaluations with the most highly circulated titles in their collections. 


Impact of Age on Deterioration 


For tapes under testing in this experiment no noticeable trend can be asserted 
with regard to the impact of age alone on quantifiable deterioration factors, such as 
dropout counts. However, insofar as both used and unused copies of all titles failed 
manufacturer quality-control standards, age likely plays a core contributing role in 
the deterioration-associated “alteration processes” taking place in the tape binder. 
While this may not be so surprising a finding, it does directly suggest that libraries 
wishing to detect deterioration in their circulating VHS tapes begin such evaluations 
with the oldest titles in their collections. 


Impact of Environmental Conditions 


The possibility that NYU-Libraries’ environmental conditions played a corrupt- 
ing role and contributed to potential tape hydrolysis remains. It is clear that such 
environmental conditions may fall short of ISO 18934 recommendations for tem- 
perature, relative humidity, and stability (which should not exceed 73°F for 20% RH 
and should not exceed 52°F for 50% RH) (ISO, 2006). However, the conditions at 
NYU-Libraries are most likely in keeping with those conditions at other circulating 
video libraries, and it is unlikely that many other libraries store their circulating VHS 
collections in ISO 18394 compliant conditions. 


GENERAL IMPRESSIONS AND CONCLUDING OBSERVATIONS 


The dropout count testing results of these experiments demonstrate that all tapes 
considered failed quality standards set by tape stock manufacturers. These test results 
were surprising as the authors did not foresee such resulting high dropout counts for 
the tapes, and as we had originally anticipated some manner of specific correlation 
between circulation and dropout—namely, that highly circulated tapes would dem- 
onstrate a more contextual trend in dropout counts. While the small data set from 
this testing may make it difficult to definitively assert such trends or offer definitive 
conclusions about the deterioration status of other videotapes in the AFC circulating 
collections, it is clear that these tapes no longer meet any of the available quality and 
performance documentation that they were believed to have met in their original 





Tune In, Turn On, Drop Out 231 


state at the time of production and/or manufacture.” Whether or not a tape under 
testing in its original state adhered to such standards seems moot, as such a retroac- 
tive determination is impossible for these (and most other) tapes. Most importantly, 
the authors strongly believe that all evidence suggests that the complex factors in- 
volved in the “alteration processes” of deterioration will only serve to intensify the 
deteriorating condition of any given videotape with time. The evident deterioration 
relative to dropout counts is likely only the first in a litany of deterioration factors 
for all tapes tested in the course of this experiment. 

Unlike myriad other potential deterioration factors, dropout count thresholds 
specified by tape manufacturers actually do exist. As such, we believe that such 
quantified dropout counts offer one of the least subjective forms of noninvasive 
evidence for videotape deterioration currently available. From the legal perspective 
of justifying preservation reformatting under Section 108(c), we would recommend 
such dropout count measurement, as it can at least enable contrast with subsequent 
physical condition states, even if only at future points in time. While one can, per- 
haps, never know an original dropout count measurement for a specific videotape 
item at the point-of-creation, documenting current dropout rates for library collec- 
tions can be seen as invaluable data for deterioration assessments down the road. The 
authors also concur with the legal analysis of Counsel Robert Clarida that, given the 
absence of a specific requisite quantity of deterioration in the language of the law, as 
small a spike in dropout counts as those lasting even a few minutes would certainly 
qualify a tape as “deteriorating.” 


ACKNOWLEDGMENTS 


This testing was carried out under the auspices of the Andrew W. Mellon Foundation- 
funded project, Video At Risk: Strategies for Preserving Commercial Video Collections 
in Libraries. The authors would like to thank Counsel Robert Clarida; DuArt Film 
and Video's Chief Engineer, Maurice Schechter; NYU Libraries’ Melissa Brown, Carol 
Mandel, Craig Michaels, Ben Moskowitz, David Perry, Michael Stoller, Kimberly 
Tarr, and Kent Underwood; as well as VAR interns Federica Liberi and Kristin Mac- 
Donough for their research contributions, feedback, and assistance. Special thanks to 
Howard Besser for his continued support in the research of moving image preservation. 


Tune In, Turn On, Drop Out 241 
NOTES 


1. October 10, 2012. “[Videolib] Confronting a campus wide VHS DEATH Deadline.” 
2. See H. Rep. 94-1476 at p. 73, under the heading “Reproduction and uses for other 


purposes": 


A problem of particular urgency is that of preserving for posterity prints of motion pictures made 
before 1942. Aside from the deplorable fact that in a great many cases the only existing copy of a 
film has been deliberately destroyed, those that remain are in immediate danger of disintegration; 
they were printed on film stock with a nitrate base that will inevitably decompose in time. The ef- 
forts of the Library of Congress, the American Film Institute, and other organizations to rescue and 
preserve this irreplaceable contribution to our cultural life are to be applauded, and the making of 
duplicate copies for purposes of archival preservation certainly falls within the scope of “fair use.” 


3. The term “new” should be understood as meaning an unused copy of the same title, 
purchased in its original shrink-wrap packaging, and manufactured at the same time and era 
as the used, circulating copies. As VHS videotapes are no longer manufactured, references to 
such “new” tapes do not mean tapes that were recently manufactured. 

4. Bambi [Fully Restored 55th Anniversary Edition], directed by David Hand (1942; 
Burbank, CA: Walt Disney Home Video, 1997), VHS. 

5. Childhood: Great Expectations, directed by Geoff Haines-Stiles (1991; New York: Am- 
brose Video, 1991), VHS. 

6. Kids, directed by Larry Clark (1995; Marina Del Rey, CA: Vidmark Entertainment, 
1997), VHS. 

7. Snow White and the Seven Dwarfs [Platinum Edition], directed by David Hand (1937; 
Burbank, CA: Walt Disney Home Video, 2001), VHS. 

8. Sony Professional Grade VHS Video Cassettes, Sony Corporation specifications sheet, 
1991. Catalog no. ACG-5140-TKY-9109-P 1-005. 

9. Sony MQST S-VHS videocassettes, Sony Corporation specifications sheet, 1992. 
Catalog no. ACG-5164-TCS-9206-010. 

10. Broadcast VHS tapes, specifications Sheet, 3M, 1991. Catalog no. 84-9811-4338-5. 
According to the Test Notes, “during playback of a 50% gray level signal, the unlimited RF 
signal is monitored for dropouts exceeding 20 dB in depth and 15 microseconds in length.” 

11. Master Broadcast S-VHS Videocassettes, specifications sheet, 3M, 1991. Catalog no. 
84-9811-4340-1. Defined in Test Notes as, “any tape defect that produces a 12dB [for] 30 m 
sec or greater is electronically counted as a dropout.” 

12. Maurice Schecter, in conversation with the authors, December 2012. 

13. While Goldberg's investigation does not specify dropout dB size measured, past Video 
Review dropout count testing employed a -18 dB measurement (Measuring tape, 1990, 31-36). 

14, The severe spike in dropout at the beginning of the used copy of Childhood: Great Ex- 
pectations can be attributed to physical crinkling damage—affirmation of the dropout metrics 
ability to document videotape problems. 

15. Indeed, the potential data set encompassing any representative percentage of VHS 
tapes and stocks ever produced would be far too enormous and unscalable. 


242 Walter Forsberg and Erik Piil 
REFERENCES 


Bigourdan, J. L., Reilly, J. M., Santoro, K., & Salesin, G. (2006). The preservation of magnetic 
tape collections: A perspective. Final Report to National Endowment for the Humanities 
Division of Preservation and Access, NEH Grant #PA-50123-03. Rochester, NY: Image 
Permanence Institute. 

Bhushan, B. (1990). Tribology and mechanics of magnetic storage devices. New York, NY: 
Springer-Verlag. 

Braithwaite, L. (1989). Introduction. In Video Magazine’ official blank tape tests (pp. 3-14). 
New York, NY: Reese Communications Inc. 

Edge, M., Allen, N. S., Hayes, M., Jewitt, T. S., Brems, K., & Horrie, V. (1993). Degradation of 
magnetic tape: Support and binder stability. Polymer Degradation and Stability, 39, 207—214. 

Gilmour, I. (2000). Media testing in audiovisual archives: Why is my tape falling to bits? In 
M. Aubert & R. Billeaud (Eds.), /mage and sound archiving and access: Challenges of the 3rd 
millennium (pp. 79—87). Paris: CNC. 

Goldberg, R. (1990). Videos dirty secret. Video Review, 10, 36—39. 

(ISO) International Organization for Standardization. (2006). ISO 18934 Imaging Materials— 
Multiple media archives—Storage environment (1st ed.). Geneva: International Organization 
for Standardization. 

Lee, A., Prytherich, R., King, A. (2000). U-Matic preservation. In M. Aubert & R. Bil- 
leaud (Eds.), Image and sound archiving and access: Challenges of the 3rd millennium (pp. 
177-186). Paris: CNC. 

Mather, P. B., & Neil, D. C. (1998). Assessing video tape degradation. In Proceedings, 32nd 
Annual 1998 International Carahan Conference on Security Technology (pp. 220—224). New 
York, NY: IEEE. 

Measuring tape. (1990, November). Video Review, 11(8), 31—36. 

Osaki, H., Oyanagi, E., Aonuma, H., Kanou, T., & Kurihara, J. (1992, January). Wear 
mechanism of particulate magnetic tapes in helical scan video tape recorders. [EEE Transac- 
tions on Magnetics, 28(1), 76-83. 

Rice, D., & Elnabli, S. (2010, October). Barcode scanners, miniDV decks, and the migration of 
digital information from analog surfaces. New York, NY: AudioVisual Preservation Solutions. 
Retrieved from http://www.avpreserve.com/wp-content/uploads/2010/1 1/Migration-of- 
Digital-Information-from-Analog-Surfaces.pdf 

Simon, S. (Interviewer) & Kugler, R. (Interviewee). (2008, December 27). Ode to VHS 
[Interview transcript]. Retrieved from NPR: http://www.npr.org/templates/story/story 
.php?storyId-98754557 

Thiebaut, B., Vilmont, L. B., Lavedrine, B. (2006, September 1). Report on video and audio 
tape deterioration mechanisms and considerations about implementation of a collection condition 
assessment method. Retrieved from http://www.prestospace.org/project/deliverables/D6-1.pdf 

Video Magazine. (1989). Video Magazine’ official blank tape tests. New York, NY: Reese Com- 
munications Inc. 

Woodcock, R. (1989). How to rescue a damaged tape. In Video Magazines oficial blank tape 
tests (p. 16). New York, NY: Reese Communications Inc.