128gb SDXC, or a 64gb SDXC?

From reading on tech sites, I have found interesting tidbit about ERROR rate (not to be confused with failure rate):

As NAND manufacturers pack four or more bits per cell, error rates increase dramatically. Thus, smaller-capacity cards actually have much higher error rates than larger-capacity cards only because there are simply fewer cells in the smaller-capacity cards to begin with. Many large capacity cards also pack on more bits per cell, so it seems the lowest error rates are in the midsized cards. Granted, the errors are corrected via error-correction checks but the point is that this is actually counter-intuitive.

So, I'm all for established practices but I'm just wondering what they've been established on: hard data? Or anecdotes/fear?
 
I shoot with 32g cards, they have more than enough space and are cheap enough to have several on hand.
However I do sometimes archive onto these cards when my client wants the raw footage. Having a 128g card to do this would be nice.
Also, maybe I'm lucky but I have never had a problem with any of my SanDisk Ultra C10 32g, been using them for 6-7 months with no problem, I have and love my P2 cards and I'm sure they are more stable, but I'm sensing the SD card fear factor might be over hyped?
 
There's a ton of contradicting information out there and it's pointless to even ask since everyone is going to have a different opinion. Here are some things to consider:

-most camera ops/DP's are not really into IT/computers
-the old school guys apply the same principles they used for film/betacam to digital recording
-people can't grasp the fact that digital video is simply a collection of computer files
-everyone has their own voodoo -- some use a card only once, some only use Sandisk, some take images of cards, some wear their underwear on their heads prior to a shoot for good luck....you're going to get a wide range of completely useless things that people do to get that false sense of security.

When shooting video, basic logic comes into play and this logic has nothing to do with recording media. In IT and computers there is something called RPO and RTO -- Recovery TIME Objective and Recovery POINT Objective...The first one has to do with: "If you have a failure, what POINT in time are you satisfied with being able to restore to?"

...and the second "how much DOWNTIME are you prepared to accept in case of a failure?"

As your RPO and RTO gets lower and lower, you end up spending more money.

At the end of the day, it comes down to your risk tolerance. In any mission-critical computer system, you have points of failure. In your case, the media is a single point of failure but so is the camera. It just so happens that, statistically, the camera is less likely to fail than the card.

The moral of the story is that it depends on what you're recording. If you're recording to four 8GB cards for "risk mitigation" but the footage you're recording is mission critical (like a wedding or some other event that cannot be repeated), then the failure of a single card will mean a loss of a significant amount of data and you're screwed.

If you're recording talking heads and your card dies, then sure, I guess you can re-shoot but it's a pain.

In general terms, you should always try to record to more than one medium at the same time. Whether it's SD+Ninja, or SD+SD (such as the AC130 or AC160), you ideally want two copies of your data. The idea that 4x 8GB cards or 2x16GB cards is somehow magically more error-free and reliable than a single 32GB card boggles my mind.

There's nothing wrong with using a 128GB card...put two of them into a AC130 or 160 and you have 12hours of backed up footage. For some reason the AF100 doesn't do backup recording so I would put a Ninja on the back end.

For me, carrying 4x8GB cards instead of a single 32GB card makes no sense....in fact, you increase the possibility of failure because now you have 4 points of failure. Sure, you'd only lose a portion of your video, but 40 minutes is a significant portion.

I don't care about 11min film magazines from the 80's. I don't care about betacam or any of the other primitive technologies that existed before I got into video.....we're almost in 2012....times change constantly and you need to adapt. There's always a drawback to any decision -- most often this drawback in financial...spending $1000 on a ninja solves your problem, but do you have the $1000 for a ninja?

So now you have a ninja..great...do you have a backup cam in case your cam starts smoking in the middle of some wedding vows?

You have to decide what your risk tolerance is, and the lower your tolerance, the more cash you're going to have to spend to mitigate the risk.


Another way to look at it is to ask yourself this question: If my card fails, what will be my cost to re-shoot that footage? Is it even possible to re-shoot?

You will find that usually the cost (time + money) of re-shooting is significantly higher than spending a few grand ahead of time for a real-time backup solution.

You'd be surprised how many corporate executives don't understand this.
 
Last edited:
You seem to think that the discussion is only about RISK and FAILURE.

Part of the advice is about management of the data once you have finished shooting.

Shoot two or three different projects on one card means all the data is in one AVCHD Folder Structure. (which you cant lose or dissasemble for full functioning)

If you are ok with that when you get to the NLE and need to create three projects all from one source folder...thats fine .

Means you can have three clients work all in the same archive. Lose the archive and you have 3 pissed clients.

Means you have much larger and longer ingest times or transfer times.

Means you are shooting on a card that has other clients work on it. Good opportinty to piss 3 clients off again.

You can not partially download the files. You have to download / ingest all the files every time. So if you do it every night after the shoot, you will be replicating the storage incrementally in your system. ie you will have three folders or one that you are overwriting.

Means one shoot has filenames from say 0000.mts to 0105.mts , next shoot starts numbering from there and the next follows on.
Time code will start on shoot 1 and continue thru shoot 2, 3 ...etc...(unless you change it)

Good luck.

When do you get the camera anyway?
 
Sounds like we have different workflows, too, Shooter. I ingest everything once, transcode all my files to ProRes 422 and send the transcoded files to seperate folders on my system. The card image backups on my server, my server backs up on external HDDs. My clients don't get the SD card or source files. They don't even know what i record to. They get the final video. Besides. all their footage is on the same computer/network so who cares if it's on the same SD card. I am a homebrew shop. Yeah it's longer ingest times and transcode times, but I don't shoot news. I can go watch a movie while its busy. I work my own hours. Transcodes go pretty fast as i have a 12-core MacPro.

I get the camera and all accessories Thursday! Can't wait.

Decided to keep my HVX200 as well, just makes sense as I'll never get the value out of it and could use a second camera as it is.
 
I edit native files ( AVCHD or Prores ex Samurai) so thats different for a start.

Be interested to see how you get on with AVCHD File structure once you start using it in your workflow.
 
Shooter said:
You seem to think that the discussion is only about RISK and FAILURE.

Part of the advice is about management of the data once you have finished shooting.
Click to expand...

Valid points, however the original poster's question was more geared toward card failure instead of data management. If data management is concerned, then of course other factors need to be taken into consideration and different people employ different workflows.
 
filmguy123 said:
Sounds like we have different workflows, too, Shooter. I ingest everything once, transcode all my files to ProRes 422 and send the transcoded files to seperate folders on my system. The card image backups on my server, my server backs up on external HDDs. My clients don't get the SD card or source files. They don't even know what i record to. They get the final video. Besides. all their footage is on the same computer/network so who cares if it's on the same SD card. I am a homebrew shop. Yeah it's longer ingest times and transcode times, but I don't shoot news. I can go watch a movie while its busy. I work my own hours. Transcodes go pretty fast as i have a 12-core MacPro.

I get the camera and all accessories Thursday! Can't wait.

Decided to keep my HVX200 as well, just makes sense as I'll never get the value out of it and could use a second camera as it is.
Click to expand...

For the record, for larger projects I transcode my footage to DNxHD and, since DNxHD is a single mov file much like Prores, there's no need for a file structure. I usually name the files something along the lines of "Cam1_Card1_Clip1.mov", "Cam1_Card1_Clip2.mov", etc etc...I then use this for editing while backing up my original AVCHD footage to two locations.

I have a fast computer but for larger projects AVCHD is still nowhere near as smooth as editing with DNxHD/Prores.
 
Phil1076 said:
I have a fast computer but for larger projects AVCHD is still nowhere near as smooth as editing with DNxHD/Prores.
Click to expand...

For sure. I have a MacPro 12-core, 16gb RAM, ATI 5800 gpu... and even in FCP X (which I'm not a fan of) I would still edit in ProRes rather than h.264 or any mpeg4 codecs
 
My Windows System edits anything natively and smoothly without rendering at all. (ie. avio, prores, avchd, mpeg, QT mov DNxHD,

Even in the same timeline.

:laugh:
 
Shooter said:
My Windows System edits anything natively and smoothly without rendering at all. (ie. avio, prores, avchd, mpeg, QT mov DNxHD,

Even in the same timeline.

:laugh:
Click to expand...

So does my Mac Pro. :) Adobe Premiere Pro CS5.5 is very nice about that. The 32GB of RAM and high end video card that I have in the beast doesn't hurt either.
 
You must log in or register to reply here.
Back
Top