Cluster X: Series 1 - moon - ЅріzZ - nebula - drewnet - Slipstream

inqb8tr

@WarLightUK 12 minutes 1080i50/60 were just one span?

WarLightUK

@inqb8tr It was 1080i50... It was a single file and the file was about 2.5GB. So the recording stopped even before reaching the spanning point...

driftwood

@jclmedia @WarLightUK Try Spizz 3 GOP or moon L setting / HBR/H mode and let me know if it spans. Remember these are still trial settings. These are high quality 'FSH/SH' settings trying to get the most outta the codec - I'm not surprised spanning may be a problem - read my earlier posts a few pages back regarding spanning of moon. Spanning problems exist on all hi-bitrate settings not just Cluster. Its never been thoroughly looked at as to why it happens but it could well be over a certain amount of data rate some cards find it impossible to initiate the next file write or simply gives up before it reaches the first 4Gb.

L or FH/H settings are set at a lower bitrate to try and aid this problem as well as offer longer recording times at a reduced data rate.

Its one of those problems I must dedicate some time to. If only I had more time...! ;-)

Indeed, I would love just someone to offer their support and check thru all the Cluster X settings for spanning. Instead, people just fucking moan as though the setting was designed to piss them off! Do they really think this was the intention?

WarLightUK

@driftwood Appreciate all the hard work Nick. I'll try to test Cluster X settings for spanning at the weekend. Well, no promises that I'll manage to go through all but I'll give it a go. If only I had more time too...! ;-)

driftwood

@WarLightUK Good sport. Any help you need pm me.

vicharris

@driftwood For spanning tests do you prefer a certain lens or does it matter? Also, autofocus going or anything active? 14-140 all auto to push it to it's limits?

driftwood

Pany lenses 14-140 is a good start. Try normal style shooting - dont bother with iA modes. ta @vicharris

vicharris

@driftwood. Roger that. I have two shoots this weekend so I'll put some stuff through the ringers.

MirrorMan

Short clip testing Cluster X DREWnet Trial 5 and the Rokkor 135mm I just got. I shot this through an open window and it was probably -8C outside, hence the wavy heat effect. But I think this is a good test for the codec and it seemed to handle it well. I also wanted to see the difference between Smooth -2,-2,-2,-2 and Standard 0,0,0,0 which is slight (without grading). Fantastic work Nick.

matt_gh2

@MirrorMan That's some good looking footage. Really liked the 1st 2 shots - very cinematic.

jdenkevitz

I successfully recorded a 60+ minute clip using Moon Trial 5 with a sandisk 95/ 64gb card. 24P high mode. Panasonic 14mm lens. I just let it record my bands rehearsal (with a Zoom q3hd providing audio). Not a ton of motion to stress it out, but it spanned fine. The only 'issue' was that the files did not play (or appear) on the card after restarting the camera (but they recorded just fine and were still on the card). Amazing patch! Thank you for your hard work!

driftwood

Always record a little file after a long span set BEFORE switching off camera/attempting playback on 64gb cards.

peternap

@Driftwood I just loaded moon 5 and decided to see how or if it would span. I used a Nikkor 28mm 2.8 lens, 24P, 50FPS, Standard -2, ISO 400.....64 gig 95mb Sandisk.

It ran for 8 minutes but did not span. It cut out a little over 3 gigs.

driftwood

Look back a few pages. Moon 24p spanned for me in 7 or so 4gb files on the same card as yours. What was your avg datarate @peternap ?

brudney

Driftwood, I was wondering if 6gops settings are final or are you going to tweak them some more? Is it possible to lower quantisation even more? Thank you!

kritzresn

Quite a lot of confusion over the spanning/stability issue even with 95mb/sec cards when i went trough this thread.

To stir up a usefull discussion and provide a theory as to why we have people with failed spans i came up with the following: (let me know if this should be a seperate thread).

Let's for a second assume everyone has sandisk 95mb/sec cards (most even the 64GB one) we still see very different results on people who can span with almost identical setups. I just tested my SD cards in the internal card reader in my mac mini (which is connected over sata, and thus does not have usb limitations) and it shows very different results with the 3 cards i have.. i have 3 95mb/sec cards, 1x 32GB and 2x 64GB, when i use the blackmagic disk speed test (good ref. app) the 32GB reaches cont. 61mb/sec write speed, one of the 64GB reaches a cont. 36mb/sec write speed, and the other 64GB reaches 43mb/sec cont. write speed (tested this with 1GB,2GB and 4GB file writing, results are almost identical, formatted clean before testing, also tested under windows with atto disk speed test and results were same in range.. like max. 2 to 3mb/sec off).

What we see is large differences in actual cont. writing speeds of the cards 32GB: 61mb/sec 64GB1: 36mb/sec 64GB2: 43mb/sec

All seem fast enough to handle the GH2 data rate in cont' mode, but they usually take some ms to get up to full speed, in this moment it is probably where we can differentiate the good cards from the better cards and where the really high bitrate settings may sometimes fail.

As i was surprised about the large difference in cont. write speeds with the cards i contacted sandisk support and after validating all my cards were real sandisk cards they could not mention anything more usefull that the differences are within normal range for variety in writing performance between the same series of cards.

I will do some spanning tests with all 3 cards and see if i can find any relation from the cont. write speed to the fail/success rate of spanning. (bodysettings, patch, lens identical to keep it consistent)

Could some of you out there use the blackmagic or atto app and test your cont' write speeds of 95mb/sec marketed sandisk cards, maybe then we could get a general idea of how varied the actual performance of these identically marketed cards are.

Funny fact, when i take some of the older sandisk extreme 30mb/sec or 45mb/sec cards i have for my NEX camera, they are stuck between around 10 to 20mb/sec cont. average, the 95mb/sec cards are def way better quality overall!

orgie18

Driftwood Cluster X "moon trial 01" canon fd 50mm 1,4 and 85mm 1,8, lumix pancake 20mm 1,7.

rapsoe

@kritzresn very interesting results. thanks for sharing. did you do a low-level format of the cards with the SD Formatter just before testing? —https://www.sdcard.org/downloads/formatter_3/

If not, that might explain, I'm not sure if in part or completely, the discrepancies in your results.

driftwood

Also post general spanning stuff here;-

http://www.personal-view.com/talks/discussion/1047/gh2-avchd-spanning-/p1

Anything directly related to spanning on Cluster X settings should be posted on this Cluster X thread.

herik

Hi, I have a SanDisk 16gb Extreme Pro 95MB/s. I use Moon hack, the quality is amazing but I have some speed write error when i use the HBR mode and no error in 24p. why is the mode with the lower bitrate that make me some trouble ???

luxis

Low rider(predecessor of moon?) spanned for me in dark scene in 1600(or was it 3200!?) with AF on! with 20mm f1.7 panny lens in 24H(performance situation, changes of light,camera fixed on a tripod) for a full 16gb/95mbs Extreme pro card. Will do tests on the same card for Moon trial 5.

@kritzresn Good info there! Thanks! Will you use Moon trial 5 ?

herik

I made some test with codec chart and I confirm what I said two posts before: with moon, I have no problem in 24p (bitrate 148000Kbits) but I have a lot of write speed error in HBR or 1080i with a lower bitrate. I used sandisk 95MB and 45MB/s, same issue. I format my SD with SDFormatter. I don't understand what happening.

driftwood

Please be a bit more detailed when you mention HBR/1080i/720p modes. Is it 25HBR/50fps (PAL) setting or NTSC (HBR30p/FSH60) you're having problems with?

GravitateMediaGroup

@orgie18 the clip of the sun with the birds flying across was awesome and so was the slow mo inline clip across the sky. twixtor im guessing? it was so good it almost looks fake lol

dieter123

Hello I have a question about GH2 cluster X DREWnet Is it possible to set a smaller bitrate mode in 24L ? How high should the bitrate be for stock ~24H ? i need this for timelapse.

thx

Howdy, Stranger!

Categories

Tags in Topic

Top Posters