The crossover in projection from 35 mm film and E-Cinema (Electronic Cinema) to professional grade D-Cinema (Digital Cinema) is almost complete. For the very first time in some parts of the world, cinema Ads are being beamed through the exact same digital projectors used for trailers and features. E-Cinema’s soft, muddy, and dull looking pre-shows and Ads can no longer hold back the true potential of carefully crafted content. And tellingly, film projector weave, scratches, dust, and stains can no longer be relied upon to disguise content imperfections.
In our homes we have become accustomed to TVs that seem to endlessly scale up in size, resolution, and contrast. Yet, due to highly compressed digital transmissions, we are subjected to TV content of compromised quality. The same Ads made for TV can be distributed visually lossless to cinemas. The much larger screens are rock-steady, cleaner, brighter, and crisper with many capable of 4K resolution with available surround sound. Little annoyances on a TV can now amplified into major disturbances on the big screen.
As content can now be created and delivered as equals, stakes for Ad campaigns are so much higher. Now, more than ever, one needs to be mindful of what audiences can expect from cinema Ads. Audiences can be wowed by Ads exuding cinematic attributes, yet overtly dismissive of Ads not measuring up to the virtues of this new delivery medium. As advertisers, we understand audiences are selective as to which message they find themselves willing to spend time on. Once captivated—and everything goes right—their experiences may delight. Yet some Ads barely make it beyond the :30 starting gate, as attempts to deliver big impressions flop big time. Even though the message may be compelling, a poor quality presentation is all it takes for an audience to blow right past the brand, setting the stage for them to engage in something less obtrusive.
Adapting TV Ads to film was formally the responsibility of the transfer facility and film lab. Having done away with film transfers and prints, costs are now much less. However, exhibitors and their Ad contractors put the burden on Ad agencies to submit Ads in direct compliance with their D-Cinema projection specifications. Not surprisingly, their media specifications are the exact same formats as derived by a transfer facility’s final data conform prior to actual film printing. To expect Agencies themselves to derive pristine cinema Ad deliverables is a tall order, adding complexity, costs, and a lack of transparency in quality. By not knowing what kind of attention your Ad needs to work most effectively on the big screen, the riskiest thing you can do is think you’re playing it safe by allowing your in-house editorial or post facility push things as best they know how. Often conspicuously absent from the media buy is your true costs to conform Ads to the their specifications. By not being held accountable for quality issues on projection, exhibitors clearly do not care how you get there, so long as your Ad can be “wrapped” by them into a DCP (Digital Cinema Package) for projection.
Even if exhibitors were to get pristine TV Ad deliverables, one should question what motivates them to set your Ad apart from other Ads. Interestingly, in “wrapping” Ads into a DCP, it is all too common for exhibitors to choose a JPEG 2000 compression ratio dramatically higher than the optimum setting. As DCP file sizes are reduced, distribution across networks and ingest into D-Cinema servers speed up, saving time and money. The trade off ? Just passable results, or even worse a tainted brand.
To be clear, digital projection can be made to look far worse than what was previously available on film. Or it can be made to look far better “wrapped” elsewhere by those who have your best interests in mind; who are intimately familiar with the strengths and pitfalls of the new projection medium; that have tools at their disposal to detect and correct flaws in all but the most carefully crafted cinema Ad masters.
So when good enough is actually not good enough, Digital Devoid’s contribution is clear: we understand your brand begs attention; we know what works to help captivate audiences’ undiluted attention; and—most of all—what it takes for your big idea to become an even bigger idea when projected. For Ads as entertainment, we believe God is in the details.
1 ) How do you convert a television commercial for cinema release?
Submit your Ad through the internet, as data files, or on videotape, and we will transfer your Ad to any movie theater projection format. Among other image parameters, the frame rate and picture framing are adjusted to comply with the target medium. When feasible, audio is adjusted from stereo to full surround sound mix.
Digital Devoid specialize in D-Cinema Ad adaptations. We are pioneers in this field, and that’s all we do: TV Ads mastered for D-Cinema. We work with any movie theater chain or cinema media firm in the US or internationally. Just send us your Ad and we will take care of the cinema Ad deliverable.
2 ) In what format are movie theater commercials played?
Cinema Ads are projected digitally through D-Cinema and E-Cinema systems before the main feature attraction. Media is commonly distributed to screens via satellite, internet, or on disk.
Movies are primarily lensed for release in one of two formats: Cinema standard (1.85:1 Flat), or CinemaScope (2.39:1 Scope). CinemaScope movie features require Scope Ads, while Flat features require Flat Ads. So if the media buy covers cinemas projecting both Flat & Scope features, two differently formatted masters are needed.
D-Cinema refers to the digital cinema projection standard established by the Digital Cinema Initiative (DCI). A DCP file is delivered to the movie theaters in 2D or stereoscopic 3D. It can be downloaded by the theater chain or shipped to them on removable media such as USB memory stick or CD/DVDs. The DCP file essentially mimics the characteristics of 35 mm film and incorporates strict encoding specifications. JPEG 2000 compression is utilized for an encoding bandwidth of up to 250 Mb/sec. The DCP aspect ratio is most typically 1.85:1 or 1.77:1 Flat, or 2.39:1 Scope.
E-Cinema generally refers to inferior electronic projection formats, ranging in resolution from Standard Def to Hi-Def. E-Cinema Ads often run at a video frame rate of 60 interlaced fields per second. In the US, the vast majority of cinema Ads are MPEG-2 encoded at 720p (1280 x 720 pixels) for a maximum bandwidth of 20 Mb/sec. The content is then bumped-up to 1920 x 1080 pixels progressive scan at the DCI compliant projector. Though the projection quality of E-Cinema is not as good as 35 mm or D-Cinema, advertising media firms embrace them for their flexible designs and reduction of cost to the local advertiser.
You need to find out what format, or formats, your cinema chain or media buy requires.
3) What source materials are needed for a cinema Ad transfer?
We can adapt any format to cinema.
However, some materials are better suited than others. Because of cost constraints, E-Cinema Ads often run at a captured frame rate of 29.97/30/59.94 FPS (frames per second). At these frame rates, objectionable interlacing may well be visible as motion smearing on the big screen. As resolutions of projectors are relatively low, on close-up screen inspection the imaging arrays’ grid map (the density of individual pixels) often can be identified.
SMPTE, the newer DCI specification, is capable of encoding and projecting DCPs at 24, 25, 30, or 60 FPS progressive scan. For Ads to be projected through the older DCI standard, the MXF-Interop, frame rates must first be converted to 24 FPS progressive scan. All SMPTE DCI compliant D-Cinema servers are backward compatible and support MXF-Interop DCPs, but some older D-Cinema MXF-Interop servers are not upgradeable to play SMPTE DCPs. And as Hollywood features are shot, finished, and distributed at 24 FPS, many cinema chains opted out of upgrading older MXF-Interop equipment to SMPTE compliance. So, to prevent incompatibilities at the server ingest stage, most cinema Ad exhibitors require 24 FPS DCPs. Indeed, if you are converting to 24 FPS from any other frame rate, we suggest allowing us frame rate convert during the DCP mastering process. The quality will be much, much better through our proprietary software DC Master 4K™.
Whenever going from smaller to larger screens, such as NTSC / PAL or Hi-Def to 4K DCPs, as much of the original picture detail should be preserved. For example, it is OK to place data on a DVD data disk but it is not advisable to encode the picture and sound for DVD playback. DVDs use MPEG-2 or H.264 encoding, where compression ratios may average as high as 30:1. BetaSP tape, on the other hand, is uncompressed (a compression ratio of 1:1). The DBeta format employs 3:1 compression. Send uncompressed (lossless data compression) materials whenever possible. If an NTSC Ad originated from a PAL source, then we would prefer making the DCP adaptation directly from the PAL source. In the event the PAL to NTSC converted Ad was subsequently edited in NTSC, we can still use the PAL source to reconstruct the NTSC edit in PAL. The PAL conformed Ad will have a higher native resolution than the NTSC edit, and a camera shutter angle closer aligned to the 24 FPS projection standard in cinemas.
Here are some guidelines if planning for a DCP:
DCPs can only be created once an item in each A) and B) below are valid. However, MXF-Interop DCPs must match B) 1) 24 FPS.
A) Materials delivered at one of following 2K resolutions (in any ONE of two dimensions) –
1) 2048 x 1080 pixels (Aspect ratio – 1.90:1, Full Container, rarely used)
2) 1998 x 1080 pixels (Aspect ratio – 1.85:1, D-Cinema standard)
3) 1920 x 1080 pixels (Aspect ratio – 1.77:1, Hi-Definition)
4) 2048 x 858 pixels (Aspect ratio – 2.39:1, CinemaScope)
B) Materials delivered at any ONE of the following PROGRESSIVE frame rates-
1) 24 FPS
2) 25 FPS
3) 30 FPS
Here are some guidelines when planning for 35 mm prints, in order of preference:
1) DCP Standard 1998 x 1080 pixels (1.85:1); Film Standard 1828 x 988 pixels (1.85:1) at 24 FPS
2) 1080p, 1920 x 1080 pixels (1.77:1) at 24 FPS
3) 720p, 1280 x 720 pixels (1.77:1) at 24 FPS
4) PAL, 720 x 576 pixels (1.33:1, non-square pixel) at 24 FPS
5) NTSC, 720 x 486 pixels (1.33:1, non-square pixel) at 24 FPS
Here are some image formats we recommend, in order of preference:
1) An image sequence with 10, 12, or 16 bit pixel depths (.DPX, .TIF, or .PNG)
2) Quicktime movie codecs (10 bit uncompressed):
a) AJA Kona (RGB); Bluefish444 (RGB); Blackmagic (RGB)
b) Avid Packed (RGB)
c) Bluefish444 (YUV)
d) Avid DNxHD 444 (RGB, quality indicator set to maximum bit rate)
e) Uncompressed 422 (YUV)
f) Avid 1:1x (YUV)
g) Avid DNxHD 220x (YUV, quality indicator set to maximum bit rate)
h) Apple ProRes 444 (YUV)
3) Quicktime movie codecs (8 bit uncompressed)
i) Animation codec (RGB)
j) Uncompressed 422 (YUV)
k) AJA Kona 2vuy (YUV)
l) Avid DNxHD 220 (YUV, with quality indicator set to maximum bit rate)
m) Apple Motion JPEG A (YUV, with quality indicator set at 100%
4) Quicktime movie (10 or 8 bit compressed codec):
n) JPEG 2000 (RGB)
o) Apple ProRes 422 HQ (YUV)
p) Avid DNxHD 220 (YUV, with quality indicator set to maximum bit rate)
q) Apple ProRes 422 (YUV)
r) Avid DNxHD (YUV, with quality indicator set to maximum bit rate)
s) H.264 (with quality indicator set at 100%, a setting of 5 in some applications)
Note: If you can only send a compressed QuickTime file with lossy data encoding, choose JPEG 2000. We would like to note that many Ad agencies deliver materials in H.264. We want you to know that this may be OK—we even make DCPs from H.264. So long as the quality level is set to the maximum, a level of 5 in some applications. But of course there are preferable formats in the list above.
We can still make the conversion if you are unable to provide materials as specified above (actually, we always recommend submitting media at at the ORIGINAL capture frame rate and resolution). Ads can be delivered to us in any format, at any pixel aspect-ratio, frame aspect-ratio, resolution, color space gamut, gamma, and frame rate. But first we need to prepare a Digital Cinema Distribution Master (DCDM) through DC Master 4K™, our proprietary software. DC Master 4K™ addresses: contrast; brightness; gamma; saturation; color-space gamut; framing; resolution; sharpness; dirt, dust, scratches, and stains; noise and grain; interlacing; and compression ratios in the DCP. Through frame rate conversion with directional blur control, motion portrayal in the DCP is adjusted for fluidity in that the camera shutter angle of filmed footage emulates the projector’s electronic switching aperture (“dark time” interval). And where necessary, custom made filters are readily designed and applied.
4 ) Is a new audio mix needed for the cinema version of your TV commercial?
It depends on the mix and your budget.
Ideally, sound for cinema is comprised of a 5.1 (Left, Center, Right, Low-frequency effects, Left surround, Right surround), 7.1 (L, C, R, Lfe, Ls, Rs, Back surround left, Back surround right), or Dolby Atmos mix. The idea is to take advantage of all speakers in the theater.
However, when a TV stereo fullmix is played directly in cinemas, it usually only plays from the three front speakers. And if there is not enough separation between the tracks, or if they are mono, sound will collapse to the center front speaker only. Though the levels may be correct, the soundtrack will be perceived to play soft since it is only being heard from one speaker. So when there is too little separation between the stereo channels we recommend creating a cinema 5.1 mix from the stereo track. Through Digital Devoid, deriving a 5.1 mix from a TV stereo mix can yield truly impressive results.
The majority of high quality QuickTime codecs support stereo, 5.1, and 7.1 embedded audio. Generally, it is best for us to use the audio directly from the QuickTime in order to avoid synchronization problems. If audio is delivered as separate stems (dialog, music, and sound effects track files), a client supplied lo-res QuickTime comes in handy to verify sync.
5 ) How do you prepare sound for D-Cinema Ads?
Just like 35 mm film Ads, sound for digital Ads should ideally take full advantage of the 6, 8, or more (Dolby Atmos) channels available in movie theaters.
DCP’s sound should constitute a 5.1, 7.1, or Dolby Atmos mix. We can also use a stereo track for the DCP, however we recommend allowing us to create a 5.1 mix from all stereo sources. Sound always starts on first frame of picture and ends on last frame of picture, so there is no need to include 2-pop.
There are a wide variety of E-Cinema audio formats in use. Generally speaking, the Hi-Def E-Cinema formats sound good with 5.1 surround. Others are less sophisticated and only support stereo fullmix or stereo Lt&Rt (Dolby Pro-logic II surround sound matrix encoded). Generally, E-Cinema audio is relatively easy to supply. So, find out from your exhibitor what formats work best for them.
Submitting ProTools sessions is encouraged. If D-M-E stems or stereo fullmix are submitted, use .aif or .wav files with 16 or 24 bits at 48 kHz sampling rate. Limit average program levels to -10 dB, and peaks to -4 dB. D-M-E
6 ) How do you prepare the sound for 35 mm cinema Ads?
If you send us a videotape and only require two channel stereo sound, we can use the audio off tape. But if you are submitting separate sound files, for either stereo, 5.1 or 7.1 surround mix, use .aif or .wav mono files with 16 or 24 bits, at 48 kHz sampling rate. Each track should contain “2-pop”. Limit average program levels to -10 dB, and peaks to -4 dB.
The most common sound formats on film are Dolby Digital and Sony SDDS. Less prevalent is DTS. These are digital surround sound film formats created from a 5.1 or 7.1 mix. The mix is encoded to film as an analog and digital track. For the analog track, a Left Total and Right Total (Lt & Rt) matrix encoding is derived from the 5.1 mix. On projector playback, it is decoded as a pseudo surround mix (Dolby Pro Logic II) and played through the surround (L,C,R,Ls,Rs) speakers. The digital track has the further benefit of a low-frequency effects (Lfe) channel. Overall, the digital track provides a more distinct separation of the surround channels and yields a more concise timbre, wider frequency range, and greater dynamics. When a digital track is shot to film, the analog track is always added as a fail-safe option in case the projector’s digital sound reader has difficulties at any point decoding the digital track on the film print.
When sending audio for 35 mm film printing, always place “2-pop” EXACTLY 2 Seconds before first frame of picture. “2-pop” is one frame ONLY of 1 kHz tone located EXACTLY 2 Seconds prior to first frame of picture. THIS IS ALWAYS REQUIRED for 35 mm cinema film transfers. See the following blog post for a more detailed description of 2-pop: “The beep, audio 2-pop on film”.
7 ) Which quality level of Transfer or Mastering service should I use?
Content adhering to the specifications in section 3) above can be directly encoded into a DCP. When content falls outside those specifications, first a DCDM is created which in turn is encoded into a DCP. DC Master 4K™, our proprietary software, is utilized to create all DCDMs and encode DCPs.
We offer two levels of service for 35 mm film transfers:
FP Transfer 2K™ features the benefits of frame-by-frame processing at 2K resolution.
DP Transfer 4K™ goes further by offering deep sub-pixel processing at true 4K resolution.
All our quality conscious customers use DP Transfer 4K™, it is by far the best in the industry. But not all Ads will fully benefit from all the features available through DP Transfer 4K™. So allow us to make a recommendation upon review of your source materials.
8) Can a 2D DCP be played in a 3D theater?
Yes, a 2D DCP cinema Ad can be viewed during the pre-show of a 3D stereoscopic feature film, even with 3D glasses on, albeit with less screen brightness. This is an OK way for you to achieve the best placement of your Ad without incurring the significantly higher costs of producing a full 3D stereoscopic commercial. Your spot can now be seen right before the movie trailers begin.
9 ) How do you create a 3D stereoscopic DCP?
DCI conversions for cinema ads in 3D can be created by following the below specs:
a) Both left & right eye streams need to be submitted
b) Ideally, frame size to be at the DCP resolution, 12 bit uncompressed, 24 FPS progressive
c) A cinema surround sound mix should be made, either 5.1 , 7.1, or Dolby Atmos. If D-M-E stems are submitted, we prefer discreet mono .aif or .wav files, 16 or 24 bit, at 48 kHz sampling. ProTools sessions are also accepted.
If we receive left & right eye footage that does not meet the above specifications, then we can adapt the footage to the 3D DCP specification. This involves creating a 3D DCDM which in turn is used to encode a 3D DCP.
The 3D DCP will play on all 3D systems including RealD, Dolby 3D, and IMAX 3D. And if needed, we can use one of the streams to generate a 2D DCP or transfer to 35 mm film.
10 ) How long does it take to convert movie theater Ads?
Digital cinema mastering for TV Ads is typically 1-2 days. 3D stereoscopic D-Cinema conversions and custom DCP adaptations may take a day or two longer.
These are generally video format standard conversions and can be completed in one day.
35 mm film transfers:
Tape to film transfers for cinema Ads usually take three business days. Printing film copies can take another 1-2 days.
11 ) Can I approve the cinema transfer?
Yes, you can approve our work. Since we are mostly concerned with quality, we highly recommend that your Ad be screened on the largest screen possible, like at your local MegaPlex cinema. Small to average sized screens will not allow you to see the image in great detail.
DCPs can be screened at theaters with the appropriate servers and projection equipment. You can preview DCPs on smaller monitors in a studio but we do not recommend this approach as you will not be viewing what your target audience will see on the big screen.
These spots can also easily be viewed on studio monitors. Yet you will not see true results unless you screen at a movie theater under the same condition as would your target audience.
35 mm film:
The lab will create a Check Print before moving on to final release printing. On about 95 Percent of our work we assume all liability in approving the Check Print. But if you are interested in reviewing the spot yourself, we can ship you the Check Print before printing additional copies. If the media buy is from the same chain as your local screen, arrangements could be made to screen the Check Print at no extra expense.
12 ) What are the best materials for a cinema transfer?
Generally speaking, shoot and finish the Ad at 24 FPS. Later, a 3:2 pulldown can be added to satisfy any broadcast media deliverable. Keep in mind that framing in cinema is 1.85:1 Flat, or 2.39 Scope. Allow for at least 12 percent title safety area around the frame boundaries as movie theaters are notorious for cutting off part of the picture on projection.
Review the following blog post, “How to optimize tape to film transfers“, for more details.
13 ) Flat vs. Scope DCPs
The cinema Ad media buy, to a large extent, dictates whether both Flat and Scope DCPs are needed. Yet even when a media buy includes showing your Ad prior to both Flat and Scope features, some exhibitors are reluctant to run Scope Ads as they add to operator confusion and errors at the D-cinema ingest stage and thereafter on projection. Ingest operator proficiency will improve as more and more Ads are being lensed in the CinemaScope format.
If an Ad was lensed Scope, then a Scope DCP would be an ideal fit. On projection it would fill the entire screen just as do Scope trailers & features. However, adapting Scope lensed footage to a Flat DCP is quite challenging due to the confusion it may cause the projectionist.
This discussion focuses on the potential outcome of two fundamental choices available in the way Scope lensed Ads are adapted to Flat DCPs. It may be useful to first review the 3 DCP Formats as specified by the DCI. For 2K (For 4K, double these resolutions):
1) Flat – footage EITHER 1080 pixels high AND/OR 1998 pixels wide (1.85:1 Aspect ratio)
2) Scope – footage EITHER 858 pixels high AND/OR 2048 pixels wide (2.39:1 Aspect ratio)
3) Full container – footage 1080 pixels high AND 2048 pixels wide (1.90:1 Aspect ratio)
All DCI compliant D-Cinema servers/projectors support these three DCP Formats. But it should be noted that Flat and Scope DCP resolutions only need satisfy one, and not necessarily both, image dimensions. Full container DCPs, on the other hand, need to be encoded at both full height and full width dimensions.
The actual imaging array (DLP) of the projector is the Full container size: 2048 x 1080 Pixels for 2K resolutions, and 4096 x 2160 pixels for 4K resolutions. For dual use (Flat and Scope) cinema screens, the actual screen’s width is set to the full width of the imaging array (2048 Pixels), and the screen’s height is set to the full height of the imaging array (1080 pixels). When Scope DCPs (2048 x 858 pixels) are projected on these dual use screens, the footage is displayed from the bottom of the imaging array. Here the screen’s top curtain drops down 20 Percent, that is from full height (1080 pixels) to the equivalent of Scope height (858 pixels). So, in this projector configuration, Scope (2.39:1) DCPs fill 100 percent of the screen width and only 80 Percent of the available screen height, while Flat (1.85:1) DCPs fill 98 percent of the available screen width and 100 Percent of the screen height.
Standalone CinemaScope rooms are totally different. Here the projector is configured in a dedicated screen ratio of 2.39:1 (2048 x 858 pixels). The projector resizes (stretches) the height of the Scope DCP from 858 pixels to 1080 pixels, and projects the 2048 x 1080 pixels (1.9:1) footage through an anamorphic lens which normalizes (unstretches) the footage back to the correct proportion of 2.39:1. By making full use of the entire imaging array, about 20 Percent more light is projected onto the screen. As less power is used to achieve the desired screen brightness, the life expectancy of an expensive projector lamp is extended.
So, from the above discussion it should be clear that your Scope DCPs may be projected on dual use (Flat and Scope) screens, AND/OR on dedicated CinemaScope screens.
Ads are ingested into, and played back from, a D-Cinema server “playlist”. To make the process of ingest uncomplicated as possible, a playlist is usually comprised of a preformatted template: EITHER Flat, OR Scope, OR Full container. Each template supports different image dimensions and maps the picture to a specific area on the DLP (Digital Light Processor). A playlist is populated by the various Ad DCPs. But only DCPs in the same format as the playlist template will project correctly. In other words, there is nothing preventing an ingest operator from importing a Scope DCP into a Flat playlist. It will project, but not correctly. The same can be said for ingesting a Flat DCP into a Scope playlist, it projects, but not correctly.
Problems seem to arise when Scope Ads are projected on dual use screens. On ingest, some D-cinema servers may even recognize that the DCP resolution does not comply with that of the template’s minimum and maximum resolution limits, and will prompt the operator to either reject the DCP or select a framing option (crop, fit, pan, zoom, distort, etc). As can be imagined, this is where the inexperienced ingest operator can get very confused: producing unexpected consequences on screen by not paying close attention to the actual NAMED format of the DCPs.
As TMS (Theater Management Systems) and D-cinema servers/projectors do not differentiate between Flat, Scope, or Full container DCPs, it is up to the ingest operator to recognize the DCP format and manually enter the DCP into a playlist comprised of the appropriate template. On ingest, the operator needs to explicitly use the DCP folder/directory and DCP filename themselves to interpret the format of the DCP. To prevent confusion on ingest, Digital Devoid’s DCP folder/directory and DCP file are named identically.
The _S_ in “CC_Emocion_ADV_S_SP-XX_US-GB_5.1_2K_20150319_DDI” represents a Scope DCP
The _F_ in “CC_Emocion_ADV_F_SP-XX_US-GB_5.1_2K_20150319_DDI” represents a Flat DCP
The _C_ in “CC_Emocion_ADV_C_SP-XX_US-GB_5.1_2K_20150319_DDI” represents a Full container DCP
There are two ways to deliver flat DCPs: 1) Full container DCPs, or 2) Flat DCPs
Full container DCP:
A Full container format requires BOTH dimensions be filled. So here our Flat DCPs are actually delivered in the rarely used (and equally unpopular) Full container format. When Scope lensed Ads are adapted to flat DCPs and delivered as Full container DCPs, some ingest operators may notice black padding (up to 21 Percent of screen height) top and bottom of the live action footage (2048 x 858 pixels, 2.39:1) and mistakenly interpret it as a Scope DCP!
However, even if the Scope lensed live action footage (1998 x 836 pixels, 2.39:1) were delivered as a Flat DCP (with up to 21 Percent black padding top and bottom), the Flat DCP (1998 x 1080 pixels, 1.85:1) may still be incorrectly interpreted by the ingest operator as a Scope DCP! The choice of creating a Flat DCP with no padding, at a valid resolution of (1998 x 836 pixels, 2.39:1), may induce the same confusion on ingest.
In any of the above cases, one may receive complaints from the client claiming the exhibitor was unable to use the DCPs, or the Ads did not look correctly formatted on projection.
And the problem does not necessarily end here. To reinforce the dilemma faced at the ingest operator stage, let us take a look at another example. When adapting Hi-Def (1.77:1) footage to a Scope DCP (2.39:1), up to 26 percent of the Scope’s DCP width may need padding with black (left & right of action). We could deliver the Scope adaptation as a valid 1518 x 858 pixels (1.77:1) Scope DCP. But on projection, the projectionist may be confused when seeing that as much as 26 percent of the screen width is black. The projectionist may question whether the footage was ingested using the incorrect template! On trying to ingest the same Scope DCP through a Flat template, neither width nor height would be valid and the ingest would most likely be rejected.
To re-iterate, by creating a Full container DCP of Scope lensed footage to fulfill the flat DCP deliverable mandate, the entire height and width by DEFINITION have already been pre-filled. So one would hope that any “interpretative” ingest operator error could be completely avoided by staying with Full container DCPs. But based on the feedback we have received, this proves not to be so in all cases! In hindsight, due to the rarely used Full container playlist template, what may be occurring in some cases are erroneous operator ingests of a Full container DCP into a Flat template playlist. And from an ingest operator’s perspective, doing away with the inconvenience to cue a single Ad into a separate Full container template playlist ahead, or behind, a Flat template playlist containing all the other Ads would be a welcome relief.
Unfortunately, neither method of framing Scope lensed Ads into flat DCPs is full proof at the ingest stage. Yet there are no other alternatives. But as the d-cinema realm matures so too should the experience of ingest operators, with lower prevalence of ingest errors no matter Flat or Full container DCPs.
Rest assured, we always fill the cinema screen with as much action as possible, should it entail cropping, zooming, panning, or distorting (stretch / squeeze) without compromising the artistic intent.
14 ) 2K vs. 4K DCPs?
DCI compliant D-Cinema servers originally only supported 2K resolution (2048 x 1080 pixels). With the advent of 4K acquisition, the DCI extended support to 4K resolution (4096 x 2160 pixels). Effects and editorial facilities have embraced the 4K challenge, as have select D-Cinema servers and projectors. A 4K picture frame contains four times as many pixels as does a 2K frame.
To ascertain as to how the movie attendee benefits from a 4K projection, consideration must be given to the distance of the viewer to the screen. The viewing distance of an attendee to a movie screen is measured in screen heights. Modern day movie screens are designed so that the seating is populated between 0.8 to 3.5 screen heights. For many of us, the most immersing experience may be in a seat at a distance of about 1.5 Picture heights. So for an average screen height of 22 feet, the seat would be located approximately 33 feet from the screen.
At this distance, the screen subtends a relatively large vertical angle of view: 37°. SMPTE standard EG-18-1994 recommends a minimum viewing angle of 30° for movie theaters. With those having 20/20 vision, the smallest detail that can be resolved through the fovea of the eye subtends an angle of about one arcminute, or 60 pixels per degree (equivalent to 30 line pairs per degree). When multiplying the angle of view by the number of pixels per degree, we arrive at the number of vertical pixels required to display an image in sufficient detail. At 60 Pixels per degree, a 37° angle of view requires 2220 pixels to cover the screen height.
Where does this vertical pixel count fall in relation to the 2K projection standard? The DCI 2K pixel resolution is 1998 (width) x 1080 (height). Furthermore, CinemaScope projection may only cover 858 Pixels of screen height. At the “ideal” distance from the screen, 2K digital projection is clearly a gross compromise of viewer expectations for detail. Here the viewer may well discriminate on screen the pixel grid of the imaging array. In fact, only by having our angle of view reduced to 31° in a seat distanced 3.16 picture heights or further back would our angular discrimination be satisfied. Beyond this threshold of visibility, our visual acuity diffuses the pixel grid where even the increased pixel density of 4K projection will go undetected.
Even at 1.57 Screen heights and closer, 4K projection may be unable to satisfy our craving for pixel density. However, every attendee seated between 1.57 And 3.16 picture heights would benefit from being beamed up 4K (3996 x 2160) pixel data. In fact, for a full house of 20/20 Vision attendees seated between 0.8 to 3.5 Screen heights, only some 13 Percent may be satisfied with 2K projection, 72 Percent satisfied with 4K projection, but a full 87 Percent will still benefit from 4K projection (yet the same 87 percent may be dissatisfied with 2K projection). So it should be clear that so long as the audience is seated within their threshold of visibility, the 4K advantage is equally meaningful on screen sizes large or small.
Irrespective of source material resolutions, Digital Devoid’s 2D DCPs are mastered to 4K resolution (by design, all 3D DCPs are limited to 2K projection). All DCI compliant cinema servers have facilities to sub-sample 4K data to 2K data in the event 4K projector resolutions are not supported. On the other hand, 4K projectors being fed 2K data need to up-sample 2K to 4K internally.
Irrespective of a 2K or 4K encode, the maximum bandwidth limit of any DCP is 250 Mb/sec. So depending on the nature of footage in the Ad (CGI textures, live action, noise floor, etc), one needs to weigh in on the benefits of preserving greater spatial detail of 4K versus any trade off of having to use higher compression ratios in sustaining 250 Mb/sec encodes.
15 ) How much does a cinema Ad conversion cost?
Pricing is based on the duration of your commercial, the type of service chosen, the different Ad formats required, and the number of dubs required. There is often a reduction in total cost when there is an overlap of different versions for regional Ad campaigns.
The process of DCP mastering of Ads can be quite different from the more mundane task of feature length DCP mastering. Ads are often effects laden with multiple elements and layers originating at different frame rates. Ours is a time consuming, hands-on, process. But through our proprietary software—specifically designed for correcting the problems encountered when re-purposing content for the big screen—you are guaranteed the highest quality cinema DCPs.
Please contact us for pricing on specific projects.
Here is a recap of work on cinema deliverables for Samsung’s 3D (Stereoscopic ) Ad. This is a great looking Ad from the first company to launch a 3D Ad campaign in cinemas.
Title: Samsung 3D LED TV “A New Dimension :60”
Agency: DLB Group
Our client needed 35mm prints and digital versions for Central and South American cinemas, in Spanish and Portuguese. This might sound like a simple request, but it was not so easy to accomplish, correctly. Particularly when the only source material accessible was a DCP (Digital Cinema Package) formatted for CinemaScope aspect ratio, in English. We used our proprietary DC Master 4K™ and DP Transfer 4K™ software to create and deliver precisely calibrated Flat and Scope 2D 35mm film prints, and Flat and Scope 2D and 3D DCPs for various regions. (continue reading…)
Unlike most feature length movie projects, an Ad edited for TV rarely gets detailed planning and preparation for cinema release.
Digital Source Masters of TV Ads are notorious for containing a myriad of issues that complicate a smooth conversion to cinema. Aspect ratio adaptations need to be made, color attributes may need adjusting, the audio track remixed to 5.1… the list goes on. Making sure you resolve these issues at the first stage, in the DCDM (Digital Cinema Distribution Master), will ensure that the subsequent DCP (Digital Cinema Package) will look great.
Perhaps the most difficult issue involves variable frame rates. The industry standard DCP runs at 24 FPS (progressive frames per second) as does film. So before a DCP is made you need to create a DCDM that runs smoothly at 24 FPS. But often during the Ad editing process scene elements are introduced that disrupt motion patterns. For example, the 3:2 cadence—which is normally applied through telecine—may be broken; a graphic tag rendered at 60 Interlaced fields per second may have been added; text running at 30 FPS may have been composed on top of live action running at 24 FPS; the spot may have been converted from PAL’s 25 FPS to NTSC’s 30 FPS introducing inter-frame ghosting. These and other temporal distortions are largely disguised by the high refresh rate of the interlaced TV display. However, in progressive scan film projection, these same distortions are immediately apparent.
Once you reach the DCP mastering stage it is too late to make corrections. Take care of it beforehand, in the DCDM.
Here is a VISA Ad we worked on. It’s a good example of simultaneous conversions to various cinema formats including 35 mm film, D-Cinema, and E-Cinema. And since there’s a lot of overlap in the scope of work, the client realized significant cost savings by ordering all formats at the same time.
Title: VISA “Debito :30”
Agency: TBWA Latin America
We were given a DigiBeta master at standard def, framed for television, running at 60 FPS interlaced, with stereo sound. We remixed the soundtrack to 5.1 Surround, and through our proprietary software converted the picture to the following cinema deliverables:
1) 35 mm Academy Flat (1.85:1 aspect ratio) release prints with Dolby Digital (DP Transfer 4K™)
2) D-Cinema DCP for IMAX with 5.1 Surround soundtrack (DC Master 4K™)
3) 2K Uncompressed QuickTime at 24 FPS for E-Cinema projection (XR Convert™)
The key to any high quality film-out or DCDM (Digital Cinema Distribution Master) is proper preparation of the picture sequence. (continue reading…)
An artist’s canvas is the reference standard!
In the absence of the creator’s monitor, digital artwork is inadvertently susceptible to being perceptually distorted through any number of stages of production in preparation for release through disparate media.
So a reference standard for those who view the artwork is essential if the original artistic integrity is to be upheld. Compliance to standards and recommendations for viewing a medium is common within isolated professions, where—through common perception—reproductions can be consistently matched to the original rendition. However, practical guidelines for cross-media compliance are scarce and in the absence of the creator, codifying and analyzing artwork is often contained in the eye of the beholder!
In today’s multifaceted era of digital distribution, having a single display calibrated to reflect the contrast range and color temperature of each medium may prove indispensable.But this luxury is simply unattainable. Actual viewing conditions differ for various media so the WYSIWYG display of one particular medium is unlikely to match that of another.
For example, the American National Standards Institute (ANSI) recommends viewing color at a daylight temperature of 5,000 Kelvin (K) white point. This is both the film print projection and pre-press standard. But the color temperature found in most computer displays is closer to 9,000 K which is substantially bluer and colder than television’s 6,500 K standard. (continue reading…)
Light is emitted off the monitor, transmitted through film, and reflected off paper. Color is an additive process on film and in the monitor where primary colors are combined to produce white light. Ink on paper reflects light, and the primaries are subtracted produce black—well not quite. In fact, owing to the impurities in printing inks, they produce a muddy gray. Color gamut is the range of all colors that can be expressed in a given color system. Film, television, and print are incapable of encompassing a shared color gamut and for this reason have their own color systems.
There is clear evidence that the human visual system forms an achromatic channel and two chromatic color-difference planes at the retinal level. The original color TV transmission standard was partly adapted by capitalizing on this visual anomaly. In the three dimensional video color space, luminance is encoded in one channel, while the two color-difference planes reside in the other two channels. This technique allows the two color signals to interleave within the same high frequency luminance region, enabling a composite transmission that is compatible with monochrome home receivers. (continue reading…)
Humans have a non-uniform perceptual response to natural light over the brightness range. We are far more sensitive to small changes in darker tones than we are to similar changes in brighter tones. It’s not perfect, but our non-linear sensation to light can be represented as perceptually uniform when applying a power law (gamma) to light intensity. The primary purpose of gamma correction is for the representation, recording, and storage of imagery. Displaying gamma compensated imagery is of secondary importance.
As the sensor’s input undergoes gamma compression, a straight-line near zero prevents the undue boosting of a sensor’s noise. The curve transitions to a nonlinear transfer function for the range of measurable light, then rolls off to a constant zero slope at the highest light intensities. The power function allocates many more levels to capturing darker tones than does linear encoding. Conversely, there’s an excess availability of levels to capture brighter tones within areas that fall below our perceptual threshold. Gamma encoding ascribes to the principle of constant luminance, where tones are spread more evenly across the entire brightness range.
Unlike our eyes, CCDs (charge coupled devices) are made up of linear sensors that respond uniformly in spectral sensitivity to light intensity. Similarly, as physical models such as diffusion, illumination, depth-cuing, and anti-aliasing are defined linearly, images derived synthetically (CGI) are predominantly rendered floating point in the linear intensity domain. If these images were to be viewed on a display normalized to a gamma value other than unity (no correction), contouring and banding may well be visible. In both cases, their linear intensity domains can only approximate the principle of constant luminance by first applying gamma compression to the image data. As such, most software products map final compositions into a fixed point gamma-corrected domain that—in conjunction with the graphics adapter—matches the inverse of the output device’s transfer function.
Once a gamma compensated image is displayed, one may think gamma expansion in the display negates the effects of gamma compression. Well not exactly. In the case of CRTs, the electron gun’s reaction to voltage input inherently approximates the inverse of gamma compression. CRT gammas average 0.417 (1/2.4), whereas image file gammas may vary considerably from 1.0 for RAW, 1.6 for PC, 1.7 for SGI, 1.8 for MAC, 2.2 for NTSC, 2.6 for D-Cinema, to 2.8 for PAL. So further gamma compensation (expansion or compression) may be needed to derive the ideal unity plot of light intensity on any given display. Curiously, LED and Plasma displays react to their voltage inputs quite differently from CRTs. These displays are mostly calibrated through firmware LUTs (look up tables) to emulate a gamma power function with exponent values of around 1/2.4, slightly boosting contrast of the most prevalent embedded image file gamma of 2.2 (that of Adobe sRGB, Adobe RGB 1998, Rec. 709, Rec. 601).
Indeed, these systems compensate for a lack of dynamic range in the imagery by allowing us to create the desired contrast response through a limited number of bits. It should be noted that increasing the viewing gamma of an image artificially stretches its contrast range by discarding existing step detail to potentially within our threshold of contour visibility. Ironically, contrast ratio is a major determinant of perceived picture quality, so much so that an image with reduced contrast sensitivity—or higher contrast ratio—may be judged sharper than an image that has higher spatial resolution!
Luminance is the term used to refer to the brightness response of human vision. The range of luminance between the lightest and darkest scene element is called contrast ratio. Contrast ratio is a major determinant of perceived picture quality, so much so that an image reproduced with a high contrast ratio may be judged sharper than an image that has higher spatial resolution.
Experiments in human visual perception have shown that people are not very good at identifying intensity differences over a very wide range within a scene. But we are very capable of adjusting our visual response (eye’s sensitivity and dynamic range) to localized brightness and contrast regions where we become sensitive to slight shifts in intensity. The ratio between two adjacent steps of similar luminance can be expressed as contrast sensitivity. Remarkably, human contrast sensitivity adapts over an extremely wide range of viewing conditions.
Perceptual considerations play an extremely important role in image coding. Our attention span typically is concentrated in the middle of a picture’s tonal range. Darker shades and shadows are seen as more compressed, where reproduction of physically correct intensities will appear lacking in contrast. Here wider gradations in shading may go undetected. Highlights, or lighter shades, need tighter gradations as our perception of subtle changes is more acute. TV and film systems are designed so that the scene contrast is expanded upon display.
Linear Color Space
If an 8 bit linear scale were used to encode pixel brightness from black (level 0) to white (level 255), the brightness difference between 16 and 32 will be perceived as the same as the difference between 32 and 64, 64 and 128, 128 and 256, and so on. Each would be seen as twice as bright as the next. But 128 shades of gray appear in the lighter area (128-256), and only 16 shades of gray in the darker area (16-32). This linear approach to codifying our response to light intensity is unacceptable because our eyes expect to see about the same amount of brightness information between each equal ratio through the luminance scale. Furthermore, we perceive twice the light as being only a fraction brighter at the brighter end of the range, where intensities between adjacent steps may be indistinguishable from one another. Conversely, there could be ratios in the darker portions of the scale that fall well within our threshold of contrast sensitivity that’s about 1 percent. As step 50 out of 255 represents 2.5 percent contrast sensitivity, all codes 50 and lower would questionably fall within our smallest perceptual discrimination level, or within our threshold for contour visibility. Even though values below 50 may potentially contain discernible banding, whereas values between 50 and 255 may not, the image will certainly appear flat with overall contrast ratio limited to 5:1. Through 12 bit linear representation, the contrast range will have shifted to 82:1, still inadequate for photographic reproduction, and a lot of the steps available for light intensity representation will be wasted. (continue reading…)
In the natural realm, depth is simply a measure of our visual system’s ability to detect subtle variations in shades. Some scenes can contain a nearly infinite number of shades where our senses fail to quantify all the subtle nuances.
Picture a staircase, the bottom of which is black and the top white with gray shaded steps in between. If a giant were to reach the top in one step, no encounter would be made with the gray shades in between. For us, the same distance may be punctuated by several steps, each accentuated by a different shade. What we may find is that on a hop up we stumbled upon an edge! Where we probably wanted to land was on a step below or above the edge but we were unable to discern that step’s position relative to the subtly shaded adjacent steps.
If the staircase had been designed specifically with our stepping style in mind, the step count may be reduced, for argument’s sake, to 16 differently shaded steps covering the same distance. During the ascent, each incremental step became lighter but only barely noticeable by us. We might have made it to the top tired, but at least unscathed. Had the step count been further reduced, we might have clearly discerned step shade tolerances — a distinct transition from one step to the next.
In the digital realm a byte is a word as are bits characters. Now let us equate the gray steps in the staircase to bits in a byte. (continue reading…)