LensWork Technology Blog

VLC Player

On the advice and recommendation of a reader's comments to my post the other day about Windows Media Player 12, I downloaded the open source VLC Player. Very nice. Very universal. I threw every type of file I could think at it and it played all of them flawlessly and smoothly. Video and audio. If there are limitations to this player, I didn't find them.

   

   

So, it can be done. Which begs the question, why don't Apple and Microsoft do it? Why are QuickTime and Windows Media Player (other than version 12) compromised? I do tire of these inane turf wars — particularly as a publisher trying to make something that everyone can access. I guess the next question is how do I get the entire world to install the VLC player? Nice thought, but it ain't gonna happen.

Windows Media Player 12 plays MOV files!

I must have missed this one — this really big one. Or Microsoft did an awful job of spreading the word.

I discovered by accident that my new Windows Media Player (version 12, available only with the Windows 7 operating system) will now play MOV videos. No more necessity to have QuickTime just to play MOV format files. Who knew?

And there's a bit of icing on the cake beyond that. Windows Media Player version 12 will also now play the MTS video format files that come out of my Panasonic cameras. No more converting them from MTS to WMV just to see what's there. Again, who knew?

Way to go Microsoft. A video player that plays videos. I like it!

Unfortunately, the turf wars rage on — iPad won't play Flash; WMV files on a Mac are problematic; the masses who still have Windows XP need QuickTime to play MOV files. It's a mess and I sure don't see it getting any less messy in the near future. Imagine if we didn't have JPG and PNG and had to have different still image file formats for Mac, PC, and iPad.

A bit of video learning curve

I've been on the hunt for another camera that will do video (my Panasonic G1 cameras do not). I was intrigued by the new Sony A77 and its APS-C sensor size. I assumed its noise factor would be an improvement even compared with a Panasonic GH2, the leading candidate in my quest for a new video-capable camera. In fact, the new images posted at dpreview.com this week show the GH2 is less noisy than the Sony at ISO 1600, in spite of Sony's APS-C size sensor. What a shock. Obviously, the cost of packing in so many sensor points so densely.

I lost interest in the Sony. The GH2 still leads the field, but it is really more than I need. Great camera, but I'd rather have two or three (for multi-camera shoots) that are enough yet affordable rather than one that is great but overkill.

So, I went all the way back to the beginning and started over looking for a camera that might meet my video needs. I found it. The now almost-discontinued Panasonic G2 is the perfect fit for me. Uses my existing lenses; does 720p video (which is all I really need, see below); has the same size sensor as my G1 (12 megapixel stills); has almost the same noise characteristics at ISP 800 as does the GH2; has a stereo input jack for external microphones; and is at the end of its product cycle life so is inexpensive. I purchased a body-only for $299 at Amazon.com which will arrive here on Tuesday, just in time for my departure next Thursday to Carmel for the Uelsmann lecture and for my video tour of Chuck Davis' darkroom.

Here are some RAW noise comparisons from dpreview.com, however, in the real world I'd apply a little Noiseware Pro noise reduction.

   

Below is after Noiseware Pro noise reduction. I can live with this. No, I can thrive with this.

   

   

At ISO 1600, you can see that the G2 is a lot noisier than the GH2, but I rarely use it that high. What really surprises me is that it's about on par with the Sony A77! Of course, the Sony is a 24 megapixel image which will be a deciding factor for lots of photographers, but isn't important to me. I guess that's the cost of the extra pixels, the inevitable noise.

   

What is most impressive is how great the GH2 performs at ISP 1600. Wow! And a lot cheaper than a $2,500 Canon 5D MkII. Again, I assume, the advantage of the 5D MkII being is larger pixel count.

Relative to video, here was the learning curve from the week: If you are a Full Access member of LensWork Online, check out the interview I did with Doug Ethridge. The three-camera shoot of our interview was done with a somewhat dangerous mix of cameras. The video of Doug was shot with my Panasonic GF1; the video of me was shot with his Panasonic GH2; the wide shot video of both of us was shot with my snapshot camera, the Panasonic ZS3. All were AVCHD captured MTS files. Here are the details of the capture:

• GH2 (me) was shot at 1920x1080 at 16,589 kbps (1080p, full frame HD, AVCHD)
• GF1 (Doug) was shot at 1280x720 at 15,285 kbps (720p, AVCHDLite)
• ZS3 (wide of both of us) was shot at 1280x720 at 15,260 kbps (720p, AVCHDLite)

   

Although the GH2 has a much larger image and higher data rate that I'm sure would be an advantage for lots of full HD productions, these differences are nullified because of the size of the finished video we use. Constructed for playback on computer monitors, we size the final video at 960x540 — actually quite large for computer video and the context in which we use them. (By comparison, most YouTube videos are either 400x240 or 600x360.) In fact, any larger and some folks might find they can't see the entire video because their computer monitors are too small. So, the 720p format video camera (1280x720) are enough for my needs for our interviews and for production requirements for LensWork Extended. The new Panasonic G2 shoots 720p (1280x720) natively and will be great for this use.

There is one other cutting edge consideration that was worth checking out. I loaded up our streaming video on my Android tablet which I piped via HDMI cable into our Sony 40" LCD television. I wanted to see how the video looked there just in case people decided to view LensWork Extended on their HDTV. Looks great — not like full size HD cable, but great nonetheless. Even at that large size, the ZS3 pocket camera is only very slightly softer in focus that the other two, far more expensive cameras. The differences are so slight, I doubt anyone will even notice. I am fully satisfied.

So, a concluding word of advice from the pragmatist in me: In the race for more and more pixels and higher and higher definition video, it does appear that just as in other areas of life, "Enough is as good as a feast." I may have to fight off the seduction of 1080p video, but for actual work it appears that 720p has a lot more to offer than the camera sellers would have us believe — at least for those of us not doing feature films or episodes of House for commercial broadcast outlets.

Premiere Pro CS5.5

There are probably not many applications that will test the horsepower of a computer system like video editing. I've been frustrated with Premiere Pro throughout my relationship with it because of frequent hangs, crashes, jumpy rendering, interminable waits for video output — you name it.

Of course, I blame Adobe and Premier Pro, but it turns out it was my underpowered computer. (Sorry, Adobe.) My new Yippie Skippy super-fast Dell runs Premier Pro CS5.5 like nobody's business.

Yesterday, I edited video. Actually, I edited three videos — in one day! That's unheard of in my world. One of them was a 30-minute, three-camera shoot of my interview with Doug Ethridge. It was downright easy and the software did precisely what was supposed to do handling the multi-camera edits, when it was supposed to do it, with no lag, and not a single hang or crash. And I might add, I don't even have the recommended, super-expensive video card that can do GPU rendering of video on the fly. I can only imagine how this machine would perform far were willing to spend another $3000 for the video card. But, I'm not and am happy to report that — at least for my needs — I don't need to. I guess the 64-bits in combination with 16 GB of memory is enough to least justify paying for the upgrade to this latest version of Premier Pro.

Report on the GF1 videos

For the production of the new LensWork Visual Workshop, Finding an Audience for Your Work, we decided to step up to HD video recording. Using the same kind of rationale that I used before, we looked for a digicam with HD video capabilities — rather than purchase a dedicated video camera. After a bit of research, I was delighted to settle on the Panasonic GF1. This offered two advantages: I could get the 20mm f/1.7 pancake lens that also can be used on my G1, and it can serve as a backup camera for the G1 when I travel. A neat solution.

So, how does the video perform? I think, very well for our purposes. Check out the sample video we posted here for a peek at a video made with this camera in our studio.

A few comments that may be of interest:

• A large motivation for selecting this camera is the ability to focus manually. This is a must for studio video recording where "focus hunting" can be an awful distraction in the final product. Many digicams do not offer manual focusing, so beware of this missing feature.
• Also, the AE lock is a must. In a controlled studio environment, exposure should not drift up or down during a recording. Again, lots of digicams do not have this feature.
• I find the white balance of the GF1 slightly cool. Thankfully, it has a WB control that allows shifting by 100K units to a precise setting. Very nice.
• The 20mm lens is perfect. Because it is an f/1.7 lens, we could used a cooler 500W Tota bulb rather than the much hotter 750W bulb.
• Part of the process that makes such a camera work for us is that we don't actually use the full HD video size in our final production. Since our videos are seen on computer monitors, we don't need video that is 1280x720 — too many people would have problems with video this large on their smaller monitors. Instead, we produce our videos to 75% of that size — 960x540 — which will fit everyone's monitor (and browser) and is plenty big for our use. This reduction in size helps create an even better image with these class of cameras.
• The file format we are using is the AVCHD option at the best quality available in the camera, so-called SH for "super high." It's still a compressed format, but works great because we are not doing any post-processing other than trimming the file and plopping on some overlays.
• We are using Camtasia 7 to produce the videos, but it doesn't accept the MTS format files. (I'll have more to say about Camtasia in a future post.) We convert the MTS files from the camera into high resolution WMVs using Emicsoft's conversion software. These we import into Camtasia for final editing. It's a workflow that is clean, efficient, and produces great looking videos.

We couldn't be happier with this camera for this use. In the field? No idea yet, haven't tried it. For still images? Ditto. But, I can report that for the kinds of video we are making in the studio and at the size of our final output videos, it's a peach of a machine.

Technological Faith

Software engineers may think differently than the rest of us, but they are no dummies. I admire them. I thank them. More than that, I have faith in them. Here is an example.

During this last week I've been swamped — hence the dearth of posts here — doing some videos previews for LensWork #88 for our new website. (See the video previews on this page if you are interested.) I've spent the week deeply down the rabbit hole of Adobe's Premiere Pro CS4 doing video production.

Now, I love this software; it's simply amazing to work with. But, I became frustrated with one small aspect of it, to wit: I wanted to have a dissolve transition between each image. After placing 6-10 images in the timeline, I would then manually apply a cross-dissolve transition at each junction. Manually. One at a time. What a pain.

This is where my faith kicked in. Surely, I thought to myself, the software engineers at Adobe would have run up against this annoyance in their work, too. There must be a solution. Poking around the menus, sure enough — "Apply Default Transitions to Selection" does the trick.

My point is this: It is not a coincidence how many times in my computing life that I find myself thinking, "There must be a better way to do this" when, in fact, there is. Having a little faith in software engineers has paid off for me time and time again. If I'm having a need, probably they've observed the same need and it's their job to provide the answer. So, as a strategy, I've learned to trust them. Whenever I think, "I wish the software did this" I go searching. Far more often than not I find that it does, but I didn't know about it. It's almost as if a solution to a problem doesn't exists until you have the problem and look for the solution. Which, I think, is exactly the case.

Said a different way, when I become frustrated (or angry!) with software, the honest truth is that I am probably the cause of the problem; my limited knowledge is getting in the way. Rather than take it out on the software and complain about its shortcomings (which solves nothing), I try to remember to have faith, search for the solution in the software, and leave my emotions out of the equation so I can get back to work. It's amazing how many times I find just exactly the solution I was hoping for.

It's all about distance

This last week I've been testing a couple of microphones in our new video studio in preparation for recording a new LensWork Visual Workshop. Rather than hassle with a lapel mic, I thought I'd try a desk mic — sort of like Johnny Carson's old desk set up. (For you youngsters, he's the old Jay Leno.) The idea of a desk mic has always appealed to me because once it is set, you don't ever have to move it or do any re-testing; just sit down and start talking. It's a lovely theory.

Unfortunately, there is just no getting around the old rules of thumb. Just like in photography, it's all about distance. The closer the source is to the recording device, the better the resolution of details and the better the fidelity. Period. After four hours of testing and re-testing both the desk mic and the lapel mic, the comparisons of the two made the decision easy. We've decided to stick with a lapel mic.

Listen for yourself:

Deskmic

Lapelmic

Not a debate. A modestly-priced lapel mic at 5-inches from my mouth is simply better — by audible amounts — than a more expensive desk mic positioned 16-inches away. You can fuss all you want, but the distance is the determining factor. Four hours out of my life because I keep forgetting that some rules simply cannot be broken without a consequence.

I remember once spending days trying (and failing) to make a decent 11x14" print from a crop of about half of a 35mm negative — all because I didn't get close enough to the subject. That was in 1970. Forty years later and I'm still learning the same lesson.

Video on a budget

I've been asked why I use an amateur digicam like the Panasonic DMC-ZS3 for video recording rather than true profession gear. Other than the obvious huge cost difference, there are several reasons why we've found digicams just fine for our needs.

1. Recording length. We do video interviews with photographers for LensWork Extended. The raw footage can sometimes be an hour or more worth of conversation with an engaging photographer. I was always frustrated with the Canon S5 IS cameras because they automatically stopped recording after 31 minutes. We'd have to start them again, necessarily interrupting the interview or doing a clever, carefully timed, staggered three-camera splice. With the ZS3 using AVCHD, they will record until the card is full which means 58 minutes at "super high" quality on an 8 GB card, or an hour and a quarter in "high" quality. This is also why a camera like the superb Canon 5D MkII with its HD video capabilities is simply not practical for interviews: it tops out at 12 minutes per recording. (Even expensive dedicated DV digital video cameras are limited to 1 hour tapes.)
2. Focus lock. For recording things like interviews, it is deadly to have "auto focus" engaged. From time to time the focus hunts and is quite bothersome when viewing the final result. For a camera to be used for interviews, it must have manual focus or at the very least a focus lock. With the ZS3 the auto focus can be turned off so the focus is locked to the initial focus distance, perfect for interviews.
3. Exposure consistency. Ditto exposure and white balance. All it takes is a person with a dark or white shirt to start gesticulating during an interview and the exposure of the entire scene can fluctuate. Locking exposure is the perfect solution. Unlike the Canon S5 IS we've used in the past, the Panasonic ZS3, unfortunately, does not have this feature. It was almost a deal killer, but then I found that using the "Multiple" setting seems to eliminate the wild swings of exposure that can happen in some situations. It's not ideal, but perfectly functional.
4. What about sound? For many videographers, the single feature that makes most digicams unusable for serious video is their sound quality. The knee-jerk answer is to have a mic input on the camera. Some digicams have them, but often they are 3.5mm stereo plugs like the headphone jack on your iPod. (The Panasonic GH1 and G2 (stupidly) have a non-standard 2.5mm stereo jack.) Sure, that solves the problem by allowing external mics, but are just hokey. Worse, the audio controls in the cameras are often limited to "auto record level" — a fatally flawed feature for almost any serious audio recording. We use a different answer. We record the sound via lapel mics and independent, small MP3 recorders (Zoom H2) that are worn in belt packs by each speaker. No lengthy cables between subject and camera to trip over; no need for a mic input jack in the camera. The camera-recorded audio is only used to sync the independent MP3 audio to the timeline in the video editing software and then muted for final output. It's a slick solution that produces professional quality audio without a budget-busting video camera with XLR mic inputs.

   

    

5. Finally, the ZS3 has a nice wide angle (25mm), records in HD (1280x720), has changeable aspect ratios (16:9, 3:2, and 4:3), and even image stabilization for the occasional hand-held shot. Not bad for about $350. We can do three-camera shoots for about a third of the cost of one professional DV camera.

   The ZS3 is still available, but is being replace by a newer model, the DMC-ZS7 which is just now starting to ship. Same basic features but with a few updates. Roughly the same price.

   

   The point in all this is that, depending on your video production needs and final output, it can be a perfectly functional solution for video recording to use inexpensive digicams and MP3 recorders that have the right feature set. If you want to spend $3,000+ plus per camera for something like a Canon XL2 dedicated video camera, you certainly can.

   

   But it's not always necessary if you are willing to do some research and find a solution that fits your needs.

How Much Video Data Is Enough?

When I first starting working with compressed audio files — the now universal MP3 format, for example — I immediately learned about the importance of bit rates. I understood the concept because I'd seen essentially the same thing in film size: the larger the negative, the better the quality. Likewise, the larger the bit rate, the better the quality. Easily understood.

But, then comes the practical measurements where the proverbial rubber must meet the road. How many bits are enough for encoding music — 48k, 64k, 128k, 160k? More? And what about MP3 versus WMA, versus all the other compressed audio formats? There is only one way to find out for sure, just like there was in my early film days; you have to test for yourself. Others' opinions are just that — their opinions.

First, my criteria for "acceptable" may be different than anyone whose opinion I might read in an Internet forum. Second, my needs may be different for my particular uses. Test, cross-check, and test again. No shortcuts. A day or two later, lots of test files on several different kinds of audio, listening when I was tired and when I was fresh, listening via speakers and via headphones, and finally doing a bit of blind test checking and I had my answers.

I bring this up because I'm facing the same issue yet again, this time with compressed video. My pocketable digital camera records great HD video — really superb for my uses. It's a Panasonic DMC-ZS3 which records 1280x720 HD video, far larger than I will need. In fact, I'll end up reducing this to 75% with my final videos running 960x540, plenty large for computer screen presentations in the 16:9 ratio.

I'll use three of them for three-camera shoots. I've been doing this very thing for the last several years making 640x480 interview videos for LensWork Extended using three Canon S5 IS still cameras. (Who needs a dedicated video camera? But, I digress.)

The Panasonic ZS3 records video using AVCHD compression. It's a VBR encoder creating MTS files. Never heard of them and can't play them with my usual players. So, for archiving I need to re-encode them into a standard format that I can play and repurpose for post-production. Which format should I use? Which bit rate? Test, cross-check, and test again, keeping in mind that my answers may be only applicable to my specific needs.

Because I do all my editing on a PC, I decided to first test the standard WMV format. Well, sort of standard. I'm actually using a WMV HD format, encoding from 1280x720 down to 960x540. But at what bit rate? Since these are archive files, I want all the quality I can swing for use in post-processing somewhere down the line. I started with 15,000 kbps, the highest bit rate my converter offers. Wow, that's a lot of data and far more than I probably need. The recompressed WMV files are indistinguishable from the original MTS files. So, I started testing, compressing the same set of files as various bit rates. I finally settled at 10,000 kbps for the master archives. Why? Because I could just barely see a loss of quality at 5,000 kbps, and imagined I could maybe see a fleeting loss at 7,500 kbps. I have confidence that 10,000 kbps will give me everything I will need for editing and production for disc distribution — way more than I will need for Internet distribution.

Next, I'll test a couple of the other video formats, but I at least know I have one format that will do the job. I have confidence, but I'm not so foolish as to think I could never be wrong or missing something. I'll keep the original MTS files in deep storage, too, just in case.

Who Controls Time?

We are asked with some frequency why LensWork Extended is not available in DVD format for playing on a television like a DVD movie. The technology is relatively easy to create and it's not a matter of difficult authoring that prevents us from doing so. Rather, it's a matter of timing in the sense of the viewer's experience of time.

In the world of technology, there is what I think of as "the great divide" — those media that are viewer-timed and those that are author-timed. Movies are author-timed; PDF slide shows are author-timed; live lectures are author-timed; music is author-timed. In my way of thinking, photography is best when it is not author-timed, but when the viewer sets the pace of viewing. It's enough for me as a photographer to control what is seen, and perhaps in what sequence it is seen. But, I don't feel a compunction to control how long a viewer sees any given image. In fact, it's perhaps as a viewer of photographs that my prejudice was formed — I want total control of that characteristic for myself when viewing photographs, primarily because I find I prefer to spend far more time with an image that most people do.

Why? I've always thought that viewing photographs is more of a dialog that an entertainment. As a viewer, my responsibility is to look, ask questions, think, place the work in context of my life experiences, ask more questions, even ask that most leading of all questions, What if?

This kind of dialog only happens when there is enough time to do so. Sometimes, I feel a photograph is worth a minute to two. Sometimes an hour to two. Almost never do I think a photograph (I'm talking about serious artwork here) is worth a fraction of a second or two as is so often the case now in our age of the quick cut.

Author-timed media has accelerated in my lifetime to rates that would astound and confuse our grandparents. I'm not sure this is a virtue. There is a lot of power in the pregnant pause that is all but absent from today's author-timed media. This is where the book, the PDF, the still photograph can still be of value — providing the time to think and react that is missing from author-timed media.

Of course, there is a risk for us photographers when we let go of control of time. Viewers can be fickle. Work we intend for leisurely viewing replete with meaning and counterpoint may be consumed in a glance, discounting all the subtlety we worked so diligently to include. It's a risk I think is worth taking for this simple reason: the best strategy as a producer is to create for that subset of viewers with whom we connect at the deepest level. I know of no other philosophy of art making that promotes the core of art's dual purposes — experiencing and sharing with the pursuit of excellence.

Besides, I'm pretty convinced that the author-timed media creators have the same problem. One can watch a movie while cooking, doing homework, talking, or otherwise distracted from full immersion in the artwork. Media does not determine connection; connection determines connection. You may quote me on that.

So, I prefer to have faith; faith that some viewers, albeit a few, will want to take time to connect with what I produce; faith that allowing them control of the time they spend with my art work will foster connections that are appropriate for them, long or short; faith that what I produce will, indeed, connect with someone, somewhere, in spite of this philosophy being in such opposition to the zeitgeist of our times.