Matt Silverman
Forum Replies Created
-
Plasmas still cut off about 10%, so Action Safe is still relevant. If you saw everything you will notice a lot of boom-mikes, blanking, closed-captioning, and all the other crap that gets left in there by folks like us who know it will never be seen.
Action Safe and Title Safe have never been a “regulated” standard. If you are designing a promo for a client like Discovery Channel, they might request staying within title safe, but it isn’t like broadcast safe colors… if I have a spot airing on Discovery I can violate it all I want. I disregard title-safe all the time on TV spots. Sometimes clients like to cheat it. Sometimes I just do it for design. It is purely a recommendation to make sure that the text is going to be readable. You can ignore Action-Safe if you like as well, but I can’t see a logical reason to do so on broadcast spots. We do it all the time designing for LCD’s (which generally do not have any cutoff).
-Matt
-
None of the 3D apps use the QuickTime Video Output Components, so there is no way to mirror output. You could use the extended desktop and drag your viewer to the NTSC… I do this with Maya’s render view sometimes.
-
We have older mac based Media Composers which are 8bit. I don’t know about newer ones.
We usually render out of AE 10bit BM, then play the footage from the FCP timeline and capture live video into the MC 1:1. If we need timecode we lay off from FCP to Digibeta. This is usally for offline only, then we finish in smoke.
-
As far as I know, Digibeta cameras can not shoot progressive, so I am assuming that you have progressive footage transferred to Digibeta, or you shot it with a progressive format like DVCPro50 with an SDI out to the Digibeta.
If you are NTSC then your field order should always be set to Lower… if you are PAL it should be set to Upper. It doesn’t matter if the footage on tape is interlaced or progressive.
When rendering in AE do not field render if you want it to match the look of the progressive video source. If you have motion, then make sure to enable Motion Blur to prevent “strobing”. If you progressive footage is 30 this is easy. If your progressive footage is 24p (or 23.976), you will need to remove pulldown in the Interpret Footage window, comp in a 24/23.976fps comp, then field render and add pulldown in the Render Settings.
-Matt
-
Have you updated the firmware? If yes, I suggest re-seating the card in another slot.
-
Yes. I installed it but did not have time to test the codec. Editing works fine in FCP…
-
It depends if the filter is 8bit or 16bit/float. The FCP 3way CC is float, so your image will then drop to 10bit without downsampling to 8bit. I assume that all of the FCP plug-ins are float (I only really use the CC). If you are running a 3rd party plug-in, there is a good chance that it is 8bit only, especially if it is an AE plug-in (I don’t think 16bit AE plug-ins work in FCP, but I could be wrong.).
As for finishing, you have found the missing piece in the desktop puzzle. There is no good solution like smoke when it comes to serious client supervised finishing. We have a smoke on a quad SGI Tezro for finishing, and a second SGI running flame to help knock out effects shots. The beauty of smoke is that it is a fantastic editor with flame’s effects tools for tracking , cutting mattes, 3D, etc.
We have finished some jobs in FCP, but these are usually design/effectts intensive conforms which are already “finished” in After Effects. Conforming most live action spots is done in smoke. FinalTouch is an amazing tool which has come a long way. We have been really testing it out hard with DVCProHD 24p, since this is a format smoke currently can not deal with in our pipeline. It potentially could be a great finishing tool and compositor. In it’s current state it would have definitely helped you out on your job, since it can track b-splines on secondaries or effects like blurs, and keep chugging away in realtime. It’s definitely a lot more limited than smoke, but far better than FCP.
-Matt Silverman
Phoenix Edit
SF -
To clarify, Digibeta is 10bit 4:2:2 with adaptive compression. Some images get slightly compressed, while others are losslessly compressed. I have done some tests which are lossless after multiple generations and others which incur loss.
also, there is no “digibeta” codec. You should use the BM 10bit codec.
Regarding D1 or D5… D1 is uncompressed but 8bit only (and a dead format), which is why it is equivelant to digibeta… it’s a tradeoff of 10bit for slight compression. D5 standard def is uncompressed 10bit (HD is compressed 10bit).
SDI is just a transport stream… it is like a firewire or scsi connector. It just sends the 1’s and 0’s from the tape to your computer without losing any info. The compression and color space compression happen on the tape.
-Matt
-
We do this all the time without any issues. You should render to the BM10bit codec in AE, then open in FCP using 10bit Uncompressed sequences. Make sure your AE prefs are set to enable 64bit QT for v210. You can then select “Trillions of Colors” in the output module. If you tried using the Apple 10bit codec in AE with Trillions, you will get corrupted frames. This has apparently been fixed with the new ProApps updater, but I haven’t taken the plunge yet (doing this later today). This also fixes the gamma embedding in the Apple codecs, which should allow you to then render directly to the Apple codec in AE.
To clarify, BlackMagic created the first codec for the Kona cards, which used the v210 file format. Apple then created their own codec and also used the v210 file format. These two codecs are interchangeable, though they get confused sometimes. Overall, the BM codec had better color filtering and was “higher quality”, but they eventually killed it in FCP to prevent headaches with the apple codec. Also, by using the Apple codec they could rely on Apple’s RT effects instead of having to write their own support. Hopefully one day it will rise from the dead with a new file format (I don’t need much RT).
-Matt
-
I disagree. Photoshop can only use bi-cubic filtering for best results. Resizer or InstantHD have more advanced filtering algorithms (like Shake) such as Mitchel-Netravelli, Catmol-ROM, Sinc, etc. You definitely can see better results on extreme blow-ups. I suggest a Catmul-Rom or Sinc filter then add an Unsharp mask.
You can also check out Algolith or Teranex…