PDA

View Full Version : 1hr take: class 4 cards OK, h4 audio sync drifts



J Davis
06-23-2009, 07:35 PM
I have been running tests just because I like to know my gear before I start shooting.

SDHC Sandisk class 4 Video HD

First thing I wanted to test was my SDHC cards and these are left over from my HF100 which I have now sold. The cards I have are 8 gig Sandisk Video HD and they are rated class 4.
Longest test I did with HF100 was 1hr30m shot at 1080p AVCHD 17mbps. Used log & transfer to FCP ... no problems.

Now for the GH1 .. same card and due to less disk space than usual on my mac I cut it short at 1hr 5min. GH1 recorded fine, no problems at 1080p AVCHD 17mbps. Used the same log & transfer to take it into FCP ... took a long time .. and there was no problems. Of note - the total size on the sdhc card for this test was 4.6gig and once it was transcoded to prores it was a 67gig .mov file

So as far as my tests go, I think its fine to use the Sandisk Video HD class 4 speed rating.

Here's a link to the 8 gig card I used, costs about $30 LINK (http://www.bhphotovideo.com/c/product/541413-REG/SanDisk_SDSDHV_008GR_8GB_Video_HD_SDHC.html)
and they have the same card with 4gig, 8, 16 etc cheapest is $20 LINK (http://www.bhphotovideo.com/bnh/controller/home?O=sort&A=search&Q=&sortDrop=Price%3A+Low+to+High&bl=&atl=&pn=1&st=search&mnp=0.0&mxp=0.0&sv=sandisk+sdhc+video+hd&bhs=t&shs=sandisk+sdhc+video+hd&ac=&fi=all&pn=1&ci=0&cmpsrch=&cltp=&clsgr=)

AUDIO SYNC with H4

Well I figured if I'm going to transcode such a huge file I may as well do an audio sync test at the same time. I am using the H4 not the H4n. To save time I did not remove pulldown and I don't beleive this would have made a difference - please tell me if I am wrong. So with the prores file at 29.97 and the timeline at 29.97 the audio was out by 10 frames at the end of the hour.

I think this is good info to know especially as I only own one cam and you never know when someone is gonna want to pay you to shoot a long interview or something.

Isaac_Brody
06-23-2009, 07:37 PM
Thanks for being the guinea pig. It'll be good to save some money on SDHC cards.

Have you tested 60P yet? Wonder if Class 4 can handle that.

e-steve
06-23-2009, 07:47 PM
I saw somewhere on vimeo (I think the Zacuto profile page) where they used the H4n and showed how to sync it with the 5DMkII footage and for some reason (I haven't tried to figure out why yet) they adjusted the audio speed to 99.99%.

Perhaps there is something you need to do that is similar to ensure the audio remains in sync?

J Davis
06-23-2009, 07:54 PM
Thanks for being the guinea pig. It'll be good to save some money on SDHC cards.

Have you tested 60P yet? Wonder if Class 4 can handle that.

Yeh but not a long test. Longest at 60p was 6-8 minutes. Pushed it thru cinetools and dropped into 24p timeline, played beautifully.

Don't mind being the guinea pig as I was planning to run these tests anyway. If the cards from the HF100 hadn't worked I would have shelled out cash for the class 6 (but no point doing that till after testing).

The fact that I can shoot on a $20 4gig card is awesome news to tell a potential client. I think the limit on that card is 45 mins of 1080p.

There are some other tests I did that concern the mounting of the H4 for solo run & gun and in particular the problem of handling noise and lack of a shock mount. For those that haven't read it, the thread for that is here LINK (http://www.dvxuser.com/V6/showthread.php?t=174994)

Isaac_Brody
06-23-2009, 08:05 PM
I still have cards left from my HF10 so this is definitely welcome news to save some bucks.

J Davis
06-23-2009, 08:09 PM
I saw somewhere on vimeo (I think the Zacuto profile page) where they used the H4n and showed how to sync it with the 5DMkII footage and for some reason (I haven't tried to figure out why yet) they adjusted the audio speed to 99.99%.

Perhaps there is something you need to do that is similar to ensure the audio remains in sync?

Yep .. the .wav from the H4 was longer.

I used CTRL click on the tiime line counter to convert to frames and it was 117215 frames long.

10 frames out means the .wav should have been 117205 frames long and if you divide it out it comes to %99.991468 ... blah blah

I could have saved myself all this hassle and just watched your vid
(but I am a mistrustful soul ... I like to learn the hard way! )

Ben_B
06-23-2009, 08:15 PM
Good for left over cards but seems like no reason not to get class six, cost-co sells two 4gb 15megabyte/second cards at $30 for the pair....or 4x2gb.

J Davis
06-23-2009, 08:17 PM
Whatever the card .. its important to test the hell out of it before showing up on a shoot and disappointing a whole lot people.
Me? ... I like brand name 4gig cards ... and a bunch of them to use in rotation.

e-steve
06-23-2009, 08:22 PM
Here's the video I was referring to:
http://www.vimeo.com/zacutousa#4782593

John Caballero
06-23-2009, 08:32 PM
Regarding removing pulldown and synching with H4n audio

Will the audio from the H4N sync well to pull down removed 1080P? Thats one of the few questions I have with this combo.

J Davis
06-23-2009, 08:37 PM
Regarding removing pulldown and synching with H4n audio

Will the audio from the H4N sync well to pull down removed 1080P? Thats one of the few questions I have with this combo.

I believe that the 10 frame error I saw would have been the same before/after pulldown. As far as a difference between H4 and H4n .. that I don't know as I tested with H4.

e-steve said the zacuto vid shows that they saw the same error rate I saw and they used the H4n

EDIT
either way if you are removing pulldown to get true 24p then you are probably doing narrative film type stuff .. right?
in which case it is unlikely you would need to correct for the error as its usually lots of short takes.
I'm doing 24p narrative and I will not be correcting for the error unless its a shot that goes for 5-6 minutes and has dialog near the end

John Caballero
06-23-2009, 09:22 PM
I believe that the 10 frame error I saw would have been the same before/after pulldown. As far as a difference between H4 and H4n .. that I don't know as I tested with H4.

e-steve said the zacuto vid shows that they saw the same error rate I saw and they used the H4n


Thanks.

BTW: I just bought this tripod for when I get the GH1. Its pretty smooth for the pans but two cool things about it is first that you can lower it to 11'' and second that about 5 inches bellow the head you can unscrew it and use that portion up for handheld shots. You can put the camera on top a little sideways in between the handle bars, then press the handle bars against your chest and hold the 5'' of tube with your hand. I think is gonna work pretty good. You can take it from tripod to handheld in a minute. It is also very light but sturdy.

http://www.adorama.com/TFV3H.html?searchinfo=voyager+tripod (http://www.adorama.com/TFV3H.html?searchinfo=voyager+tripod)

Park Edwards
06-24-2009, 01:57 AM
17mbs/8=2.125MB/s which in theory a Class 2 card could do. Class 4 is 2x the speed necessary to record video.

Class 6 cards from A-data run you $9.00 over at newegg

http://www.newegg.com/Product/Product.aspx?Item=N82E16820211067

People spending money on ultra III's or whatever are either really bad at math or don't mind spending money on unnecessary gear.

J Davis
06-24-2009, 03:19 AM
17mbs/8=2.125MB/s which in theory a Class 2 card could do. Class 4 is 2x the speed necessary to record video.

Class 6 cards from A-data run you $9.00 over at newegg

http://www.newegg.com/Product/Product.aspx?Item=N82E16820211067

People spending money on ultra III's or whatever are either really bad at math or don't mind spending money on unnecessary gear.



Actually I was testing out old cards .. I haven't bought new cards for a while and had no idea about those prices.

So are A-data cards a reputable brand?
So is this the card that you use, have tested and recommend?
Or did you just go to newegg and search for all class 6 cards and order by price just to prove a point. LINK (http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&Description=sdhc%20class%206&bop=And&ActiveSearchResult=True&Order=PRICE)

Park Edwards
06-24-2009, 11:30 AM
Actually I was testing out old cards .. I haven't bought new cards for a while and had no idea about those prices.

So are A-data cards a reputable brand?
So is this the card that you use, have tested and recommend?
Or did you just go to newegg and search for all class 6 cards and order by price just to prove a point. LINK (http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&Description=sdhc%20class%206&bop=And&ActiveSearchResult=True&Order=PRICE)

A-data is the memory I use in my PC, kicking 4 gigs of it. Over 2years now?
I had planned on using it on my d90 and will use it with a Gh1.
It wasn't this particular card, but I had one under my wish list and it was only $7, same brand, but they no longer sell it.
Not to prove a point, but to point out to others they don't need to spend money on SanDisk Ultra whatevers.

J Davis
06-24-2009, 08:34 PM
Thanks for being the guinea pig. It'll be good to save some money on SDHC cards.

Have you tested 60P yet? Wonder if Class 4 can handle that.

Shot about about an hour today 720p60 on my old card without a hitch

J Davis
06-24-2009, 08:37 PM
A-data is the memory I use in my PC, kicking 4 gigs of it. Over 2years now?
I had planned on using it on my d90 and will use it with a Gh1.
It wasn't this particular card, but I had one under my wish list and it was only $7, same brand, but they no longer sell it.
Not to prove a point, but to point out to others they don't need to spend money on SanDisk Ultra whatevers.

A challenge for you -
buy a GH1 and the $9 A-data cards that you recommend, run a bunch of 1 hour tests and post the results. That way we can all benefit. Good idea no?

Isaac_Brody
06-24-2009, 08:52 PM
A challenge for you -
buy a GH1 and the $9 A-data cards that you recommend, run a bunch of 1 hour tests and post the results. That way we can all benefit. Good idea no?

Great idea. The cheapo cards are only useful if they've been tested and work fine. Until then I'll stick to the tested Sandisk whatevers.

gmoe
06-24-2009, 09:00 PM
A-data are quite reliable and have used them on the HMC150 and EX1 with K&S adapters. I think I posted that I even accidentally left the adata card in the washing machine and it still worked fine afterwards.

Park Edwards
06-24-2009, 09:26 PM
A challenge for you -
buy a GH1 and the $9 A-data cards that you recommend, run a bunch of 1 hour tests and post the results. That way we can all benefit. Good idea no?

no problemo...i'll even buy a pair of $15.00 basketball shoes and report back how they'll hold up just as good as Nike.

If the cards suck, hell I'll be out a massive 9 bux that would've probably went to some double cheese burgers anyway.

Mike@AF
06-24-2009, 10:00 PM
Is it possible the camera could detect a slower card and lower the bitrate of the recording to compensate. Meaning, it would still work but you could get better quality with a faster card?

J Davis
06-26-2009, 11:50 PM
I posted this in another thread. I think its very relevant to this discussion

http://www.dvxuser.com/V6/showpost.php?p=1674583&postcount=14

Ozpeter
07-31-2009, 03:44 PM
CardTest free testing software - http://sourceforge.net/projects/cardtest/

From this I've noticed that card speeds vary considerably from one end to the other. Always fill a card completely before deciding that it's fast enough.

Eddy Robinson
08-02-2009, 03:17 PM
I used CTRL click on the tiime line counter to convert to frames and it was 117215 frames long. 10 frames out means the .wav should have been 117205 frames long and if you divide it out it comes to %99.991468 ... blah blah

It's probably jitter rather than a pulldown issue (which would have resulted in a much higher discrepancy). a 10 frame difference over an hour is not bad for such cheap equipment...OK, that's 100 parts per million which is high in the pro world but when you consider it as an 00.01% error and which is fixable in < 30 seconds in software, that's quite acceptable.

If you have Adobe Audition, then you can easily make a script to do this and just run it as a batch process on all your audio recorded from the Zoom.

rawdod
08-02-2009, 04:57 PM
I am new to high end cameras, but I am a programmer, and it seems to me, if the camera were able to record video to the card at all, any issue with the sound sync would be the fault of the camera's recording code not the card itself.

J Davis
08-02-2009, 05:09 PM
My gut feeling is the fault is with the H4 zoom not the camera. Until someone does a long sync test with the GH1 vs a DAT or something we won't know.

Ozpeter
08-02-2009, 05:37 PM
I'm not sure why a sync drift between two unsynced devices should be considered surprising. Unless there is a sync connection between them, each will be using their own clocks and those clocks will not be the same. Close, hopefully, but not the same.

FWIW I just put the audio from an Edirol R-44 to 20 mins of GH1 1080 footage (classical music live performance) and I felt no need to adjust the end having aligned the beginning. There would appear to be no gross clock error in the GH1, so you should get no gross drift vs any other device which has no gross clock error. But you are very likely to get a small error, which of course will be more evident with very long recordings, and will be more obvious with some recordings than others depending on visual content.

J Davis
08-02-2009, 05:40 PM
My test was with the sharp 'clap' sound from a slate. Makes a nice audio peak to line up with.
If you get a chance Ozpeter, can you do a long silent test with a slate clap at each end?
If yours lines up then we know its the H4 zoom that is at fault.