Looking for SDI capture, Or I believe So.

[CJ Henke III]

I am a FCP editor and I just purchased a Sony HVR-Z1U. Instead of capturing via firewire, which has worked fine in the past - - I would like to get the best out of the footage that I shot. I was guessing that purchasing an SDI capture card would be the next best step, but which one? I'm not concerned with down converting, just more or less interested in capturing the cameras data at it's best.

 

 

My comp specs:

 

Quad Core Mac Pro, 2.66 HGz Intel Xeon

5 GB 667 MHz DDR2 FB-DIMM

4 500 GB Hard Drives.

ATI Radeon X1900

 

 

 

Would something such as this work for my needs?

http://cgi.ebay.com/Blackmagic-Decklink-SD...3A1%7C294%3A200

[Rick Martinez]

I have Blackmagic Multibridge Pro (External Unit) and work wonderful importing from camera and digibeta through SDI to my FCP RAID.

(Link here: http://www.blackmagic-design.com/products/multibridge/)

 

Aja also offer the Kona models and they are good.

(http://www.aja.com/html/products_macintosh_kona.html)

 

Check carefully the specs.

[CJ Henke III]

I have Blackmagic Multibridge Pro (External Unit) and work wonderful importing from camera and digibeta through SDI to my FCP RAID.

(Link here: http://www.blackmagic-design.com/products/multibridge/)

 

Aja also offer the Kona models and they are good.

(http://www.aja.com/html/products_macintosh_kona.html)

 

Check carefully the specs.

 

 

Thanks for the advice.

[Phil Rhodes]

Bear in mind that capturing SDI won't make the footage look any better than it would if you captured firewire, since it's compressed on tape and neither type of capture will make any difference to it. Handling the material as uncompressed thereafter, however, may make a noticeable difference, since it won't be continually recompressed whenever you make a change to it, and if you handle it as 10-bit you may do slightly better again.

 

So, if you wanted to capture it firewire then bump it up to uncompressed, you would get identical results, but that's quite an intensive rendering process. In either case, you would need a disk array capable of dealing with uncompressed HD. If you capture it as 8-bit 4:2:2 1080p, I suspect that your four disks, in RAID-0, under ideal circumstances and not otherwise loaded, would just about keep up, although really you should go for something better.

 

P

[CJ Henke III]

Bear in mind that capturing SDI won't make the footage look any better than it would if you captured firewire, since it's compressed on tape and neither type of capture will make any difference to it. Handling the material as uncompressed thereafter, however, may make a noticeable difference, since it won't be continually recompressed whenever you make a change to it, and if you handle it as 10-bit you may do slightly better again.

 

So, if you wanted to capture it firewire then bump it up to uncompressed, you would get identical results, but that's quite an intensive rendering process. In either case, you would need a disk array capable of dealing with uncompressed HD. If you capture it as 8-bit 4:2:2 1080p, I suspect that your four disks, in RAID-0, under ideal circumstances and not otherwise loaded, would just about keep up, although really you should go for something better.

 

P

 

 

Phil, thanks, that was information that I was looking for. Question then, what do you recommend? My main purposes for this work is to export for the internet (which means nothing), but I do upload TV commercials to comcast via Mpeg2 and I notice (aside from the loss of compression for MPEG2) that most video quality drops before hand. Would a Hardware down conversion piece of equipment (such as Blackmagic or AJA) be a good investment for such means?

 

Thanks Again.

[Phil Rhodes]

What exactly is the characteristic of the quality drop you're seeing? What is the route your material goes through, process by process and device by device?

 

Until I have that information, this will all be speculation, but I suspect that handling it SDI will not make as much difference as better attention to the MPEG-2 encode unless you're doing a lot of generations of recompression on the HDV on top of a lot of fairly extreme colour correction. Sometimes, software will by default use a poor-quality scaling algorithm to produce SD material from HD, and you should check all the available preferences and settings window to ensure that isn't happening. If it isn't, then you will in most cases be using a bilinear or bicubic resize, ideally the latter. These are probably identical to what a hardware downconverter would do, if perhaps not realtime - but that's not a very big deal if you're only doing TVCs.

 

Need more info.

 

P

[CJ Henke III]

What exactly is the characteristic of the quality drop you're seeing? What is the route your material goes through, process by process and device by device?

 

Until I have that information, this will all be speculation, but I suspect that handling it SDI will not make as much difference as better attention to the MPEG-2 encode unless you're doing a lot of generations of recompression on the HDV on top of a lot of fairly extreme colour correction. Sometimes, software will by default use a poor-quality scaling algorithm to produce SD material from HD, and you should check all the available preferences and settings window to ensure that isn't happening. If it isn't, then you will in most cases be using a bilinear or bicubic resize, ideally the latter. These are probably identical to what a hardware downconverter would do, if perhaps not realtime - but that's not a very big deal if you're only doing TVCs.

 

Need more info.

 

P

 

 

 

My workflow:

 

 

If video footage is involved, I capture in FCP, editing, (Make graphics in after effects, export with alpha or pre-made background in Animation files .mov), Place them into my timeline. Then export another animation file. Open up compressor, and export the .mpg file with these settings:

MPEG2 files must meet the following specifications.

 

Resolution – 720x480

 

Aspect Ratio – 4:3

 

Stream type - System stream

 

Audio type – Mpeg1 Layer2

 

Audio sample rate – 48Khz

 

Bitrate type – CBR or VBR

 

Bitrate – 5 to 10 Mbps

 

 

Or I may make the entire commercial in After Effects, either in HD, or DV NTSC, export animation file. Do the same as above in compressor. And upload the final .mpg to an ftp account.

 

I do not have betacam-anything (no doe for that), I am curious about delivering the commercial in some other means, while remaining cost effective.

 

Please note these are the accepted methods of spot delivery in order of preference.

 

Betacam SP tape

 

Betacam tape

 

Betacam SX tape

 

Betacam Digital tape

 

DVC Pro tape

 

DVCAM tape

 

Mini DV tape (except long play)

 

DG Systems electronic delivery

 

Fast Channel electronic delivery

 

Vyvx electronic delivery

 

MPEG2 file via FTP

 

¾” tape

 

I always assumed with compressor that the video files would maintain their quality, but once converted to .mpg's (they die). Would HD to SD hardware down-conversion help with this problem? Or should I look for some sort of USB MPEG2 encoder?

 

Thanks Again.

Edited March 4, 2009 by CJ Henke III

[Phil Rhodes]

Sure, OK, but in what specific way do they "die" when you compress them? What specifically is the problem? MPEG-2 isn't an amazingly smart codec, but it shouldn't completely destroy things at a reasonably healthy, better-than-DVD 10mbps.

 

At what point does your ingested HD footage become SD? In FCP or in Compressor?

 

I can only reiterate my feeling that a hardware downconverter is unlikely to be noticeably different to a software one, and any hardware-based MPEG-2 encoder that's within the budget range of an individual is likely to be far less good than a sofware one. There are software encoders for MPEG-2 that rival the most expensive, high-end hardware types, just very slowly.

 

Can you post example frames showing the problem?

 

P

[CJ Henke III]

Sure, OK, but in what specific way do they "die" when you compress them? What specifically is the problem? MPEG-2 isn't an amazingly smart codec, but it shouldn't completely destroy things at a reasonably healthy, better-than-DVD 10mbps.

 

At what point does your ingested HD footage become SD? In FCP or in Compressor?

 

I can only reiterate my feeling that a hardware downconverter is unlikely to be noticeably different to a software one, and any hardware-based MPEG-2 encoder that's within the budget range of an individual is likely to be far less good than a sofware one. There are software encoders for MPEG-2 that rival the most expensive, high-end hardware types, just very slowly.

 

Can you post example frames showing the problem?

 

P

 

 

...Excuse the large pictures...

 

I always take uncompressed files and use compressor (with the settings above) as the last step.

It mainly has to do with artifacts and anti-aliasing.

 

DV ORIGINAL

 

 

MPG via Compressor

 

 

DV ORIGINAL TWO

 

 

MPG via Compressor TWO

 

 

Can you see the differences around edges, lines, etc. (The top logo appears to be squeezed, but it is from the screen capture) The problem truly occurs when a small .mpg is stretched to fit a larger screen with these artifacts.

[Rick Martinez]

Henke,

Phil is right.

A hardware compressor won't have a big degree of difference in relation to a software based setting aside processing speed for the process.

 

Probably your quality issue is more in the side of the compressing scheme/and your defined parameters.

 

I thought your were asking more about capturing directly from camera SDI HD outputs to disk array without compression introduced for tape or solid state memory storage.

 

I can't advice much on the MPEG DVD. (Just do simple works in that area for costumers reviews).

 

Sorry

[Phil Rhodes]

It's difficult to get a handle on exactly what's going on there, becaue they're screen grabs of a video player. This means that the image has been scaled and otherwise manipulated and it's difficult to make QC decisions based on it. The silver-on-blue logo looks aliased in the original, but then, it's been scaled by the graphics hardware, so I can't really tell.

 

I'd really need to see actual frame grabs exported directly from the edit software. This means for NTSC it'll be a frame around 720x480 and it's best uploaded as a PNG or very low compression JPEG so we can be reasonably objective.

 

That said, what you're seeing there is fairly consistent with MPEG compression. It's softer, as one would expect, and it's particularly clear on graphics, which this is. What will really kick off MPEG-2 style compression is the sort of sharp edges caused by aliasing, so your graphics need to be clean to begin with. Need to see straight frame grabs.

 

P

[CJ Henke III]

Henke,

Phil is right.

A hardware compressor won't have a big degree of difference in relation to a software based setting aside processing speed for the process.

 

Probably your quality issue is more in the side of the compressing scheme/and your defined parameters.

 

I thought your were asking more about capturing directly from camera SDI HD outputs to disk array without compression introduced for tape or solid state memory storage.

 

I can't advice much on the MPEG DVD. (Just do simple works in that area for costumers reviews).

 

Sorry

 

 

Rick,

I was interested in that, but Phil was basically saying that I wouldn't see much of a difference than working with firewire. The camera I would use 99% of the time would be the Sony HVR-Z1U, HDV camcorder. I just noticed in doing some research on SDI capture and Export cards, that most- if not all of them- have hardware downconverting, and it could have sealed the deal for me when it comes to spending the extra money on a SDI capture card verses just buying more hard drives for the computer.

 

All in all, I'm looking for all of the benefits a capture card has to offer, and all the cons as well.

[Phil Rhodes]

Phil was basically saying that I wouldn't see much of a difference than working with firewire.

 

You won't. The signal is compressed by the camera and stored on tape; that compressed signal is then sent down firewire to the computer, which decodes it and displays it.

 

If you come out of the camera as SDI, it's then being transported uncompressed; the only difference is that the camera is now decoding it rather than the computer. There might be minute differences because there are various approaches to this sort of codec mathematics, but I couldn't tell you which would be better.

 

This is a different issue to that I see with your graphics, which appear to have pretty bad aliasing, although as I say it's tricky to prove without seeing native resolution framegrabs. This could be an issue with downconversion if you're creating them at HD res then downconverting, but you should be able to solve that in software.

 

P