Digigami MPressionist.X

Ten Canoes and 16 Pixels

2007-12-29T08:27:00.001-08:00

I was watching the feature film Ten Canoes on DVD with a friend. It's a fantastic movie, very unusual. It won a prize at the prestigious Cannes Festival. However, the picture quality of the DVD leaves something to be desired. I can't be absolutely sure, but I think it was encoded on a PC.

The problem in the encoding is pretty simple. The motion vector search range used during the encoding was set to a window of 16 pixels. In MegaPEG, you'd get a search window this size if you slide the Motion slider all the way to the top (very low motion).

The problem is, that although this feature does not have a lot of action in it, it does have motion. Camera motion, closeups with peoples heads moving through the frame. That kind of thing.

Here are three screen shots from the MPressionist analysis of the title. The first two show the two pages of motion vectors. Blue and green represent motion vectors that are pointed forward and backward in time. As you can see, no motion vector exceeds the 16 pixel maximum used during the encoding.

The encoder used on this job, handcuffed by the small motion vector search window, had to resort to other techniques to get the video into the required bitrate. Encoder theory predicts that in order to get higher motion passages to encode, more compression will be required. In this case, extreme amounts of compression (indicated by peaks in the third picture).

I have no doubt the encoder used on this job was capable of getting the picture quality right. Although the content is highly textured, they used an 8.6 Mbits/s bitrate ceiling, plenty of bits for something like this. Instead, the difficulties here are the result of "not knowing" what the encoder did. Without that crucial feedback, it's easy to miss the artifacts in this title, which appear only during high motion scenes and do not quite get blocky. Instead, high motion areas end up kind of a blurry mess.

Don't take my word for it. Rent it and watch it.

DVD Encoding PQA

2007-10-12T08:40:00.001-07:00

Compression Chart

First I-frame in bad GOP.
Note red bars (quant-scale) jump in thumbnails.

Overlay showing every macroblock is at max compression.

A friend and I were watching a standard def DVD (The Lucky Seven), and as we were watching my jaw dropped. "Did you see that?", I said, and we hit review on the remote and watched the scene again. I could not believe my eyes, the image had been crushed completely. It looked to me like an encoding problem and so I fired up MPressionist and took a look at the disc.

Sure enough, without even needing to scan the feature to find the timecode of the problematic scene, the compression overview graph showed an out-of-place compression peak about 25% of the way through. I moved the playback to that spot, let it run and the problem became more obvious. The encoder used in this sequence had run into a problem with a time-lapse section. The result was that an entire GOP ran at maximum compression, resulting in the blockiness shown in these images.

It is possible that the encoder they employed could not be adjusted to correct this problem. However, using MPressionist to review the charts for the disc could have saved replicating and distributing this disc, and potential embarrassment with the client.

DIRECTV HD PQA

2007-10-02T16:23:00.000-07:00

Artifacts (Showtime/DIRECTV)
in a slow moving 24p sequence
presumably encoded at 29.97i without inverse telecine.

If you're a DIRECTV viewer, don't beat up the ATSC HDTV channels for picture quality - the current state of picture quality in high-definition television is driven by commercial interests, primarily. Given the spectrum/bandwidth, broadcasters are looking to maximize the number of channels they can send, which means dividing up a 30 mbs pipe into as many chunks as possible. It looks to me like they are doing VBR encoding and crossing their fingers that the statmux won't experience two bitrate peaks at the same time.

Nevertheless, there is one thing that all the HD broadcasters are guilty of, which would make a significant impact in picture quality on most dramatic television. Currently, many HDTV shows which originate on 35mm film are broadcast in 1080i with the pulldown still present in the source material. This footage should be going through an inverse telecine prior to ingest. The reasons for this are both aesthetic and financial. Using inverse telecine at 1080i results in at least a 25-40% reduction in bandwidth. This is a real reduction. Instead of encoding 29.97 interlaced frames per second, only 24 progressive frames are sent (the original frames from film). There are rare instances where cuts done on the Avid will produce a short 2 or 3 frame sequence that is truly interlaced (pulldown sync lost during edit).

The impact on both picture quality and bandwidth would be startling. Personally, I think true interlacing in HDTV should be deprecated. Interlacing is an artifact from an era where we watched material on interlaced devices. It was a compromise based on technological constraints. A large number of HDTV screens now are LCD based. Interlaced content looks terrible on LCD screens. And many of the HDTV screens that are interlaced at 1080i do not implement 1080p correctly. Those latter manufacturers should be penalized by the consumer for failing to support the standard, and 1080p should be king of both content and display. 1080i, given the bandwidth allocation chosen by broadcasters is a complete wash. In low motion scenes, where there is no advantage to interlacing on a true HDTV, 1080p would be better looking than 1080i, and in high-motion scenes, the added complexity of encoding interlacing and an extra 6fps means that the temporal resolution promised by 1080i is in fact, transmuted into a blocky mess by the channel constraints/encoder.

If temporal resolution is what is needed (live football, soccer etc), 720p at 60fps is the way to go.

ATSC HDTV Analysis Workflow

2007-07-30T07:47:00.000-07:00

KTLA Broadasts 2 Digital ATSC Channels in Los Angeles

Export Streams for Analysis from EyeTV

Choose MPEG Elementary Streams as the Format

Change the filename extension to .m2v

EyeTV batches the job and shows you the progress.

If you're interested in analyzing ATSC HDTV picture quality, you can combine our analysis tools with a USB 2.0 ATSC Tuner from Elgato, in combination with their excellent EyeTV software.

EyeTV does some significant treatment of the image as it displays it - which is the right thing for them to do - delivering the best perceived picture quality to the customer. For technicians, however, we would rather see the bitstream images as they were encoded.

EyeTV has Tivo-like PVR functionality, and so you can easily schedule it to record single or recurring programs off the ATSC broadcast, from any number of channels.

For the best workflow with our software, you should have the following:

1. Apple PowerMac G5 Dual or Quad, as fast as possible.
2. Apple Cinema 30" Display, 2560x1600 resolution, with the fastest possible video card, preferably NVidia.
3. Elgato ATSC HDTV tuner box (includes bundled EyeTV software).

Schedule the programs you want to analyze. They will appear in the list of recordings of EyeTV (see screen captures).

Both MPressionist Pro HD and MPressionist HDTV can analyze ATSC HDTV bitstreams. Pro HD has less analysis, so it achieves a faster framerate on playback, but HDTV gives you access to all the variables encoded in the bitstream. Personally, I use them both, depending on what I'm analyzing or focussing on. If you buy our HDTV analysis software, we'll give you one or two copies of Pro HD if it helps you be more productive in your work.

Once you have a program recorded in EyeTV, you need to demux the bitstream. Although EyeTV stores it's transport streams with a .MPG file extension inside a package wrapper, our demuxer has problems with them. So far, these are the only MPEG-1 or MPEG-2 muxes we haven't been able to decode. But that's okay, because EyeTV can demux them for us, and the resulting bitstreams drop right into our analyzer.

Using the mouse, right-click on the recorded program item. Choose Export... and then select MPEG Elementary as the stream format. By default, EyeTV will affix a .mpv file extension, so you should change that now to the more appropriate .m2v.

EyeTV will batch up these export tasks and run them in parallel, and as soon as they are done, you can open them in MPressionist Pro/HDTV and begin analyzing them.

Here's some 1080i examples culled from a couple of problematic sequences. First off, we have Robert X. Cringely's PBS special on ATSC HDTV. Amazingly, this program was broadcast at around 7Mbits/s. Pretty amazing, but there are some artifacts. During his show he rebroadcasts an original television commercial about the introduction of Color TV. The compression artifacts in this segment made me want to grab and analyze it. Here are three screen grabs of the analysis:
the intro - low textural complexity, low motion and very clean picture (PBS is not employing cutscene detection/I-frame injection), Mr. Cringely - espousing the benefits of the flat screen and the Color TV commercial from the 50s with significant artifacts. On the other hand, 1080i at 7 Mbits/s is a wonder in itself.

Here in LA, of the major networks broadcasting in HD, I personally find CBS to be leading the way. In particular, their primetime leader CSI has excellent picture quality and is broadcast in 1080p. It has a very cinematic quality and demonstrates the possibilities for HDTV for everyone. Here are some analysis grabs from SportsCentral. As you can see, this program is encoded at around 12 Mbits/s (1080i). There is a significant amount of noise in the source material, especially the studio work. And the title sequences, with their rapid motion and crisp edges are giving the encoder some problems. Sports programs probably prove the greatest challenge for the broadcaster.

Every Artistic Domain has an Analytic/Quantitative Aspect

2005-12-22T00:57:00.000-08:00

My aunt, who is a sculptress, once told me that you'll never find art through science, but you can find science through art. While I'm not sure I agree now, the basic idea is that to get the best final result from a creative undertaking, often a grasp of the technical side can become a key factor in producing your best work.

And the more experienced and seasoned you become, the more you want to know. Consider all the analysis tools available in Final Cut. And Google Analytics. MPressionist is a kind of mash-up between Google analytics and the graphs in Final Cut.

DVD Verification Services

2005-12-15T08:40:00.000-08:00

The infamous Trai Forrester has launched a new service for pre-mastering disc verification. MPressionist HDTV is one of the tools they have at their disposal for discovering quality problems before replication.

dvdverification.com.

Statistics...

2005-11-06T09:19:00.000-08:00

There are three kinds of lies: lies, damned lies, and statistics.
Benjamin Disraeli
British politician (1804 - 1881) [Source quotationspage.com]

MacFair

2005-11-06T09:05:00.000-08:00

MacFair was a fun time. Unlike most trade shows that have a tendency to grind you down, it had a friendly atmosphere.

I was somewhat surprised by the attendance at my 1hr class "Understanding DVD Picture Quality". Firstly, there were more people than I expected. And secondly, I got a great reminder about how effective MPressionist graphics are at communicating difficult concepts to people who have no background in video compression.

I also used MegaPEG Pro HD to create a test movie that used various filtering and subsampling techniques to reduce the bitrate of a clip from 9.8 Mbits/s to 3 Mbits/s with no apparent change in perceptual quality. I concatenated the clips end to end and let them play through a few times, and also scubbed the timeline to seek randomly to different points in the three clips. Aside from the interlacing artifacts which were only present one of the clips it was nearly impossible to tell the clips apart under the viewing circumstances (1024x768 SVGA projector in darkened room).

Digigami at MacFair - Saturday Nov 5, 2005

2005-11-04T13:09:00.000-08:00

Digigami will be exhibiting tomorrow at MacFair - Southern California's largest Macintosh-only event.

I, myself will also be giving a short presentation (2:30pm) entitled "Understanding DVD Picture Quality" which will be an educational talk aimed at understanding the factors that contribute (or detract from) DVD picture quality. I'll be punctuating my talk with examples from commercial DVDs and self-produced encoder results.

If you're in the greater Los Angeles region, please drop by for a demo of our products or just to say hello.

http://www.macfair.org/

DVD Logo Consortium going after DualDisc?

2005-10-23T14:28:00.000-07:00

It looks as if the DVD Format Logo Licensing corporation is going after some format abusers. I'm wondering if this message targets DualDisc. If you know, post a comment.

0.6mm Thick Optical Discs

Lately, non-bonded discs with the thickness of 0.6mm are being manufactured and distributed to the public. These so called DVD discs are clearly out of Specifications and often cannot be properly played on players and drives in the market. They are reported to cause damages to the discs themselves and/or the hardware during reproduction. Licensees are reminded that the use of the DVD Format for products that do not conform to the Specifications is a violation of the DVD Format/Logo License. Such violation may lead to termination of the License.

Personally, despite the extra thickness in the DualDisc format, I think these format benders are actually very good for the music business, vis a vis mitigating problems with illegal downloading. Having a hard disc full of DRM-protected music and videos may be convenient for purchasing and playing on the iPod, but what if your laptop is stolen or you experience a catastrophic hard disc failure? Nice to have the physical discs (audio and DVD) as backup.

More information about DualDisc:
A Guide to DualDisc
EmediaLive Article

The Importance of DVD Quality Assurance (QA)

2005-10-19T17:10:00.000-07:00

The Refiner's Fire: Quality Assurance Testing for DVD
DVD production is a complex process in which many different media types, Like strands of wool, are woven into an intricate tapestry. Consequently, the Slightest glitch can cause a domino effect that ends up as a major headache.

According to Randy Berg, business development manager for Rainmaker Digital Pictures of Burbank, California, "it can take eight hours to change one video clip. So one little problem equals one day."
[more...]

HDTV buyers get fuzzy deal, Cuban says

2005-10-18T21:40:00.000-07:00

HDTV buyers get fuzzy deal, Cuban says
By Kimberly S. Johnson Denver Post
Beverly Hills, Calif. - Consumers are getting a raw deal when it comes to viewing the best quality of high-definition television possible, said Mark Cuban, co-founder and president of Denver-based HD.Net. Speaking to more than 200 members of the television-manufacturing industry last week during an HDTV conference, Cuban said most HD content is compressed or made smaller, underutilizing HDTV sets capable of showing programming with extremely high resolution.

"It's all really depressed derivatives of what we really could see," he said.

MPressionionist Pro HD will work with both HD-DVD and Blu-Ray

2005-09-27T21:05:00.000-07:00

All Blu-ray players will be required to support playback of MPEG-4 AVC and VC-1, in addition to MPEG-2.

What video codecs will Blu-ray support?

The Blu-ray Disc Association (BDA) is still in the process of finalizing the BD-ROM specification, but they have stated that MPEG-4 AVC High Profile (previously called FRExt) and Microsoft's VC-1 video codec (the proposed SMPTE standard based on WMV9) will be mandatory. They will also include MPEG-2 support for playback of HDTV recordings and DVDs. Please note that this simply means that all Blu-ray players and recorders will have to support playback of these video codecs, it will still be up to the movie studios to decide which video codec(s) they use for their releases. The BDA expects the BD-ROM specification to be finished some time in 2005.

Full text of Article.

MPressionist HDTV Competition, Sort Of...

2005-09-27T19:27:00.000-07:00

Tektronix Introduces MPEG-2 Terrestrial ISDB-T Test Capabilities
6/14/2004 - Tektronix, Inc. (NYSE: TEK), a leading worldwide provider of test, measurement and monitoring instrumentation, announced the addition of MPEG-2 Terrestrial Integrated Services Digital Broadcasting (ISDB-T) test capabilities to its leading MPEG Monitoring and Test Systems. Tektronix' new monitoring and analysis solutions enable Japanese equipment manufacturers and broadcasters to develop, deploy, manage and maintain new ISDB-T equipment used in Digital TV (DTV) infrastructure.

Full Text of Article

The Statistical World of Automagic Compression

2005-08-19T00:05:00.000-07:00

At the end of the Digital Production Buzz interview, I touched briefly upon an analogy comparing 'black-box' video encoders with so-called 'quantitative' stock trading systems used by the large investment banks to manage capital.

Statistics is a form of soft-science which attempts to mathematically describe the phenomenal world, usually by averaging a large number of numerical descriptions/samples to determine trends/patterns which are used, in combination with heuristic algorithms to make decisions about individual events, such as whether to buy/sell or hold a stock.

Statistics, however useful, tell you absolutely nothing about a single sample, or a group of samples. The reason that these heuristically based trading systems are profitable is because the large investment banks can afford to cover short term losses enroute to a larger gain from a predicted trend. It is only because they are embracing a large statistical pool of data that the systems can be profitable.

Thus, an automated trading system is a form of statistical process control, where feedback from the system is filtered statistically, and the results used to control the trading.

This is directly analagous to black-box video compression tools. These tools gather statistical data from one or more passes through the video, and employ this data by running heuristic algorithms to attempt to control the bitrate, predominantly by controlling the amount of compression, or Q (in MPEG parlance), applied to a particular scene, GOP or frame.

Once again, by understanding the nature of statistics, we can anticipate that a statistical black box compression tool, over a wide range of material and many encodes, generally speaking is going to produce good results. However, this does not guarantee that you are going to get good results for a particular encode, scene or group of frames.

Just as a human trader, equipped with both statistical tools, past experience and knowledge of the market can make better-informed decisions about how/when to trade, so then, can we expect that a human compressionist, equipped again with statistical tools, past experience and first-hand visual knowledge of the material (ie. 'knows' particular piece of footage which needs to be compressed) will be able to out-perform a purely automated or statistical system. To return to the analogy of the large investment bank's trading system, what makes that work is huge amounts of capital to risk, deep pockets, and the ability to ride out unwanted losses to reap a later reward.

This rationale is what gets us up in the morning at Digigami, and is the main reason that our encoding tools can produce quality and bitrates competitive with other 'modern' codecs (as mentioned on the show).

Long before I did RoboHelp, which came well before video compression in my life, I worked in the field of industrial process control, specifically for natural gas pipelines (college summer job), and then later for heavy manufacturing where they melt sand into glass and spin it like cotton candy into sheets of home building insulation.

There is no question that statistical methods can be used to augment human insight and experience. We look at it as a way to provide the compressionist/operator with a view of the process which they otherwise could not see. When you cannot see something clearly, you cannot make reasoned, rational decisions about it. Instead, the process becomes mired in myth, legend and religious views.

Interviewed on Digital Production Buzz

2005-08-18T19:18:00.000-07:00

I was interviewed tonight on the radio show 'Digital Production Buzz' regarding the delivery of High-Definition (HD) video content using MPEG-1 VBR produced with Digigami products. The show is podcast as well. We only touched briefly on MPressionist. Digital Production Buzz

Studio DVD Discs and QA with MPressionist Pro 3.0

2005-07-03T19:29:00.000-07:00

From the Apple dvdlist mailing list:

"Take a look at DVD for the independent feature dot com for murder, starring Nastassja Kinski. As a benchmark, compare it to One Night Stand, also starring Ms. Kinski."read the full post on dvdlist@lists.apple.com/2005-06

Superbit DVD Encoding Revealed with MPressionist Pro 3.0

2005-06-19T21:35:00.000-07:00

When I first discovered Superbit variations of Hollywood-encoded feature films, I felt obligated to bring them under objective scrutiny; what improvement, if any, in DVD picture quality did Superbit-encoded titles bring to the table? And what technical differences were there in the encoding parameters used to create the discs?

I thoroughly enjoyed answering the questions that Superbit raised. In fact, analyzing Superbit-encoded DVD discs shaped the feature set for the comparative analysis features of the new MPressionist Pro HD 3.0. For example, I initially designed the per-pixel frame differencing feature of our new product as a way to compare a source movie with a compressed version. This is highly useful by itself. For instance, you can create two compressed versions (with different bitrate and encoding parameters), and compare each of them to the original, drawing from both objective and subjective criteria for deciding which version to use in your final DVD disc. Once I had this basic functionality up and running, my own desire to compare a Superbit movie and it's equivalent widescreen version, motivated me to add MPEG-to-MPEG movie comparison (complementing MPEG-to-QuickTime comparisons). And once I had MPEG-to-MPEG comparisons working, I discovered that there were edit differences in some of the titles. In other words, SpiderMan Widescreen Edition does not contain exactly the same number of frames as the Superbit Edition; in fact, the cropping on the frame is different as well. The desire to do frame-by-frame comparisons of these different 'cuts' motivated me to add the 'sync-point' matching feature of the new product: when comparing two different cuts, one simply needs to find an exact 'match-point' (identical frames at a cut scene), and from that position forward in the movie, the two versions can be compared. The addition of this feature also allows for comparing compressed segments (ie. MPEG movies created from in and out points of a longer source movie) with source footage.

Technically speaking, Superbit titles are encoded at approximately 1Mbs higher average bitrate than the corresponding widescreen edition. As well, the Superbit titles are basically devoid of fancy motion menus and 'DVD extras'. Two audio tracks, both English, carry complete surround sound programs in Dolby Digital and DTS formats. Alternate language tracks are not included.

Under slow-motion or still-frame scrutiny, the compressed picture is indeed of a higher quality. This increased quality is most evident in features shot on a grainy film stock: in Superbit-encoded titles, the grain of the film is much more visible than in the widescreen versions - suggesting that either the Widescreen versions were filtered before encoding, or that the low-pass characteristics of DCT quantization (employed by MPEG compression) was discarding the high-frequency grain structure least visible to the naked eye.

Even more compelling are the frame-difference comparisons for high-motion regions of these films. High-motion footage poses a greater challenge to video compression, and this is evident in comparing Superbit with non-Superbit titles. Typically, in standard widescreen versions, large frame differences can be observed during high-motion (and to a lesser extent, in highly textured) regions relative to the higher bitrate Superbit title.

Of the pairs of discs that I have compared thus far, Underworld (Superbit Collection) shows the least differences with it's equivalent Widescreen Edition. This is primarily due to the extremely 'soft' quality to the cinematography, rather than other factors. I also compared these versions to the Unrated Extended Cut. Unless you're planning to exercise MPressionist on these discs, my suggestion is to avoid the extended cut, which includes new scenes that, frankly, added nothing to the story and included weak performances by the characters, especially the Kraven character.

Analyzing Superbit-encoded titles also provided me with motivation for the extra quality improvements exposed in new features of MegaPEG Pro HD. Specifically, the new mastering quality setting, which uses extremely time-expensive encoding algorithms and makes only a very, very, very small improvement (over Best quality) to the final output. I decided that if someone wanted Superbit quality for their own independent feature, there ought to be an encoder capable of delivering it - even it it meant running the encoder for afew days. For everyone else, MegaPEG's 'Best' quality will deliver a superb picture in afew hours.

Even if you never decide to compare Superbit and Widescreen DVD discs produced by major motion picture studios, I think you'll find the new comparative analysis features of MPressionist Pro 3.0 to be highly useful in your work. I have. The new HD features, for instance, make it possible to compare the quality of HDV cameras, and the output of HDV encoders. If there is one thing that I have learned in my extensive hours of analyzing MPEG video, it is that all encoders are not created equal.

The Rimbaldi Disc: Under the Knife

2005-03-04T01:12:00.000-08:00

Tonight I took an episode from Alias, Season 2, under the knife (MPressionist.X Pro) on a fact gathering mission for something I was researching.

As with many Prime Time TV shows, the principal photography for Alias is done with 35mm film, not with video. From the looks of things, it would appear that the show is edited in 24p, with the 2:3 telecine being applied when printing to the finished 29.97i broadcast master (which I imagine is Digital BETACAM or something similar).

I was intrigued by the breaks in the 2:3 pulldown flags and started studying them. It would appear that the 2:3 pulldown flags in the MPEG-2 video stream are being produced by an MPEG-2 encoder with inverse telecine on the front end which is determining the pulldown cadence by comparing the actual lines of pixels. The breaks in the 2:3 pulldown flag cadence always occur during "cuts through black" (which includes the title sequences that Alias employs when a change in locale has occurred; Paris, Rome, Istanbul, Yawn.) Presumably, upon encountering rows of identical black pixels, the inverse telecine function throws up its hands and reverts to a straight 29.97i cadence.

Which brings me to another byproduct of having breaks in the 2:3 pulldown flags when you publish MPEG-2 on DVD or DTV; the function of progressive display is impaired. When auditioning content on a high-resolution progressive display, a sequence of frames containing fields doesn't just look bad: to the non-professional, it looks bizarre. My guess is that most consumer DVD players which contain "progressive" DVD output are simply delivering progressive frames for coded pictures which include the "progressive_frame" flag in the picture coding extension. I doubt if most players have the circuitry for frame-buffer based de-interlacing algorithms - perhaps some high-end units do - but most would not.

Got coupon? Now you do. mpress-blog-0503 use it in the Digigami Online Store and get 10% off MPressionist.X Pro or the "Master Your Moving Picture Pro Bundle". Thanks for reading.

2:3 Pulldown Cadence Update

2005-02-07T16:47:00.000-08:00

In an earlier post, I described some of my findings with respect to cadence breaks and changes in 2:3 pulldown flag patterns on commercial DVD discs.

I went back to some very early (circa 1992) literature on the development of the MPEG-2 standard, and found some references to the motivations for the 2:3 pulldown features of MPEG. As it turns out, the designers of MPEG-2 anticipated these choppy cadences of 2:3 pulldown material as such sequences are commonly produced by non-linear editing systems when the final product is 29.97i.

So this brings my understanding full circle. The first time I noticed breaks and changes in 2:3 pulldown flags was in evaluating and comparing two different releases of a Hollywood DVD: the original theatrical release vs. the directors cut.

It was clear that the director's cut was reworked on a non-linear editing system, and then released straight to DVD. They probably did the director's cut in 29.97i, using additional 2:3 telecine material (ie. scenes not in the theatrical release). This would account for the fact that 99% of the DVD has a regular 2:3 pulldown cadence and just a very few scenes have true 29.97i material.

Nevertheless, DVD discs which contain mixed pulldown cadences will undoubtedly cause a great deal of trouble for people ripping discs and re-encoding.

More voOM

2005-02-07T16:30:00.000-08:00

In further evaluation of voOM, I have a specific example of artifacts in the video presentation. In this case, it is an HD channel, HBOHD. The scene in question is from the recent HD presentation of the Matrix (original 1999). It's approximately 2/3 of the way through the film, just after Neo and Trininty decimate the foyer of the building where Morpheus is held captive by the 3 agents on the upper floor. They get into the elevator, and then onto the roof of the elevator where Neo sets an explosive device. Then he says, "there is no spoon", the device detonates, and he and Trinity hang onto the elevator cable as the explosion goes down into the shaft. During the explosion (flames etc.) there was a series of frames, perhaps as long as 3/4 of a second, where serious block artifacts occur. I'm assuming the stream is 24p, although I don't have any way to verify that. So we're looking at about 18 (3/4 of 24) frames that have serious artifacts. It detracted from the presentation of the movie.

Since HBOHD is on both voOM and DirecTV, I am wondering if HBO provides a compressed MPEG-2 feed to the DTV broadcasters - so it would seem, as the header for an HD film reads "in HD where available".

voOM

2005-01-24T16:17:00.000-08:00

Recently, I've had a chance to evaluate the voOM HDTV service. While it is true that the HD channels look great, some of the SD channels "feature" serious artifacts, especially in high action sequences. One particular clip, a trailer for a Pierce Brosnan "James Bond" movie, was so bad it was almost unwatchable.

If you're a DVD SciFi fan, I think you'll enjoy a film called Equilibrium, which contemplates the future of martial arts in a novel way.

MPressionist.X Gets Aqua Makeover, JLo spotted in Salon!

2004-12-09T07:58:00.000-08:00

Just kidding about the second part.

The new build of MPressionist.X Pro is now available. HD will follow, no specific timetable, but soon. The build number (2.0.5) is synchronized with MegaPEG.X for the bundle users.

First off, there is a whole new Aqua-friendly look to the product. If it seems like the previous incarnations look like an app with one leg still in the Classic world, well, that's what it was.

In redesigning the UI, special efforts were made to make sure the product was ergonomic for PAL 720x576. Sometime this summer I realized that there was an unintended bias forming toward NTSC, so I put an end to that. If you have a PowerBook17 and are working with PAL video, you will see that there is now just room to fit the whole MPressionist window above the dock.

Also in this version is the ability to decode sequences which include changes in frame size, rate etc, including CIF, Half-D1, Full-D1 and (HD) formats. It is quite interesting to try and visually evaluate the difference between Half-D1 and Full-D1 at a regular viewing distance. I couldn't tell them apart, but I could see a slight difference between CIF 352x240 and 720x480. But it was only because I was looking for it. Under normal watching conditions I doubt anyone would care.

If there are wild changes in the frame rate at sequence_header breaks, the bitrate statistics are going to be off. The graph will still approximate the trend, but the actual numbers will be wrong. I mean, I looked at those feature films with 2:3 pulldown breaks for 2 years not realizing they had timing 'issues'.

Another new feature is the ability to take .m2v files which are only a portion of a VOB (say one segment) and open it and analyze it. If you need to RIP a full DVD, follow the instructions in the user's guide and you will get one nice long 4GB .m2v file to work with.

DVB Logo Certification

2004-11-29T14:41:00.000-08:00

Sometimes introductory pricing is kind of a marketing concept, but in the case of our products, we're quite serious. Generally, we're rewarding both early adopters and customers who have purchased Digigami products in the past.

In any event, this past weekend we received the official stamp of approval to use the DVB logo with MPressionist HD. For those of you who've never heard of DVB, it is roughly the European equivalent of the ATSC HDTV group. DVB is a satellite based digital TV system which is based on the ISO/IEC 13818 (aka MPEG-2) digital video format. Today we raised the price of HD to $1280.00.

You can be sure that the price will rise again, as MPressionist HD takes its rightful place among the big iron diagnostic tools from Tektronix and others. We figure that the average DTV broadcast engineer would much rather sit behind a gleaming, silver G5 supercomputer with a pair of Apple Cinema displays than anywhere else.

We're also planning a competitive upgrade offer for those who splurged on the ill-fated, but well-engineered MProbe product. If you're an MProbe user and find yourself chomping at the bit, contact us and we'll work something out.

The Value of Objective Measurement

2004-11-25T04:34:00.000-08:00

As far as my job goes, which includes the responsibility for improving an MPEG encoder, MPressionist plays an additional role that I want to mention briefly.

Writing new code to improve an MPEG encoder is challenging, but not that challenging. It's not as difficult, as say, sending 4 live human beings to the moon in a rocket-powered aluminum tin-can and getting them back to earth safely.

What is difficult or sometimes impossible, is changing a person's belief system. And that includes mine. So one of the main roles that MPressionist plays in my life is forcing me to come to terms with my beliefs about how well our MPEG encoder performs. It also helps me to determine what is feasible; because if we don't know what is feasible, we can't deliver on promises made to customers.

I think you'll find some places where 29.97i NTSC is promised at 4.5 Mbs with Hollywood DVD quality. Even with heavy filtering, I think this barely feasible; at least with the algorithms I've seen running in the MPEG encoders I play with. And because I don't think its feasible, we don't make that promise in MegaPEG.X; instead we provide a gamut of alternatives, which includes spatial and temporal downsampling, so that you can get the picture quality out that you want.

I never would have arrived at these conclusions without MPressionist. I figured that if a rock-video shot on 35mm film with Steadicam and cranes at 30fps (on what is probably a 5M budget) is encoded at 7-9 Mbs 29.97i, with copious amounts of filtering, then it is doubtful that a camcorder sequence shot on a handheld/tripod mounted Sony 3CCD prosumer camera and edited in FCP is going to be able to EXCEED that. I mean you can go on and believe whatever you want, but my conclusions are based on working backwards from a finished product that looks superb.

If you wanted to sum up our philosophy here at Digigami, it would be: no artifacts.

That does not mean that our encoder does not produce artifacts. It means that we give you a complete toolset to get your project into the required bitbudget WITHOUT artifacts. And you get to choose what the tradeoffs are.

Every time there's a piece of footage that makes our encoder barf, minutes later I come face to face with it in MPressionist and am forced to re-examine my beliefs about our techniques. In the long run, the more times we fail in this regard the better our encoder will be.