Capturing HDV From the Sony HDR-FX1 With Vegas

Kommentare

FrankFarian schrieb am 07.11.2004 um 01:18 Uhr
Sources at other forums have confirmed that FX1 can be captured into PC using CapDVHS and then edited in vegas 5b... search at www.dvinfo.net>HDV editing or www.sonyhdvinfo.com...

Win xp sp2 needs about 1-2 min. first time to "see" FX1 HDV via firewire/i-link - be patient.

But basically, capturing into Vegas is avail. HERE and NOW.
InterceptPoint schrieb am 07.11.2004 um 01:23 Uhr
"Win xp sp2 needs about 1-2 min. first time to "see" FX1 HDV via firewire/i-link - be patient."

What is it with Sony and Firewire? It is OK to be patient but no other device I have every used took more than a second or two to be detected via Firewore by XP - except my Sony TRV 900. And now I find that my next camera is going to have the same problem.

Methinks there is a problem at Sony not at Microsoft.
jaegersing schrieb am 07.11.2004 um 01:36 Uhr
Thanks nickle, I got it. Now all I need is an HDR-FX1!

Richard
nickle schrieb am 07.11.2004 um 01:57 Uhr
I'm waiting for the HDR-FX2 which is 3d with smellevision and comes with a 1 terrabyte hard drive.
Spot|DSE schrieb am 07.11.2004 um 03:01 Uhr
You can find the English translated version of the application HERE
(zip file)
It's also linked on the VASST Vegas Links if you'd like to see how the modified version is different from the Japanese version.
tnw2933 schrieb am 07.11.2004 um 03:30 Uhr
I can confirm that CapDVHS works perfectly to capture HDV from the Sony HDR-FX1. This afternoon I hooked up my HDR-FX1 to my P4 (2.8 GHz) using firewire. The moment I turned the HDR-FX1 to the VCR "on" position, my computer recognized the HDR-FX1 and displayed it as a device in CapDVHS. I captured 25 clips each ranging up to two minutes in length uisng CapDVHS while monitoring the video on the LCD of the FX1. The capture worked perfectly. I then transferred the clips into Vegas 5.0b, adjusted in and out points for each clip, created crosfades between each clip, a title at the begnining, and then rendered out the entire 20 minute project in a DVD Architect compatible widescreen format (slightly less than 2 hours render time) . I then used DVD Architect to burn a DVD of the project which I viewed in my donwstairs hometheater on a Runco 980 Ultra front projection CRT displaying on a 76 in. 16:9 screen. Of course, the material is in SD rather than HD when processed in this manner, but it looked very nice even at this large screen size.

Just for fun I hooked the FX1 up via its component output into my home theater system to display the video in HD direclty from the camcorder. The main differences that I noticed in additon to the greater sharpness and sense of detail in the HD video was that the colors were somewhat more saturated than in the SD material on the DVD that I had burned.

I look forward to comparing the SD material processed in Vegas to the native DV format captured direclty in the FX1 and then processed and burned to DVD to see if there is any significant difference.

Playback of the HDV stream in Vegas on my machine was limited to 6-8 fps, but I had no real problems editing the footage in Vegas.

DVHS looks like an excellent way to get HDV into Vegas until Sony-Vegas release a capture utility for HDV for Vegas. Finally, Cineform informed me that ConnectHD will not currently work with the HDR-FX1, but they are working on a solution for Vegas and the FX1.

Tom
PeterWright schrieb am 07.11.2004 um 04:47 Uhr
Thanks Tom

How good to hear about real experiences!

What sort of footage had you shot - indoors/out - natural light? - camera settings etc?

I'd love to hear more about how it was working with the camera.
farss schrieb am 07.11.2004 um 09:18 Uhr
The issue that I see is that simply removing one frame at the start end then re-encoding means all frames have to be re-encoding to maintain the same GOP structure. Mixing wavelet compression with mpeg-2 in the one stream would be a mess to handle and still doesn't address the issue of 3rd party apps such as AE.
What I see as significant also is what the camera records. It does a great job for what it is but shoot the wrong stuff and the images can get pretty horrible, I can say that having seen it on a 2K projection system.
Now decoding and re-encoding those problem frames will most likely make things even worse, the encoder still may not be able to encode without errors so things get even worse. You've only got to go down a few generations encoding problematic material to DVD, decoding that, encoding it again etc to see how fast things fall apart.
With the native DV25 codec in Vegas you can go down to 100 generations without noticeable loss, the results I'm hearing for native HDV are 3 generations and it's very noticable, this is not a situation I'm suspecting the Vegas team would want us to have to deal with, we'd be pretty much back to the days of 1" tape, and at least with that stuff the image just got softer, with mpeg-2 encoding the generational loss is visually very noticable.

Bob.
tnw2933 schrieb am 07.11.2004 um 17:37 Uhr
Peter and Others,

I have had the Sony HDR-FX1 for two days now. The majority of the footage that I have shot has been with the FX1 in full auto mode (auto focusing, auto exposure, auto white balance, etc.). The footage has been a mixture of indoor (by tugsten illumination, fluorscent illumination, and illumination provided by the Sony video light that I purchased which mounts on the FX1) and outdoor footage on beuatiful fall days here in Raleigh, NC. I have carefully watched all of the footage by hooking the component output of the camera to my Extron CVC200 transcoder which feeds a clean and very high quality RGB output to my Runco 980 Ultra projector. The footage is watched in a totally darkened room and viewed on a Stewart 76 in. wide 16:9 screen at a viewing distance of approximately 12 feet. Sound is taken out of the audio out of the FX1 and fed to my Lexicon MC-12 surround processor and distributed to the 10 speakers in my surround system. I go into this deail so the comments below can be judeged in light of the equipment and conditions under which I am vieiwng the footage from the FX1.

The bottom line is that I could not be happier with the results provided to date by the FX1. Autofocusing is extremely fast -- even under conditions where I would expect the autofocus to have problems, e.g. a goose flying over a lake at about 150 feet away with a clear blue sky as background. I tracked with goose landing on the lakel. I have viewed this 10 seconds of footage over and over in slow motion becuae it is a thing of beauty. The goose is kept in very sharp focus throughout the entire footage. There is no searching or "fading" in and out of the autofocus on this difficult subject. Slow motion playback by the FX1 is simply a thing of beauty with no jerky quality and smooth as silk.

Colors recorded by the FX1 display excellent fidelity to the subject under all ligting conditons -- even fluorescent lighting. My shots around a nearby lake were recorded during the last hour before the sun dropped below the horizon and the ligting captured by the FX1 retains a warm and very nice quality. Several shots of fall foliage with beautiful reds, yellows and greens mixed into the shot with strong backligting were perfectly exposed and stunning in quality. Pausing the FX1 results in a spledid still image on the screen that is sharp and detailed. I have shot lots of pans and subjects in motion (like the geese flying) and I see no motion artifacts whatsoever.

The detail captured by the FX1 is amazing -- and I am used to watching HD broadcasts over both C band satellite and local fiber optic cable. Even very distant subjects show remarkable detail, and medium to closeup shots are breathtaking in detail.

Overall shooting with the FX1 in full auto mode produces some very very nice footage indeed. However, under these conditions colors are highly saturated and look very "Kodachrome" like so if that is not your preference you may have to experiment with other settings.

Perhaps the most pleasant surprise in using the FX1 is the unusualy high quality of the audio recorded by the camera microphones. I have a Yamaha disklaivergrand piano in my living room and I played a piece on this acoustic piano while my wife filmed the scene from about 10 feet away. I played the piano at a comfortable leve,l and I played a piece with excellent dynamics ranging from pianissimo to forte. Weh we watched and listeed to this footage in our home theater dexribed above, the picture is not only stunning, but the sound is very natural with excellent dynamics, no hissing or pumping, and of very satisfying quality. (I do wish the pianist had done a little better job however :). Outdoors wind sounds are handled very very nicely by the cameras mics -- much beter than I have ever heard with any other consumer level camcorder. In a quiet room while recording and zooming, I cannot hear any nosie produced by the FX1 on the recordings -- no whirring of zoom or focus motors, etc.

The FX1 offers the opportunity to manually adjust everything and affords selectable Picture Profiles for storing your manual settings. Over the next several days, I will be explolring these manual settings under a variety of conditions.

Bottom line: the FX1 is a worthy successor to my Sony VX1000 and I am glad that I waited for a prosumer HDV camera to come from Sony. Sony is to be congratulated for the job they have done with the FX1.

Tom

scdragracing schrieb am 07.11.2004 um 17:52 Uhr
the way mpeg2 editing works is that breaking a gop for a frame-accurate cut means that you have to re-encode only that gop, not the entire mpeg2 file... and since you are referencing the original hi-rez frame to reconstruct the rest of the gop, there is no quality loss with cuts-only editing.

doing real editing means that only the group of gop's involved in the edit will be affected, you do not have to re-render the entire mpeg2 file... only the areas where the edit happens, just like with dv.

i don't see any issue with using seperate codecs for transitions and renders, that is how it's done already with some dv editing systems... it's no different with mpeg2.

bob, you are directly contradicting yourself when you say that we must use wavelet compression for capturing the entire hdv file, but then you claim that hdv will only take 3 generations of re-encoding... guess what, maybe that is why avid and others are capturing native hdv instead of re-encoding the entire file when capturing!

but i agree with your general points regarding lossy mpeg2 compression as a less than favorable editing medium... which is of course why you must capture native hdv data.

i appreciate the real world experiences, but was everything done in 720p or 1080i??? since displaying 1080i on a pc takes at least 6 times the processing power of an sd display, there will be a big difference between the way that 720p and 1080i function in vegas.
scdragracing schrieb am 07.11.2004 um 17:57 Uhr
tom, i have read a number of complaints about too much compression with hd broadcasts via both cable and satellite... isn't voom supposed to be da bomb for hd??

did you work in 720p or 1080i?? was there any picture breakup during hi-motion scenes with 1080i? how about during fast pans and zooms with 1080i??
farss schrieb am 07.11.2004 um 19:18 Uhr
I'm not contradicting myself at all. It's the native HDV codec that suffers generational loss, decoding and re-encoding that is what falls apart. Using an intermediate codec is done to avoid those losses.
Whilst technically you can cut mpeg-2 by healing the GOP, assuming the GOP requires a fixed length GOP as I said something as simple as removing one frame at the start means rebuilding the whole thing from the start.
But that's a pretty academic argument anyway. A cuts only edit is pretty limited, you're deny yourself color correction and composits to say nothing of standards conversion. My current project is being released in both NTSC and PAL and every frame has a bug composited over it.
To answer your other question, Yes everything done 1080i, two tracks played out with composits, with mixed media, SD, CineAlta and HDV on the one timeline with no rendering. All into 2K projector via SDI. For sure meaty PC, cost locally $4K, whole solution (minus projector) AUD 10,000. Not expensive.
Bob.
tnw2933 schrieb am 07.11.2004 um 21:06 Uhr
Scdragracing ,

Not counting the FX1, I have two available sources of HD material. I get several HD channels off of a 10 ft. C-band satellite dish. These channels are not compressed at all. I get about a dozen HD channels off of Time-Warner cable. I am sure that some compression does occur on these channels, but rarely do I see any breakup or artifacting on these TW cable HD channels. We are fortuate to have excellent cable service here in Raleigh. I do not have a small dish system at all so I cannot comment on Voom

The FX1 will capture footage only in 1080i. It will not capture in 720p although it will play back tapes recorded in 720p. Thus all of my footage shot is 1080i which was sent to my Runco without conversion to 1080p.

The only scenes that I have shot to date (only had the camera two days) that involve motion are shots of geese flying across my field of view and motion produced by zooms and pans. I have seen no evidence of breakups (which I, too, have heard is a problem with mpeg2 video) during any of this footage. However, I have not deliberately zoomed or panned the camera at ridiculously high rates just to try and produce breakups. My pans and zooms were of the speed necessary to follow the subject or to move in on the scene gracefully.

Tom
mhbstevens schrieb am 07.11.2004 um 22:10 Uhr
ANY digital format or information can, with the right processor and compilers be reduced to machine language - ones and zeros - and these digital processes can be reversed and recompiled to any other format. Why does someone not write a program to convert the mpeg2 or m2p to a format where each frame stands alone (and is not dependant on preceding frames as in mpeg) and which can be edited in any NLE. The the final product can then be recompiled to whatever final format is required. I think we don't see this done because the format owners are protecting their product by not using open source compilors as this would be. Someone out there get a programmer to write this stuff.
JJKizak schrieb am 07.11.2004 um 22:18 Uhr
tnw2933:
How can you get non compressed HD off a satellite when its transported to the satellite with MPEG2 converters? Me thinks you are hallucinating.

JJK
Hulk schrieb am 08.11.2004 um 04:50 Uhr
Does WinXP recognize the FX1 without any video editing software installed?

Does it recognize "unknown device" or does it specify the model, etc?

If software is required for recognition then which software deos it?

- Mark
farss schrieb am 08.11.2004 um 06:43 Uhr
You are right this all a conspiracy by the multinationals to deprive us of a basic human rights and enslave us to their desires.
Oops hang, haven't you heard of uncompressed AVI, oddly enough you can render that straight out of Vegas, perhaps working with such a radical format could be considered a sure sign of leftist tendencies or is it too a plot by the multinationals to make us buy terabytes of disks that whizz around at mind boggling speeds just so we can watch our lossless video?

Bob.
VivaVegas schrieb am 08.11.2004 um 13:02 Uhr
All in all the camera has been in development for over a year. The HDV spec has been out since the JVC HD10 and HD10. SONY has had more than enough time " to get it right "

Chris
Spot|DSE schrieb am 08.11.2004 um 14:58 Uhr
So has JVC. They've had two products ON the market for a year, and they haven't "gotten it right" yet.
If it's easy, everyone would be doing it. I know that Sony would very much appreciate ANYONE who would do difficult things for them, present them with a totally rock-solid package that would integrate with what they've got, and would do it at a cost that would be low. But you'd have to do it prior to presenting it, because if it were easy and cheap, they'd already have done it.
All I can say is if it's easy....go for it. I'll put you in touch with the head honcho at Sony once it's done.
tnw2933 schrieb am 08.11.2004 um 16:17 Uhr
Mark,

I hooked my HDR-FX1 up to my P4 2.8 GHX computer running Windows XP Pro SP2 via a standard firewire connection, and it was recognized virutally instantly by Windows XP (new hardware found dialog box) and a driver installed for it by XP. When I opened up CapDVHS to capture the raw HDV footage from the FX1, CapDVHS immediately recognized the Sony HDR-FX1 by name as a device and I ws able to capture some 25 clips of for a total of about 20 minutes of raw footage without dropping a single frame. I then edited the material in Vegas 5.0b.

Tom
tnw2933 schrieb am 08.11.2004 um 16:21 Uhr
Spot,

I wonder if you would care to comment on whether Pnnacle Systems has got it "right" in their Liquid Edition 6 which as near as I can tell from http://pinnaclesys.com/ProductPage_n.asp?Product_ID=2452&Langue_ID=7 does appear to capture and edit the HDV stream from the Sony HDR-FX1. It appears that Pinnacle System's Liquid Edition 6 is available NOW.

Tom
Spot|DSE schrieb am 08.11.2004 um 16:43 Uhr
Sure it is available now. When and if you can get Edition to run, and when you can get it to capture HDV. Neither of which is a winning prospect. Check out Pinnacle's forums, don't just take my word for it.
That's the thing....it doesn't work. Sony also has a lot of other considerations.
Plus, as I've said at least a dozen times before....Sony hasn't announced their professional HDV camera.
Ulead, king of MPEG editing software, doesn't have this 'right' yet either, they've announced an upgrade that will carry some slight cost, but it's not released yet either.
It's like 24P. Easy as hell to announce. FCP announced it at NAB 2003. It wasn't working solidly until just before NAB 2004. Adobe announced it at NAB 2003. It STILL isn't working solidly.
Vegas 4 came out of the gate with solid 24p, as did Vegas 3 or e, I've forgotten which final build of 3 had it.
But when Sony announces it, you'll see it working, and working correctly. These guys aren't dummies in engineering, and their marketing department also knows that it's clearly a mistake to announce something "just to be first" and then not deliver.
BTW, I'm not basing my Pinnacle stability comments on their forums. I've got Edition on a beta machine, I've run it several times with tools I can't comment on, and it fails. Regularly.
John_Cline schrieb am 08.11.2004 um 16:59 Uhr
Tom,

The key word is "appears." As a former long-time Pinnacle user, I can assure you that Pinnacle hasn't "gotten it right." Liquid Edition is the most ill-conceived, user-unfriendly and BUGGY NLE out there. I was a beta tester for Edition 5 and I could just never wrap my head around their way of doing things and I had used all different types of NLE software. I've checked out Liquid Edition 6 and it's no better.

I haven't been there in a while, but perhaps you should go check out the Pinnacle Web Forums and see what the users are saying. The last time I checked, any of the Pinnacle forums, they were full of people complaining about bugs, usability issues and non-existant support. This has been the theme of the Pinnacle web forums for years.

Pinnacle Edition Web Forum

This is in stark contrast to the Vegas forum which is mainly about using Vegas, not complaining about it.

Since I'm not going to buy the FX1, as long as Sony has an HDV editing solution for Vegas by the time they release the pro version (which I am going to buy), I'll be happy.

John
scdragracing schrieb am 08.11.2004 um 20:58 Uhr
what a great thread!

there is a nice article or two in the december dv mag about hdv, wm9 hd dvd creation, etc... contrary to what i posted yesterday, they claim that even a cuts-only edit could cause up to half a second of re-encoding while healing the gop... the wm9 hd menus are html-based!

the gop size may or may not be an issue if you wanted to copy the finished edit back out to an hdv tape, but since most delivery formats will require re-encoding anyway, you wouldn't think that mandatory gop sizes would matter.

the problem with using vegas for hdv is that the entire timeline must be totally re-encoded, there is no hdv out, and there is no mpeg2 integration within the app itself... hence the need for the mainconcept encoder... but that all may be o.k. if you deliver to sd.