Twixtor is a plug-in from ReVision Effects that allows you to change the frame rate of your footage thru motion compensated frame blending.  While it can be used to add frames for slow motion effects, I have found that I get much better results when removing frames, such as when dropping from 30fps to 24fps.  Regardless of the specific settings, Twixtor takes a lot of time to render.  In our first tests on 8-Core Xeon systems, processing one minute of source footage required one minute of render time.  Now with Intel’s new Nehalem based CPUs, and recently their even newer Gulftown 6-core chips, we have seen that reduced by about fifty percent, to a half hour per minute of source footage, which is still a long time, but feels great compared to where we were a year ago.  Since our footage re-link process is based on frame counts, we have to process our entire source clips in order to take advantage of that level of workflow automation, even if we are only using the last ten seconds of a 14 minute take.  Obviously there are ways around this, but we currently have more render time available to us than man hours, and it gives us more flexibility later on anyway, so we just let it go.  We took advantage of every night and weekend during creative editorial to Twixtor every clip that made it into the rough cut, and now we just have to link to that bank of processed footage to conform our cuts in CS4.  The fact that all of Canons DSLRs now support 24p should alleviate most of the frame rate and Twixtor issues in future projects.

Besides frame rate issues, Canon DSLRs present another unique challenge, in regards to color space and bit depth.  Many professional video codecs store the color values in the range between 16 and 235, of the 256 possible 8bit options.  (The reasoning for this is fairly complicated, and relates primarily to legacy analog video signal issues)  This limits pixels to 220 levels for each color in most 8bit codecs, but the MOV files from the Canon DSLRs use the entire 256 possible options (0-255) for each color.  This increases the number of possible values for each three color pixel by over 50%, (220^3 vs 256^3) but also means that converting your DSLR footage into most other 8bit formats will result in one of two issues: either the extreme values will be clipped, losing detail in the highlights and the shadows, or all of the dynamic range will be squeezed into the reduced sample-space, meaning certain intermediate values are going to be merged together if you edit in an 8bit codec.

Clipping was the most likely possibility in most existing applications prior to the release of Quicktime 7.6.2 in mid 2009.  Previous to that point, Quicktime displayed Canon clips incorrectly (clipping the values beyond 16-235) but after that update was released, most applications that used Quicktime to decode DSLR footage, were able to access the entire dynamic range of the source clips.  This support is not a foregone conclusion though, since DSLR files could be imported with a more generic MPEG4 decoder without Quicktime, and still be displayed incorrectly.  Even with properly calibrated import processes, compressing the 255 possible values for each color channel into the limited 220 values that most 8bit video formats offer, will lead to a loss of precision, and a potential increase in color banding, especially if you plan to color correct the footage later.  A 10bit video format will offer four times as many possible legal color values, and will be able to store all of the original image data with precision to spare.  Once you have color corrected your footage, and any visual effects are complete, an 8bit distribution format may be sufficient for most uses, but any image processing that takes place on the original files before you apply the “look” that you want, should definitely be processed in at least 10bit color space to preserve as much of your original image information  as possible.

When editing DSLR footage in Avid, DNxHD is the recommended intermediate format.  DNxHD files can be encoded in either HD 709 (16-235) or RGB (0-255) color space, but any DNxHD files encoded in RGB are converted to HD 709 upon import into Media Composer, regardless of the original output setting.  Therefore any DNxHD MOV files generated elsewhere for ingest into Avid should be exported at 16-235 to match Avid’s target color space, for a lossless “Fast Import.”  On the other hand, when importing DSLR footage into Avid, you should select “Computer RGB (0-255)” as the SOURCE color space, in the “File Pixel to Video Mapping” options.  (Rec. 709 is always the TARGET color space for DNxHD MXFs in Avid)  While importing with the 0-255 setting retains the full dynamic range, it still squeezes the entire range into the 16-235 gamut.  That loss of precision should not be as significant as viewable dynamic range for an offline edit, but if you planning to export your Avid sequence as your master without a separate conform, you should consider using a 10bit codec in Avid, like DNxHD 175x.  That will allow you to maintain both the original dynamic range and the bit depth, at the expense of higher storage space requirements.

Once you have a re-linked timeline of high quality 24p footage, there are still a few more steps that can be taken to cleanup the footage.  Dead pixels should be the first thing on the list to deal with.  Dead pixels can be caused by physical debris on the sensor or lens of the camera, or by an electronic malfunction with one of the photo-receptors on the CMOS sensor.  The result is the same regardless of the cause, with one of more pixels locked at a static value throughout the shot.  The simplest way to fix this is to cover the effected pixels with information from the surrounding area.  One procedural way to fix this is to duplicate the layer of footage in an AE comp, and mask out a similar section nearby and cover it.  (If you have a horizontal row of three dead pixels, mask the three pixels above them on a second layer, and then drop the top layer down one to cover the spot)  In most cases the duplicated data will be totally invisible, but be sure to QC the result.  If you are Twixtoring your footage to a different frame rate, fix the dead pixels before applying the rate change, otherwise the motion compensation process will cause the dead pixels to move around, making them much more difficult to remove in a procedural fashion.  The next step is to look for any rolling shutter artifacts, caused by the slight difference in time between when the top and bottom of the frame are sampled.  This difference in time can manifest itself in a number of interesting ways, including distortion, with the top of the frame seeming to “lead” the bottom.  It can also cause horizontal bands of brightness with quick flashes of light only being recognized by part of the sensor.  The Foundry has a plug-in called Rolling Shutter that can help reduce the image distortion caused by motion of the camera on smoother shots.  The horizontal bands have to be removed manually in a VFX process if you want to get rid of them, borrowing data from the preceding or following frames if needed.  The Canon DSLRs also exhibit some moiré and other aliasing issues due to the way they sample the low resolution video from the high resolution sensor.  The only way to really get rid of those artifacts is to selectively mask and blur the effected sections of the frame.  Lastly, if you are using Twixtor, QC the output for corrupt frames caused by the interpretation engine being unable to guess the proper motion of the moment in the shot.  If re-rendering with different settings doesn’t help, covering the bad frame with an original frame of footage from that moment usually solves the single frame issues.  Luckily the most difficult sections of footage to calculate motion compensate for, are usually segments of where using frame dropping conversion instead is undetectable, since the extreme motion should hide any stutter caused by the missing frames. (This is coming from a guy who is processing a lot of handheld combat footage)  Once these steps, as well as the rest of your visual effects work, are finished, you are ready to export and color, which should be similar to most other workflows at this point.

Dual 3.0 Ghz 5365 Clovertown Intel Xeon Processors (8-Cores)
8GB Ram (Once I get a 64bit OS)
QuadroFX 4600 Graphics Card (To power my 30″ LCD)
2x300GB 15k RPM SAS Drives (125MB/s each)

The XW8400 is currently one step behind the new top of the line XW8600, but the PCI-X slots, wider 3rd party support, and much lower price, weighted in its favor.  I will be doing some tests and benchmarks in the near future, so stay tuned for the results.  I am also curious to see if the Quadro4600 properly allows fullscreen overlay output in Prospect2K’s CineformRT Premiere mode, fixing the Geforce8 overlay problem.

I got a ridiculously good deal on everything, since I waited nearly a year for the perfect opportunity, but I now have a few extra parts that I am not sure how to best utilize:

I already had all the disks I needed, with a 74GB 10k rpm SATA drive for OS and 4x 500GB SATA RAID5 for Data. My Raid5 is only getting 60-80MB/s with the integrated Intel Matrix storage controller, which is less than I was hoping for, but my data will be secure. If I continue to use the Raid5, I won’t have any place for my new, SAS drives worth $1000 bucks apiece. Alternately I could replace my Raid5 with a single 1TB disk for quantity storage with less security, and use the SAS disks in Raid0 for high-speed capture storage.

So that leaves me with a problem/question that I am looking for creative solutions to:
I need a good external enclosure for either 2 SAS drives, or 4 SATA drives. With SAS, the key factor I would be looking for is speed, and with SATA, the key factor would be security. Drobo comes to mind for the SATA drives, but is there a cheaper option with the same level of security, or better bandwidth?  Anyone know any good ways of externally connecting two SAS drives in Raid0?

Anyhow, export and render benchmark results to come, specifically with Premiere, AE, RedCine, and Cineform, plus a few FPS benchmarks for fun.

Surely the biggest issue at the moment related to consumer electronics would be the Blu-Ray vs. HD-DVD competition.  It looks like Blu-Ray has a majority of the market, and recently more companies have been dropping HD-DVD in favor of Blu-Ray.  As a consumer who doesn’t yet own an HDTV, and is not interested in paying $30 or more per movie, it doesn’t really effect me as much as it could, but I have been rooting for Blu-Ray from the side lines.  I have a Blu-Ray burner, and Adobe Encore CS3 supports Blu-Ray authoring, and I also just think the technology is better.  50GB on a dual layer disc, more capacity in the future, it is just superior all the way around, higher priced or not.

 A number of manufacturers are offering 120Hz LCD screens.  This technology is advantagious in a number of ways.  24p content will look smoother at 120Hz, as each frame is flashed 5 times, instead of alternating 2 and 3 refreshes on current 60hz screens.  Certain 3D TVs combine 120Hz with internal polarization to provide multiple image streams that can be separated with passive filters.  My primary application for this would be stereoscopic 3D video display, but there is also talk of future gaming consoles offering each player a dedicated fullscreen view.  Instead of 3D glasses, you would have Player 1 glasses and Player 2 glasses, each filtering out the other’s view.

Westinghouse is showing off their 8MegaPixel 4x1080P display, which I see being useful to all those new Red enabled 4K film-makers.  3840×2160 isn’t technically 4K, but it is close, and for $15K, what a bargain.  No indication of how they intend users to input to it, most likely through two channels of dual-link DVI.

Dell is developing a 16×5 wide monitor under the Alienware brand, targeted at gamers, but it might make for an interesting NLE workspace at 2880 pixels wide.  On a related note, Dell updated their 30″ LCD to the 3008WFP, which added analog inputs, HDMI, and the new DisplayPort, which is currently useless.

Intel’s new Skulltrail platform is what I call clever marketing of an existing product.  It is a dual socket design, identical to the professional Xeon workstation platform.  The only real addition for the consumer is overclocking capability.  I will take my 5000 series Xeon workstation instead thank you.

On that note, Apple released the details for its long awaited refresh of the MacPro lineup.  No surprises there at all.  8 cores at 2.8-3.2 Ghz, 1600MHz FSB, video options are ATI HD2600XT or faster Nvidia 8800GT, in a reversal from last release where ATI was the faster upgrade option.  Same case, 4 drive bays, now with SAS options, 4TB SATA max.  Nothing there was unexpected.

Intel also released more info on its upcoming consumer level 45nm CPUs, both desktop and mobile, but nothing much of interest in the specific details.  Maximum clockspeed still hovers around 3GHz.  Blame AMD for no serious competition in that regard.

The biggest change in the new generation, is that these chips are based on Intel’s new 45nm process.  This will allow them to run much more efficiently, requiring less power, and generating less heat.  There are also some marginal increases in maximum clock speed, but it is rumored that the new process has much more headroom for future speed increases.  The primary reason I can imagine why Intel might be waiting to push the speeds higher, is that the competition from AMD doesn’t require them to increase speeds to stay ahead, and they are holding that capability in reserve, in the event that AMD does release something powerful, so that they can respond quickly with a faster product.  Regardless of how far Intel is currently pushing the envelope, they currently provide the most powerful processors for most applications and uses.  The new release bumps the top models from the X5365′s 3.0Ghz to the new X5482′s 3.2Ghz (or 3.4Ghz for a Dual Core X5272).  The highest end chips also have a FSB increase from 1333Mhz to 1600Mhz (333Mhz to 400Mhz if you don’t count Intel’s Quad Buffering marketing terms).

The new line of CPUs are cheaper than the older generation at equivalent speeds, which is always an advantage for end user.  Having new products on the market should also bring down the prices on the older generation of CPUs, which will hopefully lead to some bargains in the near future.  The new processors are pin compatible with the previous three generations of Xeons, all using the LGA771 socket, so it should be possible to upgrade existing Xeon workstations and servers with the new chips, but the FSB will be limited to 1333Mhz.  Any 5000 series Xeon system or above should be compatible with some of the new chips.  There are also some new single socket CPUs that use the more common LGA775 socket.  Many more LGA775 compatible 45nm varients will be released early next year under the “Core2″ label.As of today, it is very difficult to find any way to actually purchase these new units, as both HP and Dell have not begun to include them as options on their current systems.  HP has announced that it will soon be replacing their XW8400 series workstations with the XW8600 series, which will bring all of the advantages of the new CPUs and chipsets to their line.  I expect a similar announcement from Dell, replacing their current Precision 490 and 690 lines in the near future.  The new XW8600 will be based on the 5400 chipset, and will offer 2 PCIe x16 slots for dual graphics, as well as an x8 and dual x4 PCIe slots, and an older PCI-X slot for compatibility with older components.  It currently lists a maximum FSB of 1333Mhz, but I expect that to be upped to 1600Mhz as the CPUs that support that speed become available.  RAM will still be based on fully buffered DDR2 technology at 667Mhz, with up to 128GBin 16 DIMM slots, but still limited to 4GB for 32bit work.

Currently, information on HP’s site implies that the new systems will not be compatible with XP, and only with Vista, but I cannot confirm that.  I have seen it implied elsewhere that Intel’s new chipset will have Windows XP support.   Windows XP still offers many advantages, especially when viewed in light of current post-production software options, so that is an important question that remains unanswered.

As newer post-production software is developed to squeeze every last bit of available performance out of available hardware, the GPU is becoming a more important factor in building a high performance workstation.  A few pieces of software that I use that depend on the GPU are: Matrox’s AXIO-LE, Red Giant’s Magic Bullet (especially Colorista), and Iridas’ SpeedGradeHD.  Each has a list of supported cards, and hopefully there will be some intersection in those subsets, or these software applications will be incompatible with each other.  For any given product, there are usually a variety of options, sometime ranging in price from $50-$2500.  Determining which of these options best suits your needs is an important decision, and sometimes the best choice is not immediately apparent.

The competition between NVidia and ATI used to be much stronger, but recently, NVidia has pulled ahead significantly.  I am not sure if this is related to ATI’s abrupt acquisition by AMD last year, or anything else, but NVidia’s development has been consistently resulting in products that are much more capable than ATI’s.  In the professional arena, ATI doesn’t even offer features like SDI outputs and Genlock, to compete with NVidia’s offerings.  These specific features are very relevant to the utilization of these cards in the post-production workflow.  SLI is another NVidia development that ATI has no answer for in their professional line, but implementations of that technology are more tailored to 3D animation and scientific applications.  Stereoscopic output has been offered by NVidia’s QuadroFX line for many years, although their solution is a bit outdated at this point.

ATI has few advantages to counter with.  The most significant one I am aware of, for post-production, is that the ATI architecture is better optimized for returning processed images to the system bus.  Certain applications are able to pass more data to and from ATI cards than their Nvidia counterparts, which is beneficial if you plan to do more than preview the results on screen.  This is why Matrox’s AXIO-LE gets better performance when paired with ATI cards than much more powerful NVidia solutions.  Another issue I have seen with Cineform’s RT engine in Premiere is a color shift between between the video overlay and still frames.  According to David Newman at Cineform, this is due to an inconsistent implementation of YUV overlay on NVidia cards (See his comment on the ProspectHD post) and ATI cards, to their credit, do not suffer from this problem.  There are very few other features in ATI’s favor that I am aware of, but I am always open to being enlightened in that regard if I am overlooking something significant.  Given the current state of things, my recommended choice in most cases, would be to go with an NVidia based card.

Choosing between NVidia and ATI solutions is not the only significant step in the selection process.  Frequently, the most confusing aspect of choosing a new display card, is motivated by chipmakers’ desire to make higher profits from business customers, in that “professional” 3D graphics cards are much higher priced, than seemingly identical consumer gaming 3D graphics cards.  The actual specific differences are rather vague in many cases, and will depend on the requirements of your application.  Certain features such as SDI output and Genlock are clearly exclusive to professional hardware, and product support is much better for the professional lines, but when it comes to GPU processing, the differences are not so obvious.  This is especially true since both companies utilize a unified driver architecture, allowing the same drivers to support almost any of their cards.  Both companies throw around the term OpenGL in regards to their professional cards, but most of the same features are available from the consumer cards.  I have used OpenGL acceleration in After Effects, and have found no real differences, but I am not a professional animator, so higher end 3D animation and modeling programs might see certain advantages.

ATI has their FireGL line of professional cards to compare to their Radeon series.  I have used very few of these cards, so I can offer little in the way of advice.  They are rarely recommended or required by post-production software solutions.  My primary experience with the Radeon line has been in conjunction with the Matrox AXIO-LE, and I have not been impressed with the stability or features of the cards.  The most important feature that I find totally unsupported is the hardware spanning of two displays.  I also have occasional vertical sync issues when running LCDs at 1920×1200, but all this is based on my experience with two X1900 series cards.  I have much more experience, and a greater level of success with NVidia cards.

Nvidia’s QuadroFX line of professional graphics cards is VERY similar to their GeForce line of cards, and with even greater price differences.  In my experience, most software runs equally well if not better on GeForce cards compared to their QuadroFX relatives.  I own a QuadroFX3400 which is almost exactly identical to the GeForce6800GTX, and was four times the MSRP when I bought it.  Although the card has served me well, I have found no compelling reason to have required it over the similar GeForce option.  There is a rumor that Nvidia disabled certain functions when they released their newest generation of consumer cards, that will now only be available from the QuadroFX line, but I have not been able to confirm that.  Specifically they are said to have disabled hardware support for full screen video overlay, (allowing full screen preview in an NLE) which I intend to test once I get a working GeForce8 card.  I would appreciate information about anyone else’s experiences in this regard.  If that is true, it means that we might soon find signicant disadvantages from using consumer cards for professional work, but fortunately, I do not think we have yet come to that point.

What all that boils down to is, currently Nvidia is the performance leader, and unless you have a compelling reason to shell out the money for a QuadroFX model, a GeForce card should be suitable for most applications.  That said,the new 8800GT is a remarkable value for almost anyone who needs a powerful GPU. (Please note I am NOT speaking of the much lower end 8600GT card)  As an added benefit, the new 8800GT should run cooler and quieter than any other card with similar performance.  I also expect that the new PCIe 2.0 compatibility should be able to be taken advantage of with upcoming release of the next generation of Intel Xeon workstation platform early next month.  If I hadn’t been in the process of acquiring the similar 8800GTX, which is at least twice the size, price, heat, power, and noise, for similar resulting performance, I would have already ordered a GT by now, and still might do so regardless.

Technology has come a long way in the last two years, especially in the CPU processing aspect of the equation.  A well equipt laptop can be purchased now that has more processing power than the highest-end Windows based workstations of two years ago, thanks to the Core2 Duo.  I bought a Xeon workstation in 2005, and one year later, bought a 12″ notebook for LESS money, that has MORE CPU power.  With the upcoming release of quad core mobile CPUs, we can remove processing power from the list of limitations that mobility imposes.

Next is RAM, and we are in a unique situation in that regard.  Most systems still use 32bit OSes, and are limited to 4GB of RAM.  This software limitation has allowed notebooks to catchup with desktops in this regard, as demand has not climbed as much past 4GB in the desktop sector, and notebooks were under no similar limit until they caught up.  4GB of notebook RAM can be had for under $200.  Obviously mobile solutions will not be limited by the maximum available RAM. (Any more than a desktop)

The first area where we encounter trouble is with storage, in both capacity and transfer rate.  While there are solutions that allow uncompressed HD speeds and capacities on a laptop (ExpressCard to CalDigit HDPro), that solution is not very portable.  The capacity issue can be solved via 1TB drives connected via firewire, or internal RAIDs or 2.5inch hard disks in large laptops, but there is currently no way to provide the required transfer rates for realtime uncompressed high-resolution editing, in a mobile solution.  This leads to a need for a different solution.  We can utilize the extra CPU now available by using a compressed video format to decrease the strorage requirements.  (As an aside, my FREE IDEA of the day is: A mobile array of 8x250GB 2.5 inch disks with an external PCIe interface for use with x1 PCIe or ExpressCards would offer 2TB at 250MB/s.  It would be an interesting solution, and if a company decided to create it, all the technology already exists.)

 Since there are no reasonable storage solutions for uncompressed, we must examine compression options.  The first standardized option is HDV.  The advantages are low-bitrate, wide support, and firewire I/O, which most laptops already have available.  The disadvantages are lower quality, 8bit 4:2:0 MPEG encoding, and a 1440 horizontal resolution limit.  The next option is a very recent one for Premiere Pro users, with DVCPro-HD, made possible by the release of the 3.1.0 update this week.  With datarates of 5-12Mb/s, this format is will within the 30MB/s capabilities of a single 2.5inch laptop drive.  The horizontal resolution is even more limited, to 1280 pixels wide, but it encodes 8bit 4:2:2 with DCT compression, and in general a higher bitrate should improve quality.  Since it doesn’t use MPEG compression, it should be less CPU intensive to playback and edit, leading to better performance.  The next step would probably be one of Cineform’s products.  AspectHD is limited to 1440 at 8bit, but it still the highest quality solution yet, usually at around 9-10MB/s, using Wavelet compression, which allows efficient low resolution playback as an added bonus. 

 Blackmagic Design released a MotionJPEG codec a while back that allows full 1920×1080 files at 8bit 4:2:2 to be used with a data rate of around 12MB/s.  The advantages are that it can be used for free, with out any limitations that I am aware of, but it is designed to be used with their I/O hardware for acquisition and preview.  The only disadvantage is not really a disadvantage comparedto the options below it, but image quality will not be as good as Cineform, and is limited to 8-bit.  Cineform’s higher-end product ProspectHD has few limitations, allowing 10bit 422 at full1920x1080 to be edited at around 15MB/s depending on settings.  This is easily sustainable on internal laptop disks, while an external firewire drive could increase performance and capacity.  Currently Cineform is the solution I would recommend if you need high end HD editing in a portable form factor.  The next question that Cinefrom prompts is, how high can I go, and currently 2K 444 RGB is possible at 30-40MB/s, meaning an internal RAID 0 would benefit playback on a laptop.  There is even talk of 4K realtime playback with the release of the RedOne, so it seems that the sky is the limit.

Other solutions that I am aware of, but don’t seem ideal are: Matrox MPEG I-Frame files from AXIO or RT.X2 at 12MB/s with the Matrox M.key in desktop mode, which I use, but the performance is not good for creative work without the hardware acceleration.  In the non-Adobe world, FCP offers DVCProHD, and now recently ProRES, a full frame 10bit 422 codec that runs about 15MB/s at 1080p I believe.  I have no familiarity with Avid, but XpressPro or Media Composer with DNxHD might be a portable HD option as well.

Laptops have done a lot of catching up recently, and the concept of ”desktop-replacement” is a much more legitimate now than it was two years ago.  The most important aspect we have not yet examined is HD I/O for portable solutions, especially for portable acquisition, which I plan to go over next time.

Then in the summer of 2006, Intel released two major processor upgrades back to back.  The Dempsey cores (5000 series) were a maxed out variation of the Pentium 4 “Netburst” architecture, and finally brought dual cores to the Xeon line.  The 3.73 Ghz was well above AMD’s 2.6 Ghz and the 1066Mhz FSB final topped AMD’s 1Ghz.  Only one month later, Intel released its entire new line of CPUs for all platforms, based on their totally new “Core2″ design.  The Woodcrest series of Xeon’s (5100 series) were clocked lower, were supposed to be much more efficient per clock cycle, along the lines of the Opterons.  Woodcrest had everything to finally close the gap between Xeons and Opterons, with dual 64 bit cores runnning more efficiently and already at higher clock speeds, with a 3Ghz model available.  AMD had very little in the way of improvements in their response, and were totally unprepared when Intel released their next update less than 6 months later.

The Clovertown (5300 series) CPUs were simply two Woodcrest chips in a single socket, making it a Quad Core CPU.  This allowed a regular Xeon motherboard to support 8 discrete processing cores, clearly doubling performance in high end applications.  I had the privilege to use a Clovertown system for about a month when they were first released, and it was without question the fastest computer I have ever used, by a long shot.

AMD’s response was a new line of CPUs with a new numbering scheme, but no new major features.  Then recently, a year after Intel brought Quad Core CPUs to market, AMD released their long awaited Barcelona line, which were native quad core CPUs.  I have yet to see any version of those for sale nearly a month after release, and almost every review and benchmark has been negative.

We are now a month away from Intel’s next refresh of their CPU line, and are looking forward to more L2 cache, 1600Mhz FSB, and much lower prices.  AMD seems to have nothing in sight with which to compete with, which is unfortunate for both Intel and AMD users, since competition usually drives prices down for all users.  On the positive side, Intel doesn’t seem to be using their monopoly on the ultra high end to dramatically inflate prices. 

Xeons are still lacking AMD’s integrated memory controller and Hypertransport link, but those are scheduled to be included in Intel’s next major redesign, “Nehalem” in late 2008.  It will be interesting to see what AMD brings to the table by then.  Stay tuned for details when Intel releases their new line of CPUs next month.