Correct me and fill me in please... This is NOT regarding live capture into computer, this IS regarding taped footage.

HDV is 4:2:0, so if i record to tape, and then capture footage via firewire into FCP it's regular HDV (i think)...

OR

If I were to instead, capture from tape, using HDMI and BM Intensity, is the quality going to be better or the same as the firewire capture? Or is there a quality increase similar to using SDI and how it decompresses the codec? or what???

Thanks,
Lonnie

There is a difference in decompression of the HDV content if it's managed at the camcorder vs the NLE, yes. Whether it's significant or not is a personal opinion not easily quantified. I like the chromasmoothing that CineForm applies, for instance. I also like the chromasmoothing is applied when using BMD and HDMI capture. But it couldn't be quantified as "10% better" IMO.

You'll also need something like DeckLink's Intensity:

http://decklink.com/products/intensity/

heath

Douglas,
Is the HDMI/BM Intensity workflow decompressing the footage from tape playback - in other words is it truly an uncompressed capture into an nle?
And is this a cleaner decompression than a software decompression offered via an nle render (FCP)? Or is that what you were referring to above?

(again, not interested in a live out, which bypasses the hdv codec, only interested in what's happening from the tape)

Thanks,
Lonnie

What about transport control? Does the HDMI interface include transport control? I think I heard someone at Decklink say you can control via Firewire, but capture via HDMI.

Check out the specs:

http://decklink.com/products/intensity/techspecs/

heath

Thanks. That's where I saw the control via FireWire in the Intensity specs. Anyone out there tried it?

I see one useful outcome of capturing from tape via HDMI: you can ingest the clips as "uncompressed" (although the usual uncompressed quality wont be there) and use these "uncompressed" files in the NLE to avoid the annoyingly long processing times characteristic for HDV native editing.

... and use these "uncompressed" files in the NLE to avoid the annoyingly long processing times characteristic for HDV native editing.

What "long" processing times? Playback is real-time with HDV. Multi-stream RT is better with HDV than "lossless" compression codecs because these require high-bandwidth disk systems -- often RAIDs. Real uncompressed requires RAID.

CineForm gets its RT power -- only in Premiere -- because they provide their own FX engine.

If you are talking about export -- that's got nothing to do with HDV. (Playback is real-time with HDV.) It's the encode time of the codec you are using.

And, if you use any AVID NLE you get Smart GOP Splicing so you may be able to avoid conforming on export.

Processing times=render time (some may refer to rendering as "encoding" yet "encoding" in current industry jargon is relegated to highly compressed web formats, although the word could be applied accurately to any compressed format).
I may be incorrect, but believe this is what Zsolt is referring to.

Realtime playback of HDV is *entirely* CPU dependent. Uncompressed formats don't suffer the challenges of decoding, but rather bring in its own problems via throughput/resource availablility.

Cineform gets its horsepower due to it being a lesser compression than native HDV, and being optimized to work in Sony Vegas and Adobe Premiere.
Adobe Premiere has an additional advantage with Cineform, as Cineform is able to access the Adobe Premiere HAL.

Processing times=render time (some may refer to rendering as "encoding" yet "encoding" in current industry jargon is relegated to highly compressed web formats, although the word could be applied accurately to any compressed format).
I may be incorrect, but believe this is what Zsolt is referring to.

Adobe Premiere has an additional advantage with Cineform, as Cineform is able to access the Adobe Premiere HAL.

Encoding is the term generally applied to inter-frame compression (compression expressed in bit-rate reduction). The fact some might see the term applied to the web is a side-effect of inter-frame compression being used on the web.

Compression is the term generally applied to intra-frame compression (compression expressed as a compression ratio).

Conforming is used by some to define the encoding process. This leaves the word render for the calculation of FX.

With AVID Smart GOP Splicing, there is no "conform" time if no FX are used. Conform only occurs in the areas where there are FX. Even better, where there are no FX, the original bits are exported with no decode/recode cycle. (Just like DV.) This puts the same bits back to tape as were on the source tape. (Except two GOPS at cuts.)

Since HDV export is the first step in going to BD or HD DVD, this is a huge time saver for anyone working with HDV. It also can deliver maximum quality.

HAL is Hardware Abstraction Layer. By using it, CineForm replaces Premiere FX code with their own FX code that have been written to work optimally with CFHD and Intel. So when you see CineForm's amazing multi-stream RT demos, you ONLY get this performance with Premiere. CineForm also brings 2-3 pulldown removal to Premiere for 24p editing.

so, in essence, HDMI capture from TAPE, is very similar to being a poor man's SDI - is this a correct analogy...

Lonnie

Since HDV export is the first step in going to BD or HD DVD, this is a huge time saver for anyone working with HDV. It also can deliver maximum quality.

.

There is no point in ever rendering/conforming/processing/exporting to HDV/.m2t file format output for delivery to BD or HD DVD.
A-.m2t is not a native-supported format in either HD playback system
B-rendering/conforming/processing/exporting is an additional compression step that I'm sure everyone recognizes the perils of performing. There simply is no need. Some applications do require an export, while most applications allow direct output of MPEG2, AVCHD, or VC1 directly from the timeline.
This is ALWAYS the best option. Otherwise, titles, graphics, and other 4:4:4 content on the timeline becomes 4:2:0 and is subsequently recompressed, and that's without mentioning the losses incurred in rendering an MPEG format to yet another more highly compressed format.
Every editor in the world (I believe) is well-informed when it comes to the subject of recompression and highly compressed source content.

"Compression," while being an accurate descriptor, is never used in the professional editing world. The term "render" on the other hand, is very commonly used when discussing output/export/mastering. As a great example, simply search this site for the term "render" and then again for "compress video." Do the same with Google. "Render" is the nomenclature used by professionals and consumers alike. I'm sure you're familiar with the common lingo of the industry?

Lonnie, HDMI is often referred to as "poor man's SDI" as it's quite similar to SDI for in-studio/capture/acquisitional use, but could never be used effectively in the field due to not only lack of cable length capability, but also weakness in the connector, and lack of ability to lock it down.

Those of us who have actually been burning hi-def optical discs know that because HDV is simply a "brand name" for MPEG-2 and MPEG-2 is what's used on BD and HD DVD, we can simply "export" HDV and -- with the correct application -- burn a disc. These discs will have menus, etc. And, except for the menus -- no recompression of the HDV is performed. The TS is simply repacked as PS.

And, if you use Avid NLE's, Smart GOP Splicing, can mean no "conforming" during .MT2 export!

Sony's BD burners also have a FireWire input that supports direct burning of HDV to BD -- obviously without menus.

And, we understand that BD and HD DVD use 4:2:0 sampling so exporting to HDV is the same thing as exporting to MPEG-2 at 4:2:0. Only the name is different.

And, of course, we understand that since exporting to HDV is the same as exporting to MPEG-2 -- there are no "extra" compression steps: camera encode, NLE decode, NLE encode, disc decode.

We also understand this can be done very cheaply -- no need for some expensive app to make hi-def discs. Spend $60 and you are in business.

No need to buy a $750 burner, either. Burn to red-laser discs. The 25Mbps of HDV is a perfect match -- although some NLEs let you tweek the "HDV" data rate so you can increase this to 29Mbps.

In the real world folks burn hi-def discs using HDV. And, unless one has a need to believe in the "extra" quality of inputting via HD-SDI or HDMI -- just stay with i.LINK. You are working with an under $5K HD camera who's own problems far out-weigh the issues of what type of xfer is used. So we are back on topic.

In this months DV mag, there is a review for the Black Magic HDMI card, and the reviewer says unless you are going LIVE out of the camera with HDMI into the computer, then you are better off just staying native HDV. This is in regards to quality of the image.

If you don't want to deal with long render times and other issues that have to do with HDV, then using HDMI in to another codec or uncompressed FROM TAPE is the next best option. In terms of picture quality from tape, go HDV native over firewire.

In this months DV mag, there is a review for the Black Magic HDMI card, and the reviewer says unless you are going LIVE out of the camera with HDMI into the computer, then you are better off just staying native HDV. This is in regards to quality of the image.

If you don't want to deal with long render times and other issues that have to do with HDV, then using HDMI in to another codec or uncompressed FROM TAPE is the next best option. In terms of picture quality from tape, go HDV native over firewire.

I would agree. If you're pushing color hard, or need fast renders, then the 8bit/10bit uncompressed file is the way to go. For the cost of drive space, it's a significant trade-off, but worth it if speed and quality of output/archive integrity are important to you.
I've got some footage captured via DR60 vs HDMI; I'll post a couple of stills when return home. At full rez, you cannot detect a difference on a full-screen monitor view. However, you can see a difference when you zoom in by 300%. Color is *slightly* deeper, and contrasted edges slightly smoother. This is due to the codec and management. But it's a whopping difference in HDD space.

Spot,

Ultimately, aside from space, would you even recommend getting the Intensity Card to capture HDMI vs. native HDV?

heath

For non-live capture?
Nope. I got the Intensity to go along with other HDMI products we're working with, so it was a no-brainer for us. But if all I had was a V1, a computer, and unlimited HDD...I'd probably consider it twice.
Bear in mind, we're constantly interfacing with XDCAM and HDCAM source, so it's actually a greater workload on the system to have native HDV on the timeline vs BMD or Sony YUV, simply for speed of edit/render. The content you saw for ESPN was captured both ways, and neither they nor the client was the wiser. Then again, ESPN isn't real-world. IMO, the benefit of the HDMI capture has far less to do with the camera output than it does the access to the better codec.

So, for a filmmaker like me, stick with just Firewire I/O.

Thanks!

heath

Thanks all - most helpful thread replies!
Lonnie

If you don't want to deal with long render times and other issues that have to do with HDV, then using HDMI in to another codec or uncompressed FROM TAPE is the next best option.

Long render/export time a primarily a function of the codec you are GOING TO not the codec you are COMING FROM.

The long render/export is shocking to many because they have never done HD exports before they had HDV.

Try this:
Take an HDCAM stream, render to DV. Using uncompressed HDCAM source, of course. Unlike HDV or AVCHD, there are several codecs that may be used for HDCAM on ingest over HD/SDI.
HDV stream, render to DV
AVCHD stream, render to DV.

Then render HDCAM to HDCAM
HDV to HDV
AVCHD to AVCHD

Now render HDCAM to CineForm or most any other HDI codec.

HDCAM to HDCAM will win in all renders, all things being equal at the proc side.

Heath,

If you're making a film and can control the situation, then I would think using an HDMI card and going direct from camera to computer would be the best option. A lot of extra requirements but the quality would be better. Unless you're shooting a road picture with a lot of remote locations!

I'm talking about capturing in post, not live.

heath