Personal View site logo
Cluster X: Series 1 - moon - ЅріzZ - nebula - drewnet - Slipstream
  • 1008 Replies sorted by
  • Hello. Headed up to Alaska for a two week shoot. Was going to use the GH2 for inserts and specialty shots (C300 is the A-Cam). Network requires 60i acquisition. Do you have a suggestion for a preferred Cluster X for shooting 1080i? Thanks!

  • I just finished testing Cluster X Nebula 6b in NTSC mode using Panasonic 14-140 lens with both my outdoor for detail test and indoor for spanning test. All recording modes (24H, HBR, FSH, SH) recorded with no problems including full zooming and iA mode in outdoor sunny detail scenes using Sandisk 64GB 95mbs card. Spanning worked for 24H, HBR and FSH. HBR and FSH even spanned using Sandisk 32GB 30mb/s card. However, SH stopped at 4GB mark after recording little over 15 minutes with both 64GB 95mb/s and the 32GB card. The quality is very good with high frame sizes in Streamparser.

  • Here is another Shot with Moon at 0 -2 0 -2 on standard, and using Voightlander 17.5mm lens. This time I mixed some footage I had from a couple months ago from a monorail, also using the 17.5mm lens, but it was shot with Cake not Moon.

  • Is Moon Trial 4 sharper than Trial 5?

  • @emweb

    Definitely yes, t4 is the sharpest one. Although i personally prefer the magical filmic look of 'moon' t3 and t5.

  • So far Trial 4 and the creamy look of my voigtlander 17.5 makes a great combo. T5 is also great.

  • I'm using a 64GB Sandisk Extreme class 10 card. Used Moon. Here is what I've found so far.

    24H 95Mbps and spanned for 2hrs until the battery died.
    24H 100Mbps does not span. It'll record for 8 seconds then freeze. 24H 151Mbps does not span. It'll record for 8 seconds then freeze.

    @Driftwood When is the next version of Moon coming out?

  • @T1000 Explain. Don't understand your method of testing? Which Trial etc..

  • @vicharris

    Thanks again for your help but I was always under the assumption that we transcoded files to broadcast range instead of full range in 5DtoRGB. Am I wrong in this thinking?

    Not necessarily "wrong". It depends what you want to achieve. Since you edit in FCP you certainly prefer to edit with fullswing footage. 5DtoRGB with "Broadcast Range" enabled remaps the GH2's luminance levels from studio swing ("Broadcast" Levels) to fullswing ("Computer" Levels) ... which is want you actually want.

    But ... I don't know whether or not 5DtoRGB renders this luminance stretch in 8 bit or in higher bit precision (10 bit or even 16 bit). I assume 5DtoRGB's processing is 8bit - but, again, I am not sure. I for one prefer to preserve the original luminance levels of MTS files when I transcode to MOV files. To remap the luminance levels to fullswing I do apply a curve in the editing software and set my redner settings to 16bit. This way I make sure that the transcoding does not introduce unwanted image degradation... and remapping in 16bit does not introduce image degradation either.

    But this is just a general note. As posted above I've transcoded your clip through 5DtoRGB with "Broadcast Range" enabled and the clip did not show obvious banding or posterization. So "Broadcast Range" seems to work just fine (at least in this case).

    I transcoded a few files to full (...) It also flattens out the footage which is what I assume you meant by my files looking aver saturated after I sent em through 5D.

    Your clip transcoded through 5DtoRGB (with "Broadcast Range") looks less saturated on my machine. This is why I assume your issue is not really related to 5DtoRGB but maybe to your import- or project settings in FCP ... or something else I can't figure out from a distance.

  • @Driftwood

    I'm using Moon Trial 5. I'm testing out the bitrate settings out under the 'End Users Patches'. First test in Ptools under AVCHD Compression I set the 24H bitrate to 95mbps. I set 24L, FSH/SH, and FH/H to the suggested value. For Auto Quantiizer in both modes I set to '2'. In 'AVCHD GOP Length' for all settings I set it to the highest number. This is all recorded in low light. Then recorded for 2 hrs til the battery died.

    2nd test (100mbps) and 3rd test (151mbps) I did the same thing but for some reason the GH2 kept freezing. I took the battery out and put it back in and it still did the same thing.

    I'm in my forth test now with Moon Trial 5 24H set at 130mbps. Same settings as above except the bitrate. Been recording for 27minutes. So far it hasn't froze on me. Any reason why it'd be freezing in 100 and 151mbps? Hopefully I'm doing things right? If not any advice would be great!


  • @towi

    In your opinion, what the best methods for transcoding avc, that workflow uses to get 10/16 bit, Windows/Osx

  • @T1000 I believe these patches are already opimized for maximum quality. If driftwood could squeeze out another byte in bitrate without risking stability (on Sandisk 64gb 95mb/s card..) he would do as I know him. He has been tweaking settings down to the matrix for two years now, I doubt you will find something new he did not already took into account.

    The only two things that really make sense to change yourself is lower bitrate settings (for longer recording tines when quality does not matter) and to change MJPEG FPS to 2 FPS for space saving on timelapse.

  • @maxmizer

    I think we're getting off topic. Since "vicharris" suspected his banding issues are caused by "Moon" I've looked into the file and found there must be something wrong with transcoding or the workflow… but in any case the issue is not related to "Moon". However, to adress your question: as outlined above my advise is to use a tool for transcoding that does not remap the luminance levels and first do the remapping later in the editing software in 16bit mode. I am on a Mac, so I can't comment on Windows tools. But I assume 5DtoRGB with "Full Range" mode enabled should work just fine (especially with DNxHD codecs).

  • @T1000 Its not as simple as just changing bitrate compression! The reason moon (and other settings) bitrates are set as they are is because theyve been massively tested for stability, quality assessment, quantisation, buffer stability, death charts etc.. etc... - they dont just happen! ;-) Prepare yourself for months of testing = incrementaly changing a single parameter at a time (the combinations are endless). Good luck!


    Driftwood Cluster X Series 'Slipstream #2 (Rev A) A combination of Spizz 3 GOP, Nebula 6 GOP and DREWnet 12 GOP technology.


    • 1080p24H 88Mbps GOP3

    • FSH1080i50/HBR25p 58Mbps GOP3

    • 720p50 62Mbps GOP6

    • FSH1080i60/HBR30p 44Mbps GOP12

    • 720p60 52Mbps GOP12

    • Settings should span.

    Driftwood Cluster X Series 'Slipstream #1 (Rev B)


    • 72Mbps max IPB frame 1080p24

    • 52Mbps max IPB frame FSH/HBR/720p modes

    • Span tested

    • upto Q12 low quantisation

    • iA mode support

    All settings SPAN (SH/FSH/H) according to my tests (Ive had spanning on 720p60SH at 5 files with a datarate at 50Mbps which at 12 GOP is the equivalent of nearly 100Mbps low GOP). This is a great setting which appears to pass spanning, iA modes and death charts in all REC modes. Its a mix of 'Nebula' meets 'DREWnet' and is the first in a series of combi settings that attempts to offer decent IQ together with perfect timing and good bitrates to accomodate spanning.

    Give it a try and let me know how you get on. See page 1 for details about these new settings & to download Slipstream #1 (Rev B).

    The attached graphic below shows three and a bit file spans on 720p60 (running at avg 51Mbps) totalling 40 minutes of hi detail - hi quality footage off a GH2 with 14-140mm lens.

    Slipstream ver 1b - 720p60 span on SH.png
    609 x 274 - 15K
  • In general, should any of those settings be better at dealing with the HBR 25p banding/interlacing artifacts than the others? Right now I'm on drewnet trial 8 and the artifacts still bug me more than I'd like. Damn you PAL! :)

  • @sedlak Give the new settings a trial. HBR 25p as you correctly note suffers from interlacing and noticably more banding. You may just simply want to try shooting in 24fps and transcode to 25p as there's only one more additional frame will be created - hardly noticeable unless you go looking for it. The quality will be vastly superior than HBR mode will ever be.

  • @ Driftwood

    I'm obviously new to this hacking/testing for the GH2. I'm sharing what I've experienced so far but I didn't say that it was easy as adjusting the bitrate. I'm adjusting the bitrate to compare it to stock settings. Thanks for the hacks!

    @ Meierhans 'The only two things that really make sense to change yourself is lower bitrate settings (for longer recording tines when quality does not matter) and to change MJPEG FPS to 2 FPS for space saving on timelapse.'

    I'll be sure to remember this! Thanks!

  • Hi there,

    Is timelapse MJPEG 2 FPS mode already included in these patches, or it must be done manual?

    thank you in advance

  • Its set to 30 FPS (=default). Timelapse is a special case and not everybody wants to sacrifice MJPEG mode for it.

  • ok, how do i sacrifice it :) ?

  • @driftwood Did I get this correct - Slipstream #1 (Rev B) uses a higher bitrate in FSH 1080i50 than Slipstream #2 (Rev A) while using 12 GOPs? Wouldn't it make more sense to use the higher bitrates for the 3-GOP-settings?

  • @spreeni Oops slight error in the writeup. Its been corrected. There's over 10Mbps more bitrate required in 3 GOP 1080i50 in Slipstream 2's 1080i50 over Slipstream 1 (rev B) 6 GOP variation. As far as NTSC 720p60 and FSH1080i60/HBR30p 12 GOP is used for both settings.

  • @driftwood Slipstream #1(rev. b) and Slipstream #2 (rev. a) have the same values: Video Bitrate 24H=72000000 Video Bitrate 24L=50000000 Video Bitrate FSH/SH=68000000 Video Bitrate FH/H=45000000

    It's right?

This topic is closed.
← All Discussions
Start New Topic

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In with Google Sign In with OpenID

Sign In Register as New User

Tags in Topic

Top Posters