My last post outlines the equipment I used to perform the world’s highest resolution endoscopic surgery. Here I wanted to share what we had to do the record the footage at full HD resolution to get it to National Geographic for their HD feature “Inside the Living Body”. Archiving HD surgical footage has been a significant shortcoming of many of the existing systems.
First some historical background on archiving endoscopic surgery stills and videos. Surgical endoscopy is archived mainly via still photos. Surgeons keep a few shots of the main pathology seen or a few before and after shots. Those who like to keep videos have mainly used simple consumer VHS (there is a huge medico-legal debate in the field if it is protective or dangerous to record all of your surgeries but that is another discussion). You can imagine the nightmare of storage of a library of tapes that accumulate over the years! For more than 10 years I have used S-VHS to get a slightly higher resolution capture. I was limited since the OR systems best output was S-VHS. For years (sadly more than 10 years) I have been trying to get the surgical equipment companies to put a firewire IEEE 1394 output on the camera units. This would allow me to hook up either a mini-DV deck or digital disk based recorder that did not require analog capture and digitization. They never went this route. I wanted to use miniDV because of its higher resolution, ease of digital editing, and smaller form factor than VHS. Remember back when this started there were not small affordable devices that could capture analog video on the fly and convert to digital as there are now.
Next, the surgical video companies went to digital capture and archiving. I loved the digital still capture that let me capture photos for lectures or the patient’s chart and burn it to a CD-R or DVD-R (actually when this started we were using horrible proprietary strange disk formats now extinct unreadable and sitting in my archives). The captures are ok JPEG or TIFFs. I wish they captured at higher resolution with less compression artifacts. One problem I have seen in some ORs is that the machines are actually set for a low resolution capture which makes no sense at all. They then added direct digital video capture and recording onto the same disks. The quality of these captures is hugely variable between systems. Some capture full motion 30 fps video at a decent resolution whereas others look like blocky jerky garbage. Sadly, I usually have to jump into the capture system’s configuration menu before I start a case in many OR’s. Just today I was working with a state-of-the-art system installed in one of my local hospitals just last week. It was one of the newest HD systems. In fact the manufacturer’s rep was in this morning doing an in-service and configuring the system for optimal use. As usual, once in the configuration menu I found the still captures configured for the lowest resolution 640 x 480 BMP and the video on its lowest MPEG-1 setting. As usual, configured for the lowest quality settings.
New HD Recording System we used: Since we were operating in 1920 x 1080p for my HD project we needed something that could digitally record at this high resolution. We chose the new Sony XDCAM HD Professional Disc Recorder. This systems records 1080/59.94i, which Sony calls 1080 60i.
Because our camera outputs true 1080p, in order to record we first have to scale the image down to 1080i using a Gefen DVI to HD SDI scaler. The resultant 1080i signal is what is input to and recorded with the XDCAM HD deck to a proprietary Sony Professional Disc that is actually based on a Blu-ray disk (sorry won’t play on your PS3 - the disk is in a protective hard shell holder and recorded with a different format than consumer Blu-ray).
The XDCAM HD records HD input in MPEG2 HD, a Sony format. This MPEG HD 1) is designed to yield high-quality video and audio recording and playback. The MPEG HD 1) codec provides video compression compliant with the MPEG-2@HL standard. It enables HD 4:2:0 digital compression recording in the 1080i (1,180 effective scanning lines, interlaced) format currently in use in many broadcast facilities.
There are three selectable video bitrates: HQ at 35 Mbps, SP at 25 Mbps, and LP at 18 Mbps. The resultant “clip” is an MPEG HD file which bears the suffix .MXF.
Another aspect of this recording medium is that it converts the input 16:9, 1920 x 1080 video to a 4:3 1440 x1080, then stretches the output back to 16×9 using non-square pixels (1:1.33). this is similar to the way DV format handles wide screen in Standard Definition.
Next I had to get an exact copy of this disk to National Geographic in HD. As Eric Portlow the video engineer wrote me “We can create a 1:1 copy of the original disc. Because the Deck captures an MPEG HD file to disc and can be accessed as one would an external hard drive using File Access Mode (FAM) via an firewire interface, it is possible to create duplicate discs by importing all content from one disc to the computer, then exporting all files and directories to a blank disc. The result is a lossless copy of the original. If you would like to have an exact duplicate of the original footage in XDCAM HD format, we could do this but it would require your having access to an XDCAM HD deck to view or access the file.
Our other option would have been to use HDCAM SR, DVCAM or DVCPRO
Stay tuned for the next post on HD in the OR hype vs reality




Related Articles
5 users responded in this post
[…] As part of my posts on my research on high definition surgery (and its recording) in the operating room, I posted on the limitations of still photo archiving in the operating room and potential future advances from new compression systems. […]
As a video editor at a medical device company, I have always hated the huge difference in quality and standards of video in the operating room. Some cameras will output 1024×1024 images but then the recorders dumb it down to 640×480, cropping the output a bit. A few months ago I was able to edit an HD surgery video, though it was recorded in 1080i it was really a 4×3 video with sidebars. I am excited to hear that there are finally going to be full 1080p 16×9 systems out there! Too bad most OR’s will not get these for many years to come
Good to know some surgeons grasp the concept that HD recorded to CD-ROM as MPEG-1 or DiVx is, in fact, not HD. We get discs all teh time in any number of odd formats, and agree that the “dumbing down” settings of systems is just plain silly. I have had some conversations with engineers at Olympus and Storz about capturing good HD imagery. It is funny, you go to medical conventions, and the vendors are showing HD surgery playing off XDCAM decks. How many surgeons do you know who will pay $10,000-$20,000 for a video recorder in the OR?
Teeir is only on true 1080p chip and that is a Sony sicon chip in the filming marketplace (movies). All the HD chips in the medicla market do not have a native 1080p resolution on the chips; they are all being upprocessed. period. wake up people its is called marketing.
These references about the first “true” 1920×1080 (full spec HD) chip available are just not true. Ikegami has had a full res 2/3-inch chip available for quite some time for the surgical OR. Arguably, this camera is more light sensitive than the Optronics camera that Storz uses. Also, the Optronics camera’s chip is not true 1080 - it does do some scaling.
As for the Sony HD XDCAM recorder - there are quite a few surgeons/hospitals that have been using this system for quite a long time. Again, this is old news. For quality HD video, $15K is a small price to pay.
Leave A Reply
Please Note: Comment moderation maybe active so there is no need to resubmit your comments