Skip to content

Little Frog in High Def

Musings of an NLE ronin…
description of the photo

Archive

Archive for March, 2009

The seventh episode of THE EDIT BAY is now available for download. This one is about intentionally makking mistakes. I did that on purpose…really, I did.

To play in your browser or download direct, click here.

To subscribe to this podcast in iTunes, CLICK HERE.

I rushed my cat to the 24 hour emergency vet last night. Bad wound on his tail from a battle wound he apparently got weeks ago. Infected. And the bill just completely wiped out my budget for Vegas. So, I will be staying home, petting my poor cat and attending to his wounds.

Next year.

Alex over at EDITING ORGANAZIZED has made a quick simple plugin for creating a lower third with background. Even made a gradient background possible. Oh, and this plugin is free…and we all like free, simple and easy to use.

Gotta love the simple stuff.

Thanks Alex.

Watch this video.

‘Nuff said.

Other than the site that it came from is HERE. So if you want the story, there’s where to look.

Seriously, look at it. Automatic Synchronaization by Singular Software is truly amazing.

Shooting with 2 or more cameras? Did you start and stop one or more of them a few times? Oh, did you forget to slate or clap? Tired of searching for a common event to sync to? Me too. This does all of that, automatically.

Watch the demo…then download the beta with sample software and try it for yourself. I did…pretty freakin’ amazing.

NOTE: This says it is a PLUGIN, and it really isn’t. It is a separate application that is stored in the Applications folder. Highlight your clips, then run the App.

If you go to iTunes, and go to PODCASTS>TV & FILM and choose AUDIO PODCASTS, you will see that THE EDIT BAY is “New and Notable.”

Sweet!

To subscribe to this podcast in iTunes, CLICK HERE.

To those who shoot and edit with HDV, this is nothing new. If it is, then you haven’t been looking hard enough. This is something VERY common with the GOP formats, HDV specifically. A producer of mine was seriously considering a Sony ZU1 as the b-camera to a Varicam. Until he shot some test footage, and went frame by frame on a zoom shot and saw…well…this.

This is the crap you see when you pan, tilt or zoom quickly. Now, all of you HDV fans, justify THIS. Tell me that this is acceptable. Apparently it is because Discovery cleared HDV for BRONZE level programming, and it is acceptable by other networks. BetaSP never did this. Digibeta never did this. D2, MII, Umatic 3/4″, D1…even DV never did this. Yet this is common with HDV.

To me this is unacceptable. But, this is what I am stuck with for now.

Most people think that ‘rolling shutter’ is something they can live with too. Why are our standards getting worse?

THE METAMORPHOSIS OF THE DARK TOWER

When it comes to a lot of things I am a do-it-yourselfer. One of those things has been a self made RAID. And the one I initially made gained me a little notoriety. This was called the Popsicle SATA. This all came from the need to play back uncompressed 10-bit footage, without breaking the bank.

Now that RAID didn’t hold up…structurally speaking. The rubber bands were wearing out. So I upgraded that RAID by using screws to replace the rubber bands.

But then the internal heat on the G5 was getting to be too much, so I then purchased an external PC case, and thus the Dark Tower RAID was born.

Now I have gone to the next level.

I was asked by Enhance Technology if I would like to try one of their products. I asked for the StorPac S35T…that is a 5 bay tray loaded unit that takes up 3 5.25″ slots on the front of the PC case. They sent it…I installed it…and promptly asked to keep it. Well, I said that I wanted to pay for it, because I liked it. A lot.

Here is what it looks like:

5 removable trays, and you can buy additional trays so that you can just have them on all the drives you need to put in this thing. Let’s look at the back:

Three ports for power (I have that), and then 5 SATA connections…the blue ones on top. Because these are direct SATA connections, you’ll need to get a Port Multiplier…like I did. I got this a while ago, so that I could attach the 5 drives I have in this thing to the CalDigit FASTA 4e I have. Works great.

OK, so I installed the unit, and attached the SATA connections from here:

To the Port Multiplier:

So I went from internal drives that were pretty difficult to swap out (which I needed to do a couple times as I backed up my P2 footage) to a case with hot swappable drives. From this:

to this:

Ahhh…much better. Now you see why I didn’t want to return it.

Now we run the AJA Drive Test on that and we get what I expected:

Those are the speeds that you can expect from a 5 drive Raid 0. This was how I played back uncompressed HD. Now…the scary thing is that, well, now two companies offer these speeds in a TWO drive raid. MUCH cheaper too…CalDigit and G-Technology.

I currently am using this Raid for two purposes. As my show backup (backups of all the shows I worked on) and P2 backup drives…then I remove them and put in four 1TB Hitachi drives that I am using to store the footage for a doc I am working on. All works pretty well. The fan is a little noisy…I can now tell that the unit is on. Before it was so quiet that I would forget to turn it off. But now it is a tad louder. Not good for audio editors, but fine for me. The fan keeps the drives cool and that is what I want. And it isn’t like it sounds like an aircraft carrier.

So if you are a do-it-yourselfer and want to build your own eSATA Raid 0 unit with PC cases and the Enhanced Technology enclosure…this is for you.

Get them while they’re hot! The new MXO2 1.3 drivers are out. They enable the MXO2 to do quite a lot now.

Matrox MXO2 version 1.3 provides the following:
• Scaling on the input for capture.
• Improved support for 720p workflows.
• ProRes 422 720p capture on a MacBook Pro.
• Improved support for 1080p workflows.
• Support for 1080p at 23.98 fps via the HDMI output.
• Preservation of 23.98 fps when cross converting from 720 to 1080.
• RGB component output.
• Apple Xserve support.
• Matrox MXO2 Rack support.

Be sure to download the MXO2 documentation to read up on this yourself.

So I will be attending NAB this year…ON MY OWN. I will not be working for anyone, which is a first for me. I’ll be able to take my time and walk the floor and really look at the offerings. Instead of gobbling down lunch and dashing around the hall grabbing brochures and glancing at the products. I can also stay up late partying, and SLEEP IN. Before I would still stay up late partying, but have to get up early and work the floor the next day. SO to those who I talked to first thing after the doors opened, I apologize. Now I can sleep in and get rest and lounge by the pool for a while subjecting the other poor guests to my not so hard editor’s body.

WHY AM I SAYING ALL OF THIS? Well, to let people know I will be there and to keep an eye out for me if they want to meed me (picture is here on the Creative Cow, although my hair is a bit shorter)…and to say “Hey, if you want to go to, I have a free passcode for you to use to get you onto the floor for free.” So…

Hey, if you want to go to, I have a free passcode for you to use to get you onto the show floor for free:

Free Exhibits Passport Code: TP01
This pass includes access to the exhibit floor and the opening keynote – a $150 value. Please pass along and visit http://nabshow.com/passport to redeem.

There, now your trip to Vegas just got cheaper. Now you just need to find a decent hotel. And take cabs…rental cars and driving is just too much of a pain. Plus with all the partying you do (this is Vegas)…you’ll want a cab.

Walk the floor…there is a LOT to see…covering all aspects of broadcasting. And look at the SUPER SESSIONS too…they offer valuable information on a variety of topics from the top people in their field. Yeah yeah, it is back to school in Vegas. Just know that after school activites are better than they were in high school. Well, funner.

See (some of you) there.

The sixth episode of THE EDIT BAY is now available for download. This one is about running into someone you worked with before, but never actually met.

To play in your browser or download direct, click here.

To subscribe to this podcast in iTunes, CLICK HERE.

OK, I need to clear the air with some people. On this current job, I am considered a FINISHING EDITOR…or another term for this is ONLINE EDITOR. I am not, nor do I consider myself, a COLORIST. A colorist is someone who only color corrects, and can work MAGIC with Color, or a DaVinci. Me? I can make things look good and make the signal legal for broadcast. I might be able to pull off a decent secondary or two and use a power window, or regions, whatever you know it as. But I do not know the finer details of color correction…I make no claim to that.

My job is to take the creative cut, prep it for online by dealing with the speed changes and moves on stills…add the final graphics, add the final text and credits. Media manage the project and recapture any footage that might need recapturing. Then yes, I take it into Color to color correct and send it back, but most of the stuff was shot in camera so I don’t have to treat it much. Mainly I make sure it looks good, consistent, and is legal. Then I output to tape. A lot of busy work. The work of an online editor.

And thus far I am enjoying it. Enjoying getting out of my creative head and into my technical one. At least for a little while.

FINALLY, what many of you have been waiting for…the results of my independent tests of the G-Raid 3 and CalDigit VR. I’m going to cut to the chase…if you want to know what lead to this, read this blog post.

Now…for all of these tests I relied on what EVERYONE relies on to test drive speeds, the AJA SYSTEM TEST application found at www.aja.com. But I also utilized a new tool that John Flowers has been working on…the HD SPEED TEST. I used version 1.1, beta 4. This is still in beta, so keep that in mind. To read the reasoning behind this app, and how it works, click here.

And the machine this was tested on is an early 2008 MacPro 3.0Ghz with 4GB RAM, and the eSATA controller card is the CalDigit FASTA 4e card. FYI.

OK, moving on.

First up…the G-Raid 3…because it arrived first. The FIRST thing I did was run a speed test on my original G-RAID…the FIRST one, the one with PATA drives and not SATA drives. So I could have some sort of comparison. Just for kicks and giggles really. So here we go.

The original G-RAID…via firewire 800:

About 60 write and 67 read. Not bad. This is fairly decent for DVCPRO HD…I can get 3 streams. OK, let’s look at the the G-Raid 3 via Firewire 800:

68 Write and 83 Read…significant improvement.

Now, what you have been waiting for. G-Raid 3 via eSATA. Ready? Sit down.

220.5 MB/s Read, 217 MB/s Write. From TWO DRIVES! Wow. Before this you needed 5 drives to get this performance. Now, let’s see what the HD Speed Test got us:

Well, we see a variety of numbers…one for each codec type. Which one to believe? Well, I guess that depends what format you are working with….there is no “overall” ranking. Not sure why certain numbers are lower with certain formats, but as you note, when I run the AJA test I choose the highest format possible, 10-bit uncompressed 1080.

Let’s take a look at my home built Dark Tower SATA RAID. This is a 5 drive (5×500GB = 2.5TB) port multiplied device.

SAME NUMBERS as the G-RAID 3. Five drives vs two drives…both eSATA. I did not run the HD SPEED TEST on this setup.

OK…but that was a test with the drive EMPTY…which every one likes to test their drives as because, well, they are faster. But let’s take a real world stab at this. Let’s fill the drive halfway…put 1TB of stuff one it. Music, video files, pictures…the usual stuff. Now we can see some real numbers, not marketing numbers. This is where I employ both the AJA test and the HD Speed Test. First AJA:

189MB/s Read, 185MB/s Write. Still very respectable. Lets look at the HD Speed Test:

Just a small dip in performance…like 3MB/s across the board. So those are great numbers. The thing that gets me is how the AJA tool drops so much while the HD SPEED tool remains consistent. Whatever…believe the numbers you want to believe. I might be more inclined to trust the HD SPEED TEST, because the streams of RT it shows is what I get when I run my tests.

OK, moving on.

Now onto the CalDigit VR. Just like with the G-Raid, I am going to run the numbers on the S2VR Duo first…just for comparison. Bear in mind that unlike the first G-RAID that had PATA drives, this one utilized SATA drives, and I am connecting via eSATA, so it will be a bit faster. I just wanted to compare the S2VR Duo to the new CalDigit VR. So…the S2VR Duo via eSATA:

154MB/s on both Read and Write pretty much. Very good. Now the CalDigit VR…empty:

216.6 MB/s Read and 204.4 MB/s Write. Slower than the G-Raid 3. Not by much…they are close, but it is slower. There is a TURBO mode on the drive, so I turned that on and saw if there was any difference:

Slight. Nearly the same results Read, slightly better Write. But those numbers fluctuate plus or minus 4 points anyway…I ran multiple tests and they were up and down, I just chose the average numbers.

OK, running the HD SPEED TEST:

OK…looking at all the numbers we note only a slight difference. One to two points lower than the G-Raid3. Slower, but not a noticeable difference.

Moving on to the real world tests. The CalDigit VR 50% full…exact same stuff that I put on the G-Raid 3:

178MB/s Read, 173MB/s Write. A full 10MB/s slower than the G-Raid 3 at the same capacity. Close…but the G-Raid clearly wins. Now, the Flowers HD Speed Test:

IT’S FASTER! According to this test the drive is FASTER with more stuff. So it doesn’t make much sense. I am not a highly technical guy, so I’ll leave this to the technical guys to figure out. Going by the current industry standard the AJA test, the CalDigit unit is slower. But the Flower’s Test has it faster. Who to believe? I guess this will be up to debate.

According to the AJA test, the G-Raid3 wins hands down…full and half full. Clearly faster. Is that difference noticeable? Doubtfull, but…G-RAID3 wins. According the the HD Speed Test…the G-RAID3 also wins empty, but half full, the CalDigit wins. I’ll chalk this up to the fact that the HD SPEED TEST is still in Beta.

Note that both had dips in drive performance. That is due to neither of them having a dedicated RAID controller card. Without that they are on their own and prone to those dips.

OK…there was also question of the STURDINESS of these drives. Which one was the more solidly built. Let’s take a look at them just sitting there.

The CalDigit VR is clearly bigger. And yes, it is heavier. Nearly twice as heavy as the G-RAID3. The CalDigit VR weighs a whopping 6.6 pounds…the G-RAID3 a mere 3.8 pounds. If this sits on your desk that wouldn’t really matter, but if you pack this for on the road travel, or want to throw it in your backpack, that is a BIG difference. Gotta be the added components, because this drive is very versatile. The G-Raid is a simple drive…RAID-0, plug and go. The CalDigit VR has many options. But I’ll tough upon those later.

DURABILITY. Well, by holding them and pushing on them and poking them fully assembled, they are both pretty solid. Not sure what the big issue with durability is. If you drop the drive, the last thing you worry about is the integrity of the case…the drives will be jostled so much the data will be as risk. BUT, since someone called this into question, let’s open the cases.

G-RAID3


Looks solid enough…and looks exactly like the other G-Raid cases I have had. Nothing’s changed. But, why change it when it works? Looks good, feels solid, nothing lose…even the SATA cables seem secure. I see no issues with this case.

CALDIGIT VR


Now, right away you can see something that might be cause for concern….there are no four solid walls, but a faceplate that seems to be only connected to the base. That would seem flimsy. However, that isn’t the case.

See that bracket? Well, that is the anchor. It keeps that faceplate solidly in place. Yup, I jiggled and wiggled…it is pretty firm. I have no doubt that this is solid and will hold up. Even to being dropped from a few feet. But again, if these drives are dropped from your desk, the last thing you need to worry about is the case. It is the second thing. Since I wasn’t given permission to destroy these drives, I did not perform that test…the DROP test.

Now…let’s look at the price and options these drives offer.

The G-Raid3 is simple. $499 for 2TB ($519 on B&H, and they are usually the CHEAP place to go). Quad interface…USB 2.0, FW400, FW800, eSATA. Adapter for eSATA is extra. And the only option is Raid 0…meaning the two internal drives are hardware raided together…striped for speed. Simple, cheap, sturdy, reliable…FAST. Freaking fast for a two drive unit eSATA. Worth every penny.

CALDIGIT VR is flexible. $664 for 2TB ($165 more than the G-Raid3). Again, Quad Interface….USB 2.0, FW400, FW800 and eSATA. And this unit ships with a free eSATA extender. This handy thing allows you to connect to the two spare eSATA connections on the logicboard of the MacPro…and provides instructions on how to do so. And unlike the G-RAID3, this unit isn’t only RAID-0. You can configure this Raid 0, Raid 1 (protected…mirrored data on both drives) and JBOD, meaning the drives show up as separate drives. And the LCD and interface allow you to not only monitor the heat, fan speed, raid type and a few other things (serial number, firmware version), but also works as an interface to do a few things like choose the RAID type for the drive. And the drives are removable, so you can use it as a stand alone storage case and just order more drives as needed.

OH, and both drives have the same level of noise, which isn’t much. Sure, if it is on the desktop with you, you’ll hear them. But if they are below the desk on your computer, or off to the side somewhere, you won’t really be distracted by them. Both fans are pretty decently quiet.

So…final tally.

SPEED – G-RAID3 wins. Minus that little oddity in the HD SPEED test (app still in beta), it wins using the industry standard test…both empty and real world scenario.

DURABILITY – I would say tie. They are both very solid when together, and very solid when taken apart. Since I didn’t drop either one, I can’t say for sure. But they are both solidly built, no matter what the companies say about the other guy.

FEATURES – CalDigit wins with the ability to configure their unit in one of three ways, and with the ability to monitor various things.

NOISE – Tie.

COST – G-RAID3 wins. $165 cheaper than the CalDigit…both in 2TB configurations.

CONCLUSION? If you want a nice simple drive to edit HD video with that provides lots of space and speed, the G-RAID3 is for you. It is light, simple to use, and durable. Great little unit. I lost faith in G-Tech a while back when the G-Raid 2 had issues with daisy chaining the units (performance would drop from 65MB/s to 16MB/s), but this drive has returned my faith in the company and their products. The G-Raid 3 is a great little drive that I’d recommend to anyone.

If you want a drive with more features and one that is customizable…if you have a bit of the inner geek in you and need such things (I do at times), then the CalDigit VR is the drive for you. Fast and durable, and with features that make the extra money needed to get one worth it to many people. CalDigit has always been one to turn out great products, and this is no exception.

CHOOSE ONE? I can’t. I like both…can’t I have both? If cost was a factor then the G-RAID3 will be the clear winner. But if you have the extra cash, the added features of the CalDigit VR might be worth the cost. So YOU choose. I gave you a bunch of things to take into consideration. Now it is up to you.

(NOTE: If you look seriously at the HD SPEED TEST results…these units are very close, with the G-RAID3 winning empty. But then the CalDigit VR does BETTER when it has footage on it. I ran the tests with the full 1.2 non-beta version of the Speed Test and the results remainded the same. More data on the Caldigit, the faster it gets. Now, the debate about the validity of this tool can commence.)

OK, I am posting this blog again…making a few changes and adding the fix that I got to work this morning. I originally posted in in anger and frustration, and one should NEVER post in anger and frustration. I have since cooled down, and have made changes to the post to tone down my original rantings. It is still about 90% intact, and addresses the issue. OK, here we go.

If I died and went to editor hell, I would be glued to my edit desk and forced to edit HDV for eternity. So far that is what my day feels like. Something that should have taken 1 hour, has taken 7 hours of my life. I have been beaten down by this format after 1 week of working with it.

What is my latest issue? Well, it is pretty freakin’ bad.

I am attempting to capture HDV footage into Final Cut Pro FCP 6.0.5 on a MacPro (early 2008) running 4GB RAM and utilizing an AJA Kona LH capture card. The HVR-1500 HDV deck is connected via HD SDI, 9-pin (RS-422) deck control. We are capturing HDV shot on the Z7U and we are capturing it as ProRes…some shots capturing offline as DV. Why? Don’t ask me. That was something that happened before I got here….they had their reasons…DECK issues. Something I have talked about before.

So what I needed to do was recapture all of the DV footage as ProRes, to match the rest of my footage. So I duplicated the sequence and separated all of the DV footage from the ProRes footage…and deleted all the footage in the sequence save what I needed to recapture.

From there I highlighted the sequence and opened the Media Manager, and chose CREATE OFFLINE, and chose Kona LHe ProRes 422, chose 1 second handles (to deal with time code slippage)…and clicked OK.

So now I have a project with all the media with handles that I need to recapture. But 6 times out of 10…60% failure rate…I get this error message:

Now…the media has been captured, it just won’t reconnect. If I try to capture again, it will fail again. After a few attempts it will capture and reconnect. But now I have MULTIPLE captures on my hard drive taking up space….because they were captured but couldn’t reconnect. BAD.

When I go to manually reconnect this footage, turning off the MATCH NAME option and navigating to the file that was captured for that clip (I navigated to the last clip captured) and I get this error:

If I force the issue and tell it that I don’t care…connect to that clip, it says that the output was adjusted. And now the clip is 2 to 4 seconds off. Outside my 1 second handle zone and unusable…as it doesn’t match the offline cut. It is way off.

OK…let’s look at two files that were captured for one clip. Both of these files say they couldn’t reconnect.

Two completely different file sizes. So not only is the out point incorrect, and the in point incorrect, but the DURATION is also incorrect. This is a bag of hurt.

Now, let’s open one of the failed clips and look at the MEDIA START timecode that was captured, and compare that to the MEDIA START that the clip needs in FCP…

Over 4 seconds off. VERY bad.

Here is a pic of a clip that was captured properly…the same clip as we see above:

The timecodes match.

SO here I am, on hour 6 of capturing footage one shot at a time…making between 4 and 10 attempts per clip. With a clip being captured successfully at the first attempt about one out of 10 times…causing mild surprise. Surprise that it worked the first time.

NOW…THAT WAS YESTERDAY. CUT TO TODAY. I came in early in the morning, determined to figure this out. I posed this issue to a few people, and got a few suggestions. As you can see in the comments, Andrew suggested a viable workflow that Walter Biscardi from the Creative Cow says that he does, and that works for him. I also got a couple more suggestions, that had me using the current deck, so let me go there first.

The FIRST suggestion was to change a setting in the DECK CONTROL PRESETS in the Audio/Video Settings from USE DECK CONTROL SEARCH MECHANISM to DO NOT USE…because obviously the deck control was faulty. So I tried that. And things got a LITTLE bit better. 10 clips, 7 captured fine…3 did not. Better than the 4 captured 6 did not in the nightmare listed above. But still not good. So then I hooked up firewire and tried firewire deck control. Nope…back to being bad bad bad. Even switched the firewire to DV…same issue.

Then I tried what Andrew said in the comments….that Walter Biscardi method. His method is to NOT USE THE HVR-1500 DECK AT ALL! How about that? So I grabbed our other HDV deck, the lower end HVR-M10U. This has Component outputs, so I used those to connect to the Kona LHe. Then I used firewire as deck control.

Guess what? It worked. FLAWLESSLY!

Seriously…I highlighted 10 clips, bit BATCH CAPTURE…it captured all of them, without that error popping up once, and every clip was…well, at least 2 frames in sync. Most were spot on, 3 needing shifting. Still…that rocked! Thank you Andrew for pointing out the solution that Walter came up for this. Oh, the one drawback? The M10U shuttles in near real time. Meaning that it fast forwards and rewinds maybe 1.5x to 2x faster than it would play the footage. That means that it will take some time. But it is still a much better thing that that 1500 deck and the nightmare capture issue I had. OH, and HDV does look very pretty…I’ll give it that. Although I haven’t started color correcting it yet…so…we’ll see.

OK…reposting this. But I will leave the sheep video at the bottom, for your amusement.

Thanks everyone.

The fifth episode of THE EDIT BAY is now available for download. This one is about bidding on an agency job, and losing it for the oddest reasoning.

To play in your browser or download direct, click here.

To subscribe to this podcast in iTunes, CLICK HERE.