|
|||||||||
|
Thread Tools | Search this Thread |
May 8th, 2009, 01:14 PM | #31 |
Inner Circle
Join Date: Mar 2003
Location: Ottawa, Ontario, Canada
Posts: 4,220
|
I depends whether you view re rendering as reducing the quality. I don't think that is the case at all. It takes time . Once you re-code to Canopus HQ or Cineform you are in a much more robust file format that will not degrade with further processing and is easy for NLE's to process. In my estimation the converted file is as good as or better than the original file. In my editing I have chosen to render out from the NLE as MPEG2 HD as I find this easier and faster. The quality on Bluray is the same in my opinion and if you don't need to put more than about 2 hours and 30 mins on a Bluray there is little point in going to AVC.
Ron Evans |
May 8th, 2009, 02:02 PM | #32 |
Major Player
Join Date: Dec 2006
Location: Traverse City, Michigan
Posts: 416
|
Hi Ron. Yes I do view rerendering of video as reducing the quality of that video. I have seen it. Pinnacle 12 rerenders, while Nero 9 does not. When compared on a 42" plasma, there is a difference; Pinnacle is much softer than Nero.
You say that the video quality is not degraded or is even better when you recode your file using Cineform Canopus HQ, and that it looks fine on BR. Please take me through your process, step by step, as you take your raw footage from your camcorder, to burning the final product on BR or DVD. Thanks. Mike |
May 8th, 2009, 02:37 PM | #33 |
Inner Circle
Join Date: Mar 2003
Location: Ottawa, Ontario, Canada
Posts: 4,220
|
Both Canopus HQ and Cineform codecs are high quality intermediate codecs that decode the AVCHD files to a high quaility file for editing. This file is indistinguishable from the original. Then there is the quality of the encoder to whatever file format one needs to go to.
My process is to transfer my AVCHD files from the Sony SR11 or XR500 using the Sony Motion Browser software that came with the camera. This stitches the small files together into one big file as appropriate for each take. Mine are usually quite long theatre performances so usually over an hour. I use the Edius AVC2HQ converter to convert these AVCHD files to Canopus HQ. I place these HQ files on the Edius timeline with the HDV files from usually the 2 FX1's so there are now 4 tracks on the Edius timeline. I then use the multicam feature in Edius and edit. I then render to either a Canopus HQ file in HD or to HDV depending on whether I want to keep a HDV copy to tape or not. I usually then open this final file in Vegas Pro8, mix my final audio and set markers used for chapter markers in DVD Architect 5.0. I then render output from Vegas to a Bluray preset 1440x1080 MPEG2 HD with embedded markers and take into DVD Architect 5.0 for authoring of the Bluray disc. Frankly this Bluray disc looks just like the tape playback from the FX1 or output from the SR11 or XR500. Not all encoders are the same. Those in Edius and Vegas are in my opinion a lot better than Nero for instance. I have no experience with Pinnacle to comment Ron Evans |
May 8th, 2009, 02:49 PM | #34 |
Major Player
Join Date: Mar 2007
Location: Eggertsville, NY
Posts: 528
|
Just to throw in my 2 cents....
Re-rendering as we are discussing it here is the process of taking a file captured in one format (AVCHD) and transforming it into another format (HDV), and, in some cases, converting it back to AVCHD again when a disk is being authored. The AVCHD compression format is intrinsically lossy and a large amount of data has been literally "thrown away" when the original footage is captured to make the video fit into a comparatively tiny and highly compressed format. Once the footage is converted, or even when it is retained in the same AVCHD format but re-rendered to new and different "groups of pictures", more of the original data is lost. The compression processes used for video, both mpeg2 and mpeg4, use "motion estimates" to describe how each macroblock differs in comparison to the frames which follow. The only frames which are truly accurate are the I-frames, and all of the intermediary frames (P and B) are nothing more than a good "guess" as to what the macroblocks will do. As the new re-compression takes place, artifacts get created, and details get lost in the estimation process. There is absolutely no way that a file which has been originally captured will retain its original content intact once it has been uncompressed / expanded, then re-compressed. Unlike lossless compression schemes such as used in .zip files like Lempel-Zev, the mpeg video schemes are truly lossy, and the loss is never recoverable. Can the loss be seen? It depends on how it is done, how many times it has been applied, what playback devices are used, how critical the viewer is, etc. Some people look at an iPod and see "perfect quality" video............. My own strong opinion is that re-compression is visible and undesirable. I try to make all color and white balance adjustments in the camcorder, and smart render (which does NOT do recompression or re-rendering) whenever possible. The fact that re-rendering takes 8 or 10X real time even on my Intel Extreme 4X3.0 GHz Quadcore is another factor, but I would gladly pay the time panalty if I had a video quality improvement. Such is not the trade-off in this situation however. Re-rendering costs both time and quality. Larry Last edited by Larry Horwitz; May 9th, 2009 at 06:48 AM. |
May 8th, 2009, 03:19 PM | #35 |
Inner Circle
Join Date: Mar 2003
Location: Ottawa, Ontario, Canada
Posts: 4,220
|
Larry , technically you are correct in that every re-render has the possibility of introducing losses. The real issue is are these visible and does the process improve the editing and final production to the point that they are worth the risk of introducing artifacts. Have you used Cineform or Canopus HQ in your comparisons? Using Canopus HQ and rendering to MPEG2 HD at 30mbps max. 25mbps average,18mbps min, the Bluray's are to me just like the originals as seen on my 42"Panasonic 1080P Plasma. I can also tell you that the one render I tried using Nero to AVCHD was awful.
Ron Evans |
May 8th, 2009, 05:46 PM | #36 | |
Trustee
Join Date: Mar 2005
Location: Coronado Island
Posts: 1,472
|
Quote:
I can start with 16 mbs AVCHD, convert to Cineform DI, do extensive editing, graphics, corrections & tweakings, then render it out as 30 mbs Blu Ray, burn the BD & play absolutely broadcast quality HD imagery, certainly perceived as being as good as the original footage. The digital equivalent of "free lunch"
__________________
Bob |
|
May 8th, 2009, 06:08 PM | #37 |
Regular Crew
Join Date: Mar 2009
Location: Winnipeg Canada
Posts: 68
|
It depends on what you're talking about. Nero does a good job at what now seem to be called "smart rendering" (a term actually started by Ulead). But then smart rendering isn't RENDERING at all. It's simply copying, or at worst, changing the container in which the format is contained. As for actual rendering.... Nero stinks... and rendering is REQUIRED any time you change something at the frame level so that those frames have to be re-written. A cross fade for example will require a full render because each frame through out that fade must be fully re-written. If you do nothing but very simple 'cut and butt' editing then a "smart render" system will work well for you because all you need to do is copy your work, but if you do any detailed editing then you need an editor that can do REAL rendering.
|
May 8th, 2009, 07:44 PM | #38 |
Major Player
Join Date: Dec 2006
Location: Traverse City, Michigan
Posts: 416
|
It seems that the more I learn, the more I don't know. Thanks to all of you patient people for your explanations.
Another question, if I may: If I were to use Edius Neo, does this program do the whole job? I mean, will it capture both video and audio, allow me to do my editing (video and audio), and then allow me to burn a DVD in the AVCHD format? I have tried to figure out Edius on its webpage, but each time I read it, I seem to read something different into it each time. It is a real burden being me sometimes. Mike |
May 8th, 2009, 07:51 PM | #39 |
Inner Circle
Join Date: Mar 2003
Location: Ottawa, Ontario, Canada
Posts: 4,220
|
IF you want to do simple cuts editing then "in camera" editing is even easier. Just divide clips where you want to and create a playlist in camera. I believe Sony's external burner will even create a disc for you of this playlist no computer needed and no rendering. I think it might even put a simple menu on the disc. The moment one needs nice menus, any colour corrections or effects one is likely into a complete re-render. Then one needs a real editing program and high quality encoders. This is where the intermediate files have real value.
Ron Evans |
May 8th, 2009, 08:09 PM | #40 |
Inner Circle
Join Date: Mar 2003
Location: Ottawa, Ontario, Canada
Posts: 4,220
|
Mike, Neo 2 will be out sometime in May with more support than the current version and I think will ship with Panasonic cameras like the HMC150 etc. Go to the Canopus Forum site and check out the Neo section for more information. I have the full version so do not look too closely at Neo. Generally Neo has a sub set of the full version missing things like vestorscope/waveform monitor, multicam editing and support for the pro formats like P2 and XDCam etc.
The present version will certainly do most of what you want including the HQ codec converter etc. I am not sure how feffective it is at outputting AVCHD but NEO 2 will have Bluray support( as well as support for the hardware encoder board Firecoder Blu which has the Spurs Engine for encoding) so should do AVCHD as this is a part of the spec. I think the site shows price to be $199 sometime in May. Ron Evans |
May 8th, 2009, 09:46 PM | #41 | |
Regular Crew
Join Date: Apr 2009
Location: Nelson, New Zealand
Posts: 63
|
Quote:
Cheers Bryan |
|
May 8th, 2009, 09:52 PM | #42 |
Regular Crew
Join Date: Apr 2009
Location: Nelson, New Zealand
Posts: 63
|
Hi Dale just as an after thought I did try the "same as first file" but found it did a re-compression at a much lower bitrate than the original so haven't worked that one out yet.
Bryan |
May 9th, 2009, 06:43 AM | #43 | |
Major Player
Join Date: Mar 2007
Location: Eggertsville, NY
Posts: 528
|
Quote:
Regarding Nero, my experience was the same as yours. Nero, if forced to render, has low quality codecs and the results are indeed poor. Their smart render is my only reason for using Nero, as it seems to work quite well. Regarding Edius (as well as other transcoders to HDV such as Vaast Upshift and Voltaic): Yes, I have used them, and I dislike their long conversion times and drop in quality compared to original footage. I personally do not often author BluRay disks, although I have burners and blanks here to do so. I virtually always author AVCHD disks, and therefore have no interest in mpeg2 conversion, either for ease of editing or for distribution. (I have created many hundreds of HDV/mpeg2 disks from HDV content in the 2003-2008 time frame before I switched to AVCHD with HD DVD format, and found this to be a superb way to retain quality) For AVCHD content, I capture in AVCHD, edit in AVCHD, and publish in AVCHD, and re-render this fragile format as seldom as possible. If a lot of editing is neccesary, I generally use Vegas Pro 8, and find that it handles AVCHD quite well. I used Cineform with Premiere a few years ago when it first came out and found Premiere was just full of problems, so I have not returned to using either Premiere or Cineform in recent years. It is probably rather clear I personally dislike the use of intermediate file forms and conversion times, so I try to avoid both. With all due respect to others and how workflow should be accomplished, I need to point out that I am NOT a professional, do relatively light editing work, and use only a small subset of the features of an NLE most of the time. I do, however, spend a lot of time observing, measuring, and comparing frame grabs, comparing identical clips edited in different software, and have several big hi def displays, so I have strong preferences for certain workflows based on image quality. In my view, re-rendering has more than "the possibility of re-rendering losses". It introduces losses which, to me, are visible in many if not most cases. Your milegage may vary. I think it is entirely up to each user to decide whether they are able to see differences and also judge whether they are concerned about it or not. Larry |
|
May 9th, 2009, 07:57 AM | #44 |
Major Player
Join Date: Dec 2006
Location: Traverse City, Michigan
Posts: 416
|
Larry. Since you are using Vegas Pro 8, I am going to assume that you are pleased with the final video/audio product. Is this true? How is it to work with? What is its tool suite like? Better than Neros?
Mike |
May 9th, 2009, 08:07 AM | #45 | |
Major Player
Join Date: Dec 2006
Location: Traverse City, Michigan
Posts: 416
|
Quote:
Since I am not completely happy with the NLEs that I currently have, I am wondering if Neo or Neo 2 would be a smart investment. I pretty much understand that applying my footage after Neo converts it, to say Pinnacle (if it will work with Pinnacle), should alleviate any problem with losing quality due to rerendering. But the keys here are; will it work with the NLEs I already have, and will it truely contribute to a better quality video/audio product working with these NLEs? Mike |
|
| ||||||
|
Thread Tools | Search this Thread |
|