Personal View site logo
Atomos Ninja or any external recorder?
  • Hi everyone,

    I have been scouring the site for relevant information, but am gathering it in drips and draps, so I thought I might as well start a new topic.

    I am thinking of getting an external recorder for my GH2 (i know, I am late to the game), but have read that some recorders don't work with GH2. But first, allow me to establish some background info I know (but stand corrected). GH2's output wasn't clean till Panasonic upgraded firmware to v1.1. 25p is now clean, and Atoms Ninja website has published instructions on how to integrate GH2 into the recording workflow. Testers have reported clean footage, but some testers claim that red channel is still jaggy (the extent of which i have not seen). Then I read that Black Magic Hyperdeck does not recognise GH2 output as true HDMI; some testers have reported unusable footage(?). In any case, whether 25p or not, the recorded output has been reported to be "interlaced" on the Ninja.

    Now I have also read that GH2 with hacked firmware v1.1 yields aberrations, such as colour shifts, inaccurate gamma, ghosting(?) when recording onto the Ninja. Could someone pl verify this? Does hacked firmware not work?
    I am on the brink of buying a Ninja, but thought I should look before I leap. Appreciate all yr kind views.
    Cheers

  • 62 Replies sorted by
  • The GH2 does not output sound over HDMI

  • Hallo

    If someone helps me with information,Today I tried a GH2 with Ninja Blade, just HDMI connected do not have sound on ninja blade. What do you recommend color de profile to register with atomos blade?hak??

    thanks

  • I shoot with the GH2 and BMCC, and to be honest, I've found the Ninja or any external recording device to be useful only with machines like the BMCC, which outputs native 10bit 422 signals. For this reason, I wouldn't even use the ninja with expensive cameras like the C100, which Canon has doggedly kept as an 8 bit anomaly in today's rapidly evolving market. I have found no discernible advantage using the ninja with the GH2. In fact, the clips do get quite noisy when you "up res" 4:2:0 to 4:2:2 (see my explanation above)

  • I just got hold of a discount ninja-2. Shooting HBR 25p the reds on the ninja do look more jagged, and I haven't spotted any cadence issues (but don't in truth understand what I'd be looking for). Even without any patches the GH2's own recording looks very slightly better to me. The advantage of the ninja-2 (after the GH2's hdmi output limitations) seems to be as a monitor (peaking etc), sound recorder (2ch, manual levels) and video backup (very reassuring). The ninja's recording seems 99% as good as the GH2's onboard recording but because of it's instantly editable prores movs with ready synced audio it's quicker to get into edit for fast turnaround and lower budget jobs.

  • If you treat PsF as progressive in your editor, there should be no difference. Technically there are fields, but the content is from the same point in time.

  • @peternap

    Could it be due to the use of Psf (progressive segmented frame)? GH2 records in psf, whereas the AC160 records true progressive frames

  • The DNx HD is what should make the difference but I've found it doesn't with the GH2 or 3...but it does make a big difference with my AC 160, why I don't know because it puts out 8 bit 4:2:0 also....but there s a big improvement in post.

  • Absolutely! With the current best hacks, a Ninja makes no sense IQ wise with the GH2 (other than a backup, if needed).

  • @Gavvo888

    Anything that comes out of the GH2 is 8 bit 4:2:0, whether it is written on the card (which then undergoes further compression) or straight via HDMI. Recording HDMI signal out on an external recorder like the Ninja will not give you discernible improvement over footage written on SD cards. It's true the recorder will transcode the footage on the fly to 10bit 4:2:2, but technically, the upsampling will introduce "noise" signals. Imagine the 8 bit as a small pail, and the 10 bit as a big one. Whatever the 8bit pail contains will always be capped at a certain limit, even if you fill the pail to its brim. Now pour that content into a 10bit pail, and you'll probably get a half pail full. How then to fill up the rest of the space in the 10 bit pail? There is no more info to squeeze out of the 8 bit except to fill in the blanks with noise.

  • Fixed, the hack doesn't change it.

  • @kazuo Sorry to be coming to the party so late but....I read this entire thread and am still unclear. Does the HDMI out send a higher bitrate signal if the GH2 has hacked hbr firmware? Is the HDMI output signal firmware dependent? Or is it fixed? Do you have any updates?? Thanks

  • @mo7vies

    Driftwood's settings are brilliant for 24p. He's openly acknowledged that encoding for 25p 50i is slightly different from 24p, and he's not seen anything (including his own work) that comes remotely close to the kind of sharpness that is 24p encoding. For those like myself who live in PAL land, external recording via HDMI seems to be the best way to get quality that is equivalent to the highest 24p hacked settings.

  • IMHO, for all practical purposes, Driftwood's high-bitrate hacks, and RalphB's low-rate Sanity do cover all the bases. And with GH2 selling for less than $500 today, just go buy a second cam, and shoot something cool! :)) http://goo.gl/74DIh - I doubt there will be any lower price on this kit.

  • Yes, the Ninja is good for 1080i50.

  • Sorry for digging out this thread but I wonder if ninja or ninja 2 could make sense for shooting in 1080i50 or 1080i60. I know, these modes are not so popular because of their worse quality. Anyway, this is the format of my work so I did a short test: I connected the HDMI-out of my GH2 with the HDMI-in of my Matrox MXO2 and recorded the 1080i50 signal with the Matrox I-frame codec at 200MBit.

    After comparing it, I think the i50 quality is significant better and on the same level than 24p (interlaced of course). I couldn't see any differences in red channel other than looking cleaner with sharper borders (filmed a testchart) compared to recorded AVCHD stuff (used FlowMotion 2). But I'm not shure if this was aleady corrected by my Matrox videohardware.

    I have no Ninja for further tests, so did anybody made some experiences with Ninja and the interlaced video modes?

  • @goanna @kavadni

    I recently purchased an OEM Liteon LZT-256M3S 256GB SATAIII SSD for $169.00 AUD from a Melbourne based retailer.

    Read up to 515MB/s, Write up to 440MB/s

    I'm guessing you have the same drive?

    Windows 7 now boots in 45 secs on a 4 1/2 year old PC.

    SSD's are falling rapidly in price. I would expect 1TB SSD's be under $300 within 6 to 12 months.

  • @kavadni

    +1 re SSD prices. I have just cloned my system drive onto a lightning-fast 250GB SSD cost $AUS170.00

  • I have to dispute that SSD are expensive .. even now

    120G SSD = $200

    64G SDXC = $150

  • @kazuo Its alright, VCF is the right filetype (sharing this name with VCcard). Open Virtualdub, File --> Load Processing Settings. Then everything is preconfigured to do the job, load 24p AVI, hit F7, render it to disk. Audio will be written uncompressed.

  • FWIW I love the Ninja2. It adds a whole layer of professionalism to my rig because of the great display and professional audio. I use it with a D800E but have not tried it with the GH2.

  • @kavadni

    Thanks dude, I would love to have a Hyperdeck but SSD is expensive.

  • Hyperdeck gets just over 12 minutes on a 120G drive, and DNxHD gets just over and hour and 15 minutes on the same drive. I haven't used ProRes.

  • @Meierhans

    I tried downloading the timestretch file from the post you directed me, but all I got was a VCcard, not a VDF file that i can place in the "Plugin" folder in Virtualdub. Could you kindly repost that time stretch file for me to test. Cheers