Adobe pp2.0
 

I See Adobe pp 2.0 edits natively without any loss using codec, anyone tried this with JVC HD100 and does it support 24p? I know 1.5 did this as well a few months back but it seems pc power has come along a bit since then to cope with the larger information involved with native editing.

Cheers Greg C

Adobe PPro2 does have loss. It clips Superblack and Superwhite information out of your footage. With a rough calculation i estimate that you are losing around 10%-15% of your image's information.

It does exactly the same with DV footage.

Maybe the reason is the way PPro2 works internaly and how it converts YUV information to RGB.

I contacted Adobe about this 'problem' but i was suggested to use Cineform.

I am using Vegas now. If you are using an older computer, Vegas is not as smooth in editing HDV as PPro2 but it keeps all the footage's information.

Thanks Miltos,

Strange that adobe suggested cineform as surely this creates a slight loss of info also. Furthermore I thought the whole idea of native editing was to cut this out or have I misunderstood.


Regards Greg

Well,

I guess there are hidden meanings in words like "native" or "visualy perfect" :-)

The Superwhites and Superblacks are not supposed to be used in broadcast so Adobe decided for us that we don't need them but i can easily prove how essential they are in Colour Correction or if your work is going to end up on film.

And just to be clear, i have nothing against Adobe. I love how PPRro2 works and i find it's workflow 'easier' than Vegas'. It's also the first editing software i learned and i have a kind of affection for it, but the clipping is unacceptable for me.

Maybe other editing software do similar 'hidden' things to our footage. I only had the chance to use Ppro and Vegas so someone else might help on that!

Greg.
Keep in mind that native HDV editing is, generally speaking, a bad idea. It's imporessive that companies like Adobe and Apple offer this capability but a GOP-based codec is not designed for editing. What the software does, behind the scenes, is to decompress and recompress a lot of frames in order to allow you to do fram-accurate cuts in the middle of a GOP. While this can be convenient for short videos, it defintely pays off to convert your footage to a better codec and then continue from there with CC, effects, etc.
The time to perform these operations will be shorter and the overall quality of the result will improve.
This, again, is the nature of the codec and is not a consequence of using a given NLE.

Does FCP administer any clipping?

The Codec from Cineform does make a big difference during editing compared with native editing. That's why the Cineform converts the original m2t file in their own format to preserve the quality after multiple generations.
My backup files of projects are always Cineform. I only convert back to m2t to store on tape.
Here is also an article on the webpage of Cineform were they compare Cineform to native editing:
http://www.cineform.com/technology/H...tyAnalysis.htm

HDV Editing
 

I read very often about the poor editing qualities of HDV even though all the NLE advertise that they can handle native HDV etc.
I use Avid Xpress Pro 5,5,1.
We capture and cut HDV 720p 30fr/sec, as 24 25 are not supported yet.
The workflow has no problems at all, we can run 4 layers of HDV plus titles and effects (Titles in DnxHD 110 codec-resolution) in real time without a mojo. The effects rendering is relatively fast, even faster than DV sometimes and the export to M2t is about 2,5 X realtime.
I am waiting for a macbook pro next month w/ FCP and I hope that will have same performance.
In my opinion there is no point compresing and recompresing HDV.
I think that the best way to keep it clean from artifacts, color loss etc. it is to edit it natively and export it back to tape without any further manipulations. Avid also suggests to transcode HDV into an DnxHD codec and edit using this 1 frame based format.
I tried it and I realized that the trascode process is tooooo long without any benefits on the final editing experience. Same with DVCPro HD.

If you're happy editing in native HDV, that's great because it makes life very simple and I'm glad that there are systems out there that make this possible. However, technically speaking, HDV is far from ideal for editing. As many people on here have already pointed out at some stage, this is because it puts a high load on the system (Mpeg2/IBP and all that) and because it is extremely lossy on re-compression. Cineform and other intermediate codecs such as Canopus HQ and Avid's DNx try to address this issue. As an Avid user, I do not have extensive experience dealing with the first two but I must say that the difference between finishing with DNx and native HDV is very clear to my eyes, at least. Of course, whether or not a native HDV post process is adequate is entirely down to the demands of any particular job but I would argue that you'll get far better results working in DNx for the finishing process and that transcoding to DNx is well worth the wait and effort in most cases. Of course, there are simple and effective methods on all serious NLE systems to limit the amount of transcoding/conforming necessary for longform projects or those with high shooting ratios.

Re-editing HDV
 

Hi Antony, as an Avid user you know that for HDV 720p we get only the DNxHD 110 that is supposed to be if not worst... equal to HDV 720p.
Even MC provides only this resolution for HDV 720p projects.
So to tell you the truth, I do not see any difference with my eyes.
The only difference is that I have to wait long hours for the transcoding.
Avid is supposed to use the splice tech that is processing faster the gop of HDV for rendering and previewing purposes.
For me again editing native HDV is the best way to go.
I have experienced loss in quality using both Cineform w/ APPr and Canopus AVI HQ.
Maybe there are ways to control those codecs that I ignore but native looks so easy and the footage makes no different from capture to final export to tape.

Panos, because of its nature, when you edit in HDV you end up recompressing the footage. There is not escape from it. If you apply a cut in the middle of a GOP, and this is going to be the most likely case, the NLE has to generate new GOPs whenever it needs to render the footage. Some effects/filter can be applied in real-time but whenever that is not possible, new GOPs have to be created and this means that the NLE will uncompress the HDV footage and then re-compress it with a resulting quality loss. Now, if your footage is simple and short this is unlikely to happen but if you apply grading and/or special effects, the recompression will be unavoidable.
My personal, totally personal interpretation of this feature is that HDV native editing is really meant for consumer-grade HD cameras and to give the occasional editor the chance to edit simple videos without being overwhelmed with technical details.
For complex editing work is best to stay away from it.

This is a little above my head at the moment, but say FCP version 5.1.2 delivers native 25p editing for example am I still best following one of workflows/work arounds that have been posted on these boards?
Majority of work I will be doing is HDV 25p short documentaries/films/music promo's and corporate vids...nothing too long at this stage.

Does this put Avid Liquid in the consumer camp then?

I have Liquid, but only have used it for DV. Apparently with HDV you can have the frames render in an uncompressed format However, I'm not sure it's possible to get the video out of Liquid in an uncompressed format that other programs can read.

It has nothing to do with the frame rate. And yes, you can edit directly in HDV, it's your call. It's a matetr of "best practices", not a technical requirement.

Think in this way, you drop a clip in the timeline, the clip starts and there is a GOP: 1 frame is captured complete and 5 frames are stored as the difference between the previous frame. You apply a cut on the 3rd frame and discard the previous two. How does the NLE show frame 3, 4, 5, and 6 if the starting point has been removed? Of course, since a cut is just a pointer, the program can "rewind", find the reference point and then compute the following frames. Seems convoluted? It is, but this is exactly what happenes when editing a codec with temporal compression. Now, take this clip and imagine that you want to export it as a QuickTime movie using the original encoding: HDV. After all you think: "If I don't transcode, I don't loose image quality". That would be the case with a codec that stores all your footage as discreet frames. The frames would be simply "dumped" in to the new file. With HDV you have to rebuild a new GOP, because you just made a cut in the middle of one and discarded some frames. So, the NLE has to "rewind" again, compute the frames that you have in the sequence and then re-compress them in order to create a brand new GOP in the file. This time frame 3 will be the reference frame and frame 4,5,6, plus 2 frames from the following GOP will be stored in the file. I'm oversimplyfing here but you get the gist of it.
Moral of the story, you end up with a recompressed, re-encoded file, even when you don't change codec in export.
Now, take this scenario and apply it for effects, transitions etc. Not a pretty picture. The time saved in avoiding a conversion to, say, Black Magic 4:2:2 10 bit or AIC, is paid down the line, when the footage is spread across hundreds of cuts and several transitions, filters etc. And I'm not even considering chromakey work. This is for your everyday, 30 minute interview or similar footage.

Hope this helps.

This is sound advice/information from Paolo. There is a fundamental misconception in the idea that editing native HDV is the best solution for maximising image quality. Native HDV support is great for simplicity of workflow and for ingest of material but it is most certainly not ideal for manipulation of images in post. Of course, decompressing to an uncompressed codec and carrying out all finishing work uncompressed would yield the best results but most people do not have this option due to system and drive speed. Intermediate codecs such as those discussed - depending on the quality of their implementation - will deliver a good compromise.

For any Avid users out there, DNx 110 should certainly give you better results on re-compression than native HDV. So, if you're doing a lot of colour correction, compositing or other post manipulation, you will most certainly see the difference. If your system is fast enough to work well in native HDV it is fast enough to transcode to DNx efficiently if you consolidate and transcode (or decompose offline SD, conform and transcode) for finishing.

It's all about what is necessary, really. If you're happy working native HDV and it gives you what you need without the hassle of transcoding or re-capturing then why worry? It is, however, simply incorrect and misleading to suggest that native HDV is the ideal solution for best image quality as Paolo has so thoroughly explained.

Man, it just gets better by the second...
 

Same question for PPro 2 - I had no idea about the quality loss. How does/will FCP handle these various questions (native HDV, quality loss, etc)? What would you use for narrative film work?

What diference does all this mean when considering a film out vs normal DVD, or HD-DVD? THanks for all the knowlege -
john
evilgeniusentertainment.com

FCP allows you to select the codec used for each sequence so you can capture the footage directly in FCP, good for clip logging, and then drop it into a sequence that will hadle the cuts in uncompressed format. When you drop the clips into the sequence they get converted automatically. The render files, now in the new codec, will be used from that point on. You can use any supported codec. The Black Magic codecs are built-in, they are called "Uncompressed 4:2:2 8/10 bit" or you can download the latest versions from BM's website. AIC is another option that is provided as part of the FCP package.
You can also use an external package like MPEGStream Clip to batch convert the .mpeg files explicitly.

re re editing HDV
 

If this is the case then all the argument from the official NLE marketing is bull...
As Avid suggests for Xpro and MC "Edit native HDV" this is all about editing native... without any loss and bla bla.
As I use the JVC HD101 I am limited in Avid Xpro, as well MC and cannot change resol. in my timeline.
As far as now I went native HDV and I really had no problems with quality losses.
We work on short and long projects, one of our documentaries was projected with an HD projector on a big screen and the resulting screen was really amazing.
I will try to transcode into DNxHD110 for a doc that we work now. I want to test the difference with CC etc.
As soon as I will get my macbook with FCP I will come back for advises for that environment.

For Paolo: I shot the first footage of the sea nomads of east Sulawesi with your TC3.
I really liked the results same did my Japanese partner. I will definetely use the same set up again when I will go there on the 15th of November for the final movie. So "Grazie caro Paolo".
If you have any other suggestions I would like to know and try them too.
I can send you frames from the first shooting at an e-mail adress if you are interested to see. Everything within sea and sky... looks really great.

Many thanks for that thorough explanation Paulo...I'd almost considered the Sony Z1E at one stage and that has quite a few more GOP's if I'm not mistaken?
I'll soon be taking hold of FCP on an iMac - I guess I'll be following the workflows (which I may need to come back to for confirmation if ok with you folks!) posted here rather than anticipating native HDV editing on a rumoured update. Not too concerned about time saving if the results are not as I would've hoped. From the footage/shorts posted on these threads would I be correct in stating that they all have been edited non-natively?
...and by the way - how do I link Black Magic and or AIC into FCP - are these plug ins I could get hold of or are they part of the package anyhow?

Many thanks for the help.
dave.

I know what you mean about the marketing spin, Panos. Avid (and others) make a big deal about native HDV support yet they don't actually fully support HDV1 (I notice they changed their marketing text after a barrage of complaints on the Avid forums); they make a big deal out of working 'native' and then they make a lot of noise about DNx being the best thing since sliced bread. Of course, they keep pretty quiet about uncompressed HD in MC and AXPro marketing because - if you want that - you have to buy Symphony or DS Nitris (around £60,000) when FCP will support it as is and the extra hardware you'd need is far cheaper. I think FCP plus the Blackmagic hardware and a SATAII array is a wonderfully cheap solution for uncompressed HD by the way.

As to the whole native thing, it is great to have native support for ingest of material and ease of workflow. In the case of Avid Xpress Pro, it means that you can ingest your material via firewire and then transcode to DNx without needing MC Adrenaline and the HD board (oh, and an HDSDI converter for the HDV camera or deck). Native HDV also keeps data rates and storage requirements low and means much cheaper kit, so it can make a lot of sense if you get the results you need this way. However, many (and by no means all) people working with DV for broadcast choose to work on an uncompressed timeline for finishing to avoid fx/titles/cc re-compression and output to, say, DigiBeta for online, and the underlying motivation to work uncompressed for HDV material is even greater. DNx stands up over re-compression far better than HDV and also DVCPro HD but - especially on the 8-bit variants - it does eventually fall apart.

It all depends on your needs, of course.

... and, on the subject of Liquid and PP2.0 I think it would be pretty snobbish for anyone to say they were firmly in the consumer camp. They are both good products, IMHO. With HDV, I believe both offer the option to work uncompressed HD and PP2.0 also supports Cineform as a compromise. You'd need extra hardware and fast storage for uncompressed HD, of course. I'm not experienced with either application so can anyone confirm this?

Paolo, please read my current HDV@Work Newsletter (http://digitalcontentproducer.com), because will find that the ENTIRE de-compression + re-compression is, with FCP and Liquid, a myth. It CAN happen with an Avid, but it can be easily prevented.

PS: Yes Premiere clips the signal, something Liquid and FCP need not do.

Michael - Would I be right in suggesting that I wouldn't have the facilities to use the Blackmagic hardware/SATAII array via in iMac? I was investigating the possibility of using an HD connect box but because this mac does not have a card slot facility this would be out of the question - same apply to the black magic hw?

Time saves/workflows/native aside I'm really looking at which method will give me the best results from HDV footage using FCP on an iMac - it may be the extra hardware is not suitable for this or at this time it might be a little expensive but certainly worth considering in the near future.
Many thanks.

Well, it' just some clever marketing. There is a need for that feature and we defintely need the ability to "ingest" HDV. For example, with our camera is very helpfull to be able to log the clips so that you can re-create the project later on, after you removed all the working (huge) files. Where you go fom there is a different story. BTW, this is hardly a new trend. "Digital Moviemaking" by Scott Billups spends a lot of word on encouranging people to verify and test all product claims. It also spends a lot of time demystifying the "uncompressed" claims of several solutions. It's just the nature of the beast, you have to do your homework. The truth is defintely out there but you can trust no one ;)
Who knows, the final result of HDV-to-HDV rendering can be pretty good. The only way to find out is to cut a bit of footage and then compare pixel-by-pixel the resulting frames.

I don't doubt it. I watched some footage shot by Jody Eldred with the Sony F-350, which, BTW, uses a variant of HDV, and projected with a 2K digital projector on the screen of the DGA theater and it looked fantastic. It's a different thing when your digital footage is transferred to film, though. Filmout can highlight artifacts that you don't see in the digital projection.

And regarding TC3, you are very welcome (don't know how to say it in Greek :), I'm glad that it works for you and I'd love to get some still or any other material that you shot.

Take care.

The "best" quality you'll get is the original transport stream imported via FireWire. It doesn't get any better than this. It only expands in size with an Intermediate codec -- and with uncompressed video, a RAID becomes necessary.

With FCP, all processing is done with uncompressed 4:4:4 video. There's no need to decode HDV until the point it is processed in some way. It's just data in a file.

In most cases, FX are RT and so renders are not needed. And, after being decoded and processed, FCP will not use a rendered file unless you want to use it. And, renders are only for a faster playing PREVIEW. It has nothing to do with final quality.

The only compression or recode is when you EXPORT -- and that's going to be true even if you work with uncompressed.

The same is true of Liquid.

Avid, Premiere, and Vegas users, however, can benefit from an Intermediate codec. But, these can be "nearly lossless" codecs and so far more efficient than uncompressed.

In short, your iMac is perfect -- unless you need many HDV streams. Then a Mac Pro will offer more CPU power.

Steve, I'm not sure this is true. When I export a sequence from FCP to QT, using "Current settings" the export is done in 3-4 minutes for about 3-40minutes of footage. When opening the resulting reference movie, press Cmd-J to get the file info and you'll see that it references all the render files. Futhermore, it could not physically export that much footage in such short time without referencing the render files. I verified this by switching the sequence back to the original HDV configuration and the export took about 4 hours. This is as expected since at that point the NLE cannot reference the HDV files but it has to recalculate the new GOPs.

Interesting stuff, isn't it :) ?

I know this is probably not the place for this, but here it goes, i also edit with APP and Matrox RTX100, i'm thinking in upgrading to RTX2 and APP2 to edit in HDV (i have a JVC HD100), is it a waste of many in buying a video card would it be better to use just APP2 or another NLE like FCP or AVID? Thak you guys.

Luís Ventura Santos

Luis,
I had the same question before the RTX2 was out, I still own a RTX100 Extreme myself. I choose to testout the Cineform codec which gave me near real-time editing with High quality conversion from m2t to AVI to be used in PPro 2.0.
I capture everything in HDV now-a-days even if it's for SD-DVD purpose, and don't really miss the RTX100. After I finished and converted some old projects, I will sell the card.

Marc,
thank you for your help, as i understand you are going to keep only APP2, about FCP you don't have any expirience.

Luís

I never use "current settings" as I either make an HDV movie where everything must be re-rendered OR I'll make a DV movie where everything is re-rendered OR I'll export an MPEG-2 file to go to iDVD OR I'll export an uncompressed file. All of these will not use render files because I'm changing the nature of the codec. I believe COMPRESSOR will not, for the same reason, use render files.

But, Liquid is so much better for HDV that I haven't used FCP for months. FCP is a decade behind Liquid for editing MPEG-2.

Steve Mullem
"ProHD Handbook"
www.mindspring.com/~d-v-c

Well, we have different workflows. I pre-render the sequences since I often need to do it anyway when using plug-ins like Magic Bullet. Given that I have to render the effects, I select the codec in the sequence, usually AIC, let it go, when I'm done I use "Current settings" to export the whole footage in a couple of minutes. Then I drop the resulting QuickTime file into compressor and do the final transcoding.

I tried MB once and couldn't see any reason to wait for rendering just to get a "look." The audience will never know I chose not to spend time getting look "A" if they see look "B." So, if B is RT -- why not?

In short, if it's not RT -- forget it. :)

Oh Brother Where Art Thou used special coloring techniques that took weeks to do. They did a test render on certain frames from scenes and then decided on a final look, and did it.

Maybe you could do something similar...certainly MB doesn't take weeks to render...

Of course this is now entering the field of image grading/color correction, aestethic choices etc. etc. What I like about the digital grading techniques like Magic Bullet is the the ability to shift the gamma settings further and more finely of what you can do in camera. It's the digital world equivalent of, partially, what is done in the film lab. There is a free plugin for FCP, http://www.pureandapplied.com.au/ , that has a "film gamma" effect that can be used to esperiment with the idea. MB gives you more control but the concept is similar. The way that you can shift the gamma and then tweak it until it gets right is a way of improving the look of you footage that I appreciate a lot. One of the great luxuries of shooting in digital. Another feature that I like is to bet able to finely control the look. Color shift (warm/cool), diffusion and many other factors. It's just a matter of getting the image where I want and one of the ideas behind TrueColor: get the most neutral look in camera in order to keep all the bits intact and ready for digital grading.