MEMORANDUM OPINION AND ORDER
This case concerns the application of modern copyright law to designs embodied in an ancient medium: carpets. The plaintiff, David Shaw Nicholls, alleges that the defendants, Tufenkian Import/Export Ventures, Inc., d/b/a Tufenkian Carpets, and its principal, James Tufenkian, infringed his “Prado” design with two similar designs known as “Eclipse” and “Total Eclipse,” and thereby violated the Copyright Act of 1976, 17 U.S.C. § 101 et seq. This case initially went to trial before the *517 Honorable William H. Pauley III, U.S.D.J., and a jury, but a mistrial was declared when an overzealous juror conducted internet research related to the action. The parties then consented to proceed to trial before me pursuant to 28 U.S.C. § 636(c), and they elected to waive their right to a jury. Accordingly, I presided over a bench trial on November 16, 17, and 18, 2004, and this opinion constitutes my finding of fact and conclusions of law pursuant to Rule 52(a) of the Federal Rules of Civil Procedure. 1
Background
David Shaw Nicholls began designing carpets in about 1993 or 1994. (Tr. at 42). 2 He created a number of designs that incorporated circular motifs, including the Rav-enna, Sikri, Umbria, Cosmo-Cities, Constantinople, and Argyle designs. (Tr. at 48-49; PL Exhs. 10,12,13,15,16,17,101). According to Mr. Nicholls, one precursor of his Prado design was a carpet identified as Maui, which features large circles in which there is a spiral or scroll feature. (Tr. at 48-49; PI. Exhs. 14, 101). The Maui design progressed to the Rio, where the circles were made smaller and arrayed in a symmetrical lattice. (Tr. at 52; PI. Exh. 5). The Rio was never produced because Mr. Nicholls felt that it was not sufficiently unique. (Tr. at 52-53). Instead, he began attempting to soften the profile of the circles by transecting them, so that each half would be a different color. (Tr. at 53-54; PI. Exhs. 5 at DSN000054, 6).
This concept evolved, into the Prado, which can be described as including 20 equal size circles arrayed in five rows and four columns. At each end there is a rectangular band running the width of the carpet. At one end the band appears to cover the last row of circles so that only a row of hemispheres is visible. At the other end, the circles appear to overlap the rectangle. There are also rectangular bands of a different color running the length of each outer edge up to the point where they meet the bands running across the width. The lengthwise bands continue in a narrowed form across the width of the carpet just inside the bands at each end. These narrower bands are overlapped by the last row of circles at either end. Within these framing bands, the circles are arrayed on a background of a single color. Finally, each of the full circles is transected such that approximately three-quarters is one color and one-quarter is a different color. The transections are made width-wise, and they alternate so that if the one-quarter section is on the “bottom/’ in the first row, it is on the “top” in the next. The effect is as if each circle in one row is “facing” the corresponding circle in the adjacent row. (Pl.Exhs.1, 2,101). 3
The Prado was created in 1998 and went into production that year. (Tr. at 48). On January 18, 2002, Mr. Nichplls registered the design with the United States Copyright Office and received copyright number VAu 524-511. (Pl.Exh. 1). Together with the application, he submitted a wire frame diagram and a color reproduction of one sample of the rug. (Tr. at 45-46; PI. Exh. 1). The wire frame diagram is a *518 schematic drawing or blueprint that shows the precise dimensions for each geometric element and identifies in generic terms where each of four colors would appear. (Tr. at 45; Pl. Exh. 1). The color reproduction depicted the carpet as produced in the color way designated “Ermine.” (Pl. Exh. 1).
According to Mr. Nicholls, the Prado design received a good deal of publicity in the trade and in the general public. For example, it was depicted in a photograph accompanying an article in , the May 28, 2000 edition of the New York Times Magazine about a house that the architect Philip Johnson designed for Blanchette Rockefeller. (Tr. at 75-76; Pl.' Exhs. 35, 103). Mr. Nicholls created a “tearsheet” or advertisement from the Times Magazine piece and circulated it to perhaps 3,000 potential clients in July or August of 2000. (Tr. at 76-77;. Pl. Exh. 36). Similarly, he created a postcard of the Prado design and mailed out some 3,000 copies to his customer base in June 2000. (Tr. at 78-79; Pl. Exhs. 42, 43). Thereafter, the Prado appeared in the October 31, 2000 issue of Interior Design magazine. (Tr. at 78; Pl. Exh. 38). The Prado design was also displayed at the Architectural Digest trade show at the Jacob Javits Convention Center in Manhattan for four days at the end of October and beginning of November of 2000, and a poster for the show that included a Prado rug was disseminated to Mr. Nicholls’ customers some time in advance. (Tr. at 85-87; Pl: Exhs. 49, 101). Mr. Nicholls does not know whether Mr. Tufenkian ever saw an image of Prado before he created the allegedly infringing designs. (Tr. at 107).
Beginning in 1996, the parties had a business relationship pursuant to which the defendants manufactured carpets for Mr. Nicholls’ company. (Tr. at 87-88, 267; Pl. Exh. 50). The arrangement ended acrimoniously, however, when a dispute resulted in a lawsuit. (Tr. at 89-90, 267-70). That litigation ended in a settlement. (Tr. at 90, 267).
Mr. Nicholls first became aware of the allegedly infringing designs when he encountered an advertisement in the December/January 2004 edition of Elle Decor magazine showing a rug identified as Total Eclipse. (Tr. at 90-91; Pl. Exh. 52). He subsequently learned of the Eclipse design by viewing Mr. Tufenkian’s website. (Tr. at 96).
When Mr. Nicholls first saw Total Eclipse, he sent a letter to Mr. Tufenkian arguing that it was substantially similar to Prado and demanding that he cease marketing it. (Tr. at 91-92 & Exh. 53). A copy of this letter was sent to dealers whom Mr. Nicholls believed were Mr. Tuf-enkian’s customers. (Tr. 91-92 & Exh. 53). In response, Mr. Tufenkian’s attorney sent a letter to Mr. Nicholls’ counsel indicating that Total Eclipse was a variation of Eclipse, and that the Eclipse design was registered with the Copyright Office in early 2001. Mr. Tufenkian’s attorney expressed the belief that that date preceded the creation of Prado, and concluded that- “[o]ur concern, as you can imagine, is that the ‘Prado’ design may prove to have been derived from the ‘Eclipse’ design, which was registered long before the ‘Prado’ design was published.” (Tr. at 271-72, 275-76; Pl. Exh. 58).
At trial, Mr. Tufenkian testified that he had created the Eclipse design himself and, prior to doing so, had never seen Prado during the course of his business relationship with Mr. Nicholls or otherwise. (Tr. at 225, 240, 270). He did not see the New York Times Magazine article and never reads that publication. (Tr. at 265). Until this litigation began, he had never seen the postcard of the Prado design that Mr. Nicholls sent to his customers. (Tr. at 265). He did not view Mr. *519 Nicholls’ website until after he had received the cease and desist letter from Mr. Nicholls’ attorney. (Tr. a 265-66). And he did not attend- the trade show at the Jacob Javits Center in 2000 where Prado was displayed. (Tr. at 266-67).
According to Mr. Tufenkian, he designed Eclipse with the assistance of Elizabeth Artinian, an artist skilled in computer graphics. (Tr. 240-41). Previously, Mr. Tufenkian had marketed designs that incorporated rectangular grids, on which geometric forms like circles or representational figures such as fish or leaves were superimposed. These included the Tumbledown, Hothouse, and Lagoon designs. (Tr. at 242-44; Def. Exh. KKKK at 124-25, 136-37, 202-03). He was also becoming interested in incorporating more circular figures in his designs. (Tr. at 249-52). 4 And, as early as 1992, he had experimented with what he calls the “magic tape” effect in the Pasha Smoke design. (Tr. at 252; Def. Exh. CC). That concept also appears in his Stripe Rag Flag design. (Tr. at 251-52; Def. Exh. EE). The effect is created by “superimposing” a geometric pattern — generally of rectangles — over an underlying design. The colors of the underlying images “change” as they are viewed through different portions of the overlaid grid. It is as if the designer took an image and covered it with translucent colored tape or with rectangles of different colors of cellophane. (Tr. at 252).
In the summer of 2000, Mr. Tufenkian worked with Ms. Artinian in Armenia, presenting her with his ideas and then making suggestions based on the computer images she created. (Tr. at 253-55, 307-09). The result was a series of images showing a grid of twelve circles overlayed by various patterns of rectangles and incorporating the “magic tape” effect. (Tr. at 305-06; Def. Exhs. QQQ, RRR, SSS, TTT, UUU). 5
As finally manifested in Eclipse, the four rows of circles are evenly spaced one from the next, but the three columns are not; they are slightly asymmetrical. The circles are large and almost fill the rectangular space of the carpet. There is no “frame” because the circles extend to the edge of the rug on all sides, though rectangular bands that run widthwise at each end create a sense of boundary on those sides. The circles are transected by the rectangular shapes but not always in the same fashion: to the extent they are transected widthwise, some are divided once and others twice, and the proportion of each section varies from row to row. In addition, because some of the rectangular elements do not extend the full width of the carpet, some of the circles are transected lengthwise as well. (Def. Exhs. A, S, U, KKKK at 70-73). A total of eight colors are utilized, and because of the “magic tape” effect, both the circles and the “background” are composed of multiple colors.
Mr. Tufenkian testified that he went on to design Total Eclipse in approximately January 2003. (Tr. at 259). . He was attempting to modify the Eclipse design by making it less “edgy” and more suitable for a somewhat older market. (Tr. at 259-60,.264). He achieved this effect by eliminating the widthwise breaks in the rectangular bands. In other words, all of the rectangular elements now extended the en *520 tire width of the carpet. (Tr. at 261-63; Def. Exhs. C, T, KKKK at 94-95). Discussion
A. Validity and Scope of the Copyright
In order to establish copyright infringement, a plaintiff must demonstrate that he owns a valid copyright and that the defendant engaged in unauthorized copying of the copyrighted work.
See Jorgensen v. Epic/Sony Records,
In this case, it is undisputed that Mr. Nieholls received copyright Registration No. VAu 524-511 for the Prado design. The defendants suggest, however, that the copyright extends only to the “Ermine” color way, which was depicted in the photograph that was deposited with the Copyright Office along with a wire frame diagram. The force of this argument must be assessed in light of the dual purposes of the deposit requirement, which are: (1) to provide the Copyright Office with “sufficient material to identify the work in which the registrant claims a copyright” and (2) “to furnish the Copyright Office with an opportunity to assess the copyrightability of the applicant’s work.”
Data General Corp. v. Grumman Systems Support Corp.,
On the other hand, a copyright does not encompass designs that vary in essential respects from what was presented to the Copyright Office. Otherwise, the purposes of the deposit requirement would be nullified.
See Three Boys Music,
Registration creates a presumption not only that a copyright is valid, but also that the work is original.
See Boisson v. Banian, Ltd.,
Nevertheless “[s]imply because a work is copyrighted does not mean every element of that work is protected.”
Id.
Rather, protection extends only to “those components of a work that are original to the author.”
Id.
Thus, in this case Mr. Nicholls’ copyright does not provide him with an exclusive right to the use of circles in carpet design, because, although the Prado incorporates circles, it is not an original element. But a combination of otherwise unprotectible elements warrant copyright protection if the elements are selected,- coordinated, or arranged in an original way.
See Tufenkian,
B. Evidence of Copying
Having demonstrated that he owns a valid copyright, Mr. Nicholls must now establish the second element of copyright infringement: unauthorized copying. “ ‘Actual copying may be established by direct or indirect evidence.’ ”
Jorgensen,
1. Access
To support a finding of copying, “there must be evidence of a reasonable possibility of access. Access must be more than a bare possibility and may not be inferred through speculation or conjecture.”
Gaste,
Here, while it is a close question, Mr. Nicholls has not presented sufficient evidence of access. He notes that the Prado was displayed at an interior design show at the Jacob Javits Center. Yet, Mr. Tufenkian did not attend that show, and, in any event, the show took place after Mr. Tufenkian had designed Eclipse.
Cf. Odegard,
In sum, although images of the Prado design were disseminated, distribution was not so broad that Mr. Tufenkian was likely to have encountered it. This conclusion is reinforced by the fact that while Mr. Tuf-enkian’s Eclipse design was featured in advertisements and editorials in magazines such as Elle Decor, Architectural Digest, and Interior Design beginning in the fall of 2001 (Tr. at 232-34; Def. Exh. CCCC), Mr. Nicholls did not come across it until he first discovered Total Eclipse in December 2003. (Tr. at 207-08).
A final piece of evidence that rebuts any inference of access is the letter sent by Mr. Tufenkian’s counsel in response to Mr. Nicholls’ initial charge of infringement. (Pl.Exh. 58). Mr. Nicholls cites this letter as an admission by Mr. Tufenkian that the designs are similar, an issue that will be addressed below. But it is also relevant to the question of copying, since it would be curious indeed if Mr. Tufenkian, knowing that he had copied Prado when he designed Eclipse, would nevertheless authorize his lawyer to write a letter suggesting that the designs were similar.
2. Probative Similarity
Access alone, of course, would not demonstrate copying if the alleged infringer’s work did not resemble the copyrighted work. Therefore, to establish an inference of copying, the plaintiff must prove similarity as well as access. Similarity in this context is often characterized as “probative similarity” and must be distinguished from the “substantial similarity” necessary to prove infringement.
See Jorgensen,
The similarities here — the arrangement of transected circles in a grid pattern — are not so novel that coincidence could reasonably be ruled out. Even if the works in this case did cross the probative similarity threshold, this would not suffice to establish copying. “There is an inverse relationship between access and probative similarity such that the stronger the proof of similarity, the less the proof of access is required.”
Jorgensen,
C. Substantial Similarity
Even if there were sufficient evidence of copying, the works at issue are not substantially similar. That there are some differences is, of course, not dispositive. In determining substantial similarity, “ ‘[t]he key ... is ... the similarities rather than the differences.’ ”
Attia v. Society of the New York Hospital,
At the same time, one who copies the work of another will avoid liability for infringement to the extent that the new work
transforms the copied expression into a design that in some respects resembles the original, yet does not actually excerpt ... a more-than-de-minimis protected portion of the original. Such designs may be termed “inexact copies,” in recognition of the fact that they alter the prior image yet mimic its structure in some fashion.
Id. (footnote omitted). At most, Eclipse and Total Eclipse might be considered inexact copies , of Prado. Each design includes a rectangular array of transected circles. However, because these elements, taken individually, are within the public domain, the copyright protection afforded to their selection and arrangement is somewhat thin. Id. at 136 & n. 13.
Moreover, Mr. Tufenkian transformed these elements in significant ways in creating Eclipse and Total Eclipse. Where Prado incorporates twenty circles, Eclipse and Total Eclipse include twelve. The circles in Prado are smaller, creating more of a sense that they are independent of the background than is true of Mr. Tufenkian’s works. The transections of the circles in Prado are all proportional — each circle is divided approximately one-quarter/three-quarters — while the circles in Eclipse and Total Eclipse are transected in varying proportions and sometimes more than once. And, while the rows and columns of circles in Prado are evenly spaced, there is a slight asymmetry to the spacing between the columns in the Tufenkian designs.
Where, as here, a defendant’s work is dissimilar in the very respects in which it is claimed to be similar, that can obviously influence the conclusion as to whether the claimed similarity in fact exists or is substantial.' Furthermore, dissimilarity in realization or expression can make clear, as it does in this case, that the similarity extends only to un-protectible concepts or ideas.
Attia,
Of course, “while the infringement analysis must
begin
by dissecting the copyrighted work into its component parts in order to clarify precisely what is not original, infringement analysis is not
simply
a matter of ascertaining similarity between components viewed in isolation.”
Tufenkian,
As noted above, Mr. Nichqlls relies on the defendants’ response to his cease and *524 desist letter as an admission that the designs are similar. (Pl.Exh. 58). It is no such thing. The responsive communication simply points to a presumed chronology of the creation of the respective works and suggests that if they could be considered similar, then it is Mr. Nieholls who would be in legal jeopardy. The communication at issue makes no positive assertion of similarity.
D. Independent Creation
Finally, even if Mr. Nieholls had established access and substantial similarity, the defendants can rebut the inference of copying by proof of independent creation.
See Three Boys Music,
In general, if a party fails to raise an affirmative defense in its pleadings, the defense is waived.
National Market Share, Inc. v. Sterling National Bank,
In this case, there is no doubt that both parties understood that they were trying the defense of independent creation. At the first trial before Judge Pauley, counsel for the defendants alluded to the defense in his opening statement and presented extensive testimony on this issue. (Trl. at 26, 35, 241-53). Indeed, Judge Pauley instructed the jury on this issue without objection. (Trl. at 468). At the trial before me, defense counsel again addressed independent creation in his opening (Tr. at 21-26), and he again introduced evidence supporting that defense. (Tr. at 240-51). And, while plaintiffs counsel objected to specific exhibits of which he had not received adequate notice, he did not object to consideration of independent creation generally. Thus, the plaintiff impliedly consented to trial of this issue, and he has not been prejudiced. Accordingly, the answer is deemed amended to conform to the evidence by incorporating the defense of independent creation.
*525 The testimony presented by Mr. Tufenkian at trial on this issue was persuasive. He presented a coherent picture of the evolution of his thinking that resulted in Eclipse and Total Eclipse. Precursors of these designs such as Pasha • Smoke (Def. Exhs. CC & DD) and Stripe Rag Flag (Def. Exhs. EE & FF) embody the concepts of transecting geometric figures including circles and of overlaying the figures with the “magic tape” effect. While the plaintiff rightly points out that these designs have less superficial similarity to Eclipse and Total Eclipse than does Prado, this is simply the result of the grid pattern common to Prado and to the accused works. That does not undercut Mr. Tufenkian’s testimony with respect to how he developed his designs.
Moreover, Mr. Tufenkian’s testimony was corroborated by that of Ms. Artinian, who described the process of reducing Mr. Tufenkian’s concepts to computer renderings. By the time of trial, Ms. Artinian was no longer employed full-time by the defendants (Tr. at 298-99), and although she had done some free lance work for them (Tr. at 310-11), any personal interest she may have had in the litigation was extremely attenuated. The only significant inconsistency between her testimony and that of Mr. Tufenkian related to the time it took to design Eclipse: she testified that she worked on the project for only a day (Tr. at 311), while Mr. Tufenki-an described a design process lasting much longer. (Tr. at 240-45, 322-28). But this disparity is easily explained by their different roles. Although Ms. Artinian was involved in the final and most tangible aspect of the design process, Mr. Tufenkian had previously worked through a series of conceptual iterations before he presented her with sketches from which they drafted the final product together. Thus, there is ample evidence of independent creation that has not been effectively rebutted.
Conclusion
For the foregoing reasons, I find that the plaintiff has failed to demonstrate access or substantial similarity and that, in any event, the defendants have established independent creation. Accordingly, the plaintiff has not demonstrated that the Eclipse or Total Eclipse designs infringe his copyright for the Prado design. The Clerk of Court shall therefore enter judgment in favor of the defendants dismissing the complaint. 6
SO ORDERED.
Notes
.Prior to trial counsel submitted motions in limine and at trial they raised various eviden-tiary objections. Because this was a bench trial, I generally accepted the evidence, reserving decision on its admissibility. In this opinion I will only make those evidentiary rulings that are material to the determination on the merits.
. "Tr.” refers to the trial transcript.
. Defendants moved prior to trial to supplement the pretrial order to include among other things full-size samples of each carpet at issue. The carpets as manufactured are plainly relevant and no prejudice could result from their admission, so I am considering those exhibits.
. At trial, the defendants sought to introduce . source materials that Mr. Tufenkian allegedly relied on during the design process. (Tr. 245-49; Def. Exh. XXX). These exhibits were not identified in the pretrial order and were not produced in discovery and are therefore excluded from consideration.
. Although not identified in the pretrial order, these exhibits were admitted without objection in the first trial (Trial transcript for September 13, 14, 15, 16, and 17, 2004 (“TrL”) at 326-27), and will be considered here.
. Prior to the first trial, the plaintiff had made a motion for a preliminary injunction, and Judge Pauley had consolidated that motion with the trial on the merits. Accordingly, that motion is now denied.