220 Mbps I-Frame-Only In Next Release

Lance Librandi · October 17th, 2009, 07:15 AM

Hello Dan,
While testing my Sandisk Extreme 60MBs 32GB cards I have noticed that my max record time while recording 100 Mbps I-Frame was 1hour 18 minutes and 100Mbps Long GOP was 1hour 20 minutes. From what I have read here that "1. Long GOP is typically 2 to 2.5 times more efficient than I-Frame Only".

Given that Long Gop is 2 to 2.5 times more efficient than I-Frame Only then is not reflected in record duration?

Mike Schell · October 17th, 2009, 08:41 AM

Quote:

Originally Posted by Lance Librandi

Hello Dan,
While testing my Sandisk Extreme 60MBs 32GB cards I have noticed that my max record time while recording 100 Mbps I-Frame was 1hour 18 minutes and 100Mbps Long GOP was 1hour 20 minutes. From what I have read here that "1. Long GOP is typically 2 to 2.5 times more efficient than I-Frame Only".

Given that Long Gop is 2 to 2.5 times more efficient than I-Frame Only then is not reflected in record duration?

Hi Lance-
You need to compare 50 Mbps Long-GOP to 100 Mbps I-Frame. Both will provide the same level of video quality, but the 50 Mbps rate will give you twice the record time.

Best-

Daniel Symmes · October 17th, 2009, 09:52 AM

Lance -

What format/speed were you shooting at for the 1:18 @ I-frame?

John Mitchell · October 17th, 2009, 11:06 AM

Quote:

Originally Posted by Lance Librandi

Hello Dan,
While testing my Sandisk Extreme 60MBs 32GB cards I have noticed that my max record time while recording 100 Mbps I-Frame was 1hour 18 minutes and 100Mbps Long GOP was 1hour 20 minutes. From what I have read here that "1. Long GOP is typically 2 to 2.5 times more efficient than I-Frame Only".

Given that Long Gop is 2 to 2.5 times more efficient than I-Frame Only then is not reflected in record duration?

Lance - 100Mbs refers to the data rate, not the quality - get it? At the same data rate both codecs will take up the same amount of space on a card. 100Mbs Long GOP should be higher quality than I frame (but not twice as good :)

Lance Librandi · October 17th, 2009, 04:32 PM

Thanks to everyone I am still trying to get my head arround this HD world.

Mike - I now understand that You need to compare 50 Mbps Long-GOP to 100 Mbps I-Frame I will record Both and see if the video quality between 50 Mbps Long-GOP to 100 Mbps I-Frame is noticeable. I have read that if you edit with FCP and use multilayer's on your timeline that it is best to record in I-Frame and not Long-Gop. Would you give me your view. I frequently edit with up to four video layers on my timeline.

Quote:

Originally Posted by Daniel Symmes

Lance -

What format/speed were you shooting at for the 1:18 @ I-frame?

Daniel - I am still trying to find the best solution for my work, I have now settled on 1080/50i and now tweaking the profiles to get the best results.

Daniel Symmes · October 17th, 2009, 04:38 PM

Lance -

Thanks. Just trying to get a handle on capacity.

I shoot 23.976 so our figures won't jive.

Arguing with myself (who else?) about the value of I-frame, which it SEEMS makes editorial happier. Leaning toward 1080 23.98 I-frame.

As for the rate, I'll see what they come up with in the 64GB cards and 100+ Mbps. Not looking for uncompressed (yet), but my Hollywood clients aren't compression fans (film out, efx work, etc.).

Mark Job · October 17th, 2009, 07:41 PM

Quote:

Originally Posted by Lance Librandi

I have read that if you edit with FCP and use multilayer's on your timeline that it is best to record in I-Frame and not Long-Gop.

...Correct. Do you know why ? We all need to take a deep breath when it comes to Long-GOP. Yes, Long GOP is more efficient, but you can't edit it in post properly *because it is a Delivery Format* and not an editing format. Long-GOP was never designed for post production manipulation. I Frame was. The difficulty with I-Frame is you have to roughly double the data rate at encode to equal the quality of Long-GOP. The argument has already been made for high I-Frame data rate settings to be made available on the XDR/Nano, because as good as Long-GOP is (And it's fantastic !) who cares if you can't edit it properly on post and deliver it to the TV screen. (??)

* However, one could set up for capture via HD-SDI and simply play that beatiful 100 Mbps Long-GOP codec into your NLE either uncompressed in your editing project and post and deliver in this manner, or use a lossless compression like Avid's DNxHD 220 X (10 Bit codec). The extra two bits in the color space really comes into play in the color correction stages ;-)

**Play out the Long-GOP via HD-SDI into FCP and use the Pro Res HD 4:2:2 HQ or uncompressed, or use XDCAM HD 4:2:2 (The codec of the XDR and Nano)

Daniel Symmes · October 17th, 2009, 08:19 PM

One has to try to avoid playing the numbers game.

I did a 3D feature film in '05 (first DIGITAL, LIVE ACTION) shot with Sony Z1's.

We shot a test with F900 and Z1 side by side. Low light, high contrast, movement, etc.

Technicolor did a FILM OUT, split screen. We looked at the result, and even the technical types there would go "Ummmm, I think...ummmm, the one on the left is the Z1."

Since then, I've very much espoused the "if the audience can't tell" philosophy.

For most of my "Hollywood" clients, I go uncompressed for post. I'm covered up to 2K.

But I'm finding producers are FINALLY starting to get it that compression CAN work.

What bugs me is the PC/MAC divide. I dislike the warfare angle. Simply, I use PC since there's more choice (and usually lower cost). But the MAC world likes to exclude PC standards, so we're still running in circles.

Long GOP isn't my friend YET. When I have equipment in hand, I'll make that decision.

For now, the cost of cards is coming down, and shooting 160/I-frame is possible and likely even my most strict client SHOULD warm to it with comparison.

True enough I do NOT like the formats of the nanoFLASH. Neither is PC. As they say, "What's up with THAT?"

Recording to QT in a "platform independent" format would be the BEST for all. I do ALL my PC work as QT (or DPX sequences). I exchange QT with MAC systems all the time (as long as you watch out for the gamma flag).

Special flavors of QT/MXF aren't the solution.

Mark Job · October 17th, 2009, 08:53 PM

Hi Daniel:
You wrote: "Special flavors of QT/MXF aren't the solution."

....Yeah, for now they are :-) BTW. Avid runs on the PC or *MAC* and is MXF for both platforms. FCP is QT, because Apple invented QT, FCP and the MAC. MXF is not used by Avid because MXF is PC standard specific. Avid determined the MXF file format to be the best way to arrive at a file format that was resolution independent and could work across the PC/MAC devide.

Rafael Amador · October 18th, 2009, 04:57 AM

Mark wrote:
"Yes, Long GOP is more efficient, but you can't edit it in post properly *because it is a Delivery Format* and not an editing format. Long-GOP was never designed for post production manipulation. I Frame was".
MPEG-2 have been widely used for delivery, but this doesn't means that was designed as a Delivery Format.
GOPs is a new and only digital concept. Intraframe is nothing but the natural way to digitize an analog video signal (made out frames). Nothing related with being or not editable.
NLEs developers haven't update their software fast enough. They have no idea of the lands slide of GOPs that was coming.
Remember just few years ago. You couldn't get the mouse close to an MPEG file in a Mac. The ball would spin for half hour. Software, processors, video cards were no ready to handle this kind of files.
Things have changed and half of the world is acquiring MPEG-2 Long GOPs.

We are in a time were recording 10b Unc is easier than never before. However you heard to talk about 10b Unc less than never before. People simply don't care so much about 10b Unc. Definitively there are other options.

I've been having a look to what is available to record in a field with a quality close or better to the CD machines.
I found those two portable 10b Uncompress recorders:
This first one can even records 4.4.4. pairing two machines.

KG UDR-D100 Portable Uncompressed HD Recorder :: Equipment Sales :: Abel Cine Tech

http://www.mikrom.de/download/MVR100..._Flyer_Web.pdf

Then apart of the Ki Pro, I found this device:
Fast Forward Video
It could be better than the MPEG-2 Long GOPs if the JPEG2000 its records was 10b (supported in JPEG2000) but the info doesn't say so. I guess is only 8b at 100Mbps.
I'm sure that soon will be more recorders to choose from. The trend is that: A lens with a captor and an SDI/HDMI recorder.
Best,
rafael

October 17th, 2009, 07:15 AM	#31
Lance Librandi Major Player Join Date: Aug 2009 Location: South Australia Posts: 374	Hello Dan, While testing my Sandisk Extreme 60MBs 32GB cards I have noticed that my max record time while recording 100 Mbps I-Frame was 1hour 18 minutes and 100Mbps Long GOP was 1hour 20 minutes. From what I have read here that "1. Long GOP is typically 2 to 2.5 times more efficient than I-Frame Only". Given that Long Gop is 2 to 2.5 times more efficient than I-Frame Only then is not reflected in record duration?

October 17th, 2009, 09:52 AM	#33
Daniel Symmes Regular Crew Join Date: Jun 2007 Location: Los Angeles, CA U.S.A. Posts: 146	Lance - What format/speed were you shooting at for the 1:18 @ I-frame?

October 17th, 2009, 04:38 PM	#36
Daniel Symmes Regular Crew Join Date: Jun 2007 Location: Los Angeles, CA U.S.A. Posts: 146	Lance - Thanks. Just trying to get a handle on capacity. I shoot 23.976 so our figures won't jive. Arguing with myself (who else?) about the value of I-frame, which it SEEMS makes editorial happier. Leaning toward 1080 23.98 I-frame. As for the rate, I'll see what they come up with in the 64GB cards and 100+ Mbps. Not looking for uncompressed (yet), but my Hollywood clients aren't compression fans (film out, efx work, etc.).

October 17th, 2009, 08:19 PM	#38
Daniel Symmes Regular Crew Join Date: Jun 2007 Location: Los Angeles, CA U.S.A. Posts: 146	One has to try to avoid playing the numbers game. I did a 3D feature film in '05 (first DIGITAL, LIVE ACTION) shot with Sony Z1's. We shot a test with F900 and Z1 side by side. Low light, high contrast, movement, etc. Technicolor did a FILM OUT, split screen. We looked at the result, and even the technical types there would go "Ummmm, I think...ummmm, the one on the left is the Z1." Since then, I've very much espoused the "if the audience can't tell" philosophy. For most of my "Hollywood" clients, I go uncompressed for post. I'm covered up to 2K. But I'm finding producers are FINALLY starting to get it that compression CAN work. What bugs me is the PC/MAC divide. I dislike the warfare angle. Simply, I use PC since there's more choice (and usually lower cost). But the MAC world likes to exclude PC standards, so we're still running in circles. Long GOP isn't my friend YET. When I have equipment in hand, I'll make that decision. For now, the cost of cards is coming down, and shooting 160/I-frame is possible and likely even my most strict client SHOULD warm to it with comparison. True enough I do NOT like the formats of the nanoFLASH. Neither is PC. As they say, "What's up with THAT?" Recording to QT in a "platform independent" format would be the BEST for all. I do ALL my PC work as QT (or DPX sequences). I exchange QT with MAC systems all the time (as long as you watch out for the gamma flag). Special flavors of QT/MXF aren't the solution.

October 17th, 2009, 08:53 PM	#39
Mark Job Trustee Join Date: Sep 2006 Location: Montreal, Quebec, Canada Posts: 1,138	Hi Daniel: You wrote: "Special flavors of QT/MXF aren't the solution." ....Yeah, for now they are :-) BTW. Avid runs on the PC or MAC and is MXF for both platforms. FCP is QT, because Apple invented QT, FCP and the MAC. MXF is not used by Avid because MXF is PC standard specific. Avid determined the MXF file format to be the best way to arrive at a file format that was resolution independent and could work across the PC/MAC devide.

October 18th, 2009, 04:57 AM	#40
Rafael Amador Major Player Join Date: Jun 2009 Location: Vientiane (Lao PDR) Posts: 349	Mark wrote: "Yes, Long GOP is more efficient, but you can't edit it in post properly because it is a Delivery Format and not an editing format. Long-GOP was never designed for post production manipulation. I Frame was". MPEG-2 have been widely used for delivery, but this doesn't means that was designed as a Delivery Format. GOPs is a new and only digital concept. Intraframe is nothing but the natural way to digitize an analog video signal (made out frames). Nothing related with being or not editable. NLEs developers haven't update their software fast enough. They have no idea of the lands slide of GOPs that was coming. Remember just few years ago. You couldn't get the mouse close to an MPEG file in a Mac. The ball would spin for half hour. Software, processors, video cards were no ready to handle this kind of files. Things have changed and half of the world is acquiring MPEG-2 Long GOPs. We are in a time were recording 10b Unc is easier than never before. However you heard to talk about 10b Unc less than never before. People simply don't care so much about 10b Unc. Definitively there are other options. I've been having a look to what is available to record in a field with a quality close or better to the CD machines. I found those two portable 10b Uncompress recorders: This first one can even records 4.4.4. pairing two machines. KG UDR-D100 Portable Uncompressed HD Recorder :: Equipment Sales :: Abel Cine Tech http://www.mikrom.de/download/MVR100..._Flyer_Web.pdf Then apart of the Ki Pro, I found this device: Fast Forward Video It could be better than the MPEG-2 Long GOPs if the JPEG2000 its records was 10b (supported in JPEG2000) but the info doesn't say so. I guess is only 8b at 100Mbps. I'm sure that soon will be more recorders to choose from. The trend is that: A lens with a captor and an SDI/HDMI recorder. Best, rafael