Good or Best for Cineform Render?

johnmeyer wrote on 1/11/2006, 12:52 PM
The Vegas 6.0c "HDV 1080-60i intermediate" template for creating Cineform intermediate files from m2t captures uses "Good" for the "video rendering quality."

I assume that Sony engineers chose this correctly, but is there any advantage to using "Best?" I am pretty sure the answer is no, since there is no down-sampling, which seems to be where Best is needed, but I thought I'd ask. I used Best last night when creating intermediary files for my first major HDV project. It took a long time, and I would like to reduce that time as long as quality won't be noticeably affected.

Of course, when doing my final render to SD MPEG-2 files, I have to use Best because the 1440x1080 is being downsampled to 720x480.

Comments

johnmeyer wrote on 1/12/2006, 6:23 PM
I've never bumped before, but here goes ...

Bump.
Padre wrote on 1/12/2006, 8:34 PM
I use best for all my renders.... doesnt do me any harm...
fldave wrote on 1/12/2006, 9:06 PM
I use Best for everything, also. Since editing with Vegas isn't my day job, I have lots of time. If my day-to-day income depended on Vegas, then I would do some tests with various clips to see if there is an effect.

I guess the best way to tell is if you can tell a difference on the preview monitor between Best and Good, you should use Best.
Serena wrote on 1/12/2006, 9:20 PM
I suggest that a default isn't necessarily chosen to produce an optimum result. Probably produces the quickest result for users not critical about the outcome (Note: this is intended as a generic comment about software). I use "best"
johnmeyer wrote on 1/13/2006, 10:43 AM
I suggest that a default isn't necessarily chosen to produce an optimum result.

Serena, good point. This is actually what motivated me to ask the question. The best example is that the "Default" template for MPEG-2 produces results that are awful beyond description. It is NEVER the right render option (and should be modified for future releases, but that's a story for another time).

I am concerned that the Good setting for the Cineform intermediate may also have been chosen as the default in order to get better ratings from reviewers (who seldom delve into the details of these settings, and just simply time how long something takes, which is why, I believe, that the MPEG-2 defaults are set the way they are: the settings produce very fast renders).

However, I DON'T have all the time in the world, and if I am wasting that time rendering Best when Good will do, then I'd like to know it. Sometimes render tests don't immediately reveal problems (I've done my own tests) which is why I was hoping for an answer from someone that had special knowledge of the subject.
Chienworks wrote on 1/13/2006, 11:55 AM
In general, if the frame size isn't changing, Good and Best will produce identical results. The only time Best helps is when rendering to a different frame size.
johnmeyer wrote on 1/13/2006, 7:06 PM
The only time Best helps is when rendering to a different frame size.

Kelly, thank you very much. That is exactly what I thought as well, but I wanted to hear it from someone else that "knows the score." Everything I've read about Best rendering is that it is designed to help reduce the problems when sampling down to a lower frame size, as with still pictures and, now with HDV, from 1080i down to SD NTSC.

Based on your input, I will be using the default "Good" for my intermediate renders from now on.

Of course, when I then render the final project down to SD NTSC (for delivery to 99.99% of the world that only has SD DVD players), I will then use "Best." Based on input from other users and from Sony, I will do these renders directly from the Cineform intermediates as it appears that there is no advantage -- and might actually be a disadvantage -- to replacing the intermediates with the original m2t files.
fldave wrote on 1/13/2006, 7:25 PM
johnmeyer,
Sorry, I missed the disadvantages to switching back to the original m2t for final render. Can you elaborate?
Thanks
jrazz wrote on 1/13/2006, 7:55 PM
yeah, I would be curious to know the disadvantages as well.

j razz
johnmeyer wrote on 1/13/2006, 8:36 PM
Sorry, I missed the disadvantages to switching back to the original m2t for final render. Can you elaborate?

yeah, I would be curious to know the disadvantages as well.

Well, the obvious disadvantage is that it requires an extra step.

Next, as I posted in another thread a few days back, Vegas is not optimized to edit the m2t files, and therefore I think there are times when what you see is only an approximation of what you'll get when rendering. Virtually everything I've read, and all the posts here in this forum advise to use the intermediates if you're interested in frame-accurate (or close to it) editing. Now, I can't vouch for this directly, but the concern is that if you cut on a boundary between two scenes, and then switch back to the m2t, you might get a frame or two from the previous scene. I want to emphasize that I have only read about this, but not observed it myself, so this may be bogus. If so, I apologize.

Next, Sony replied to one of my posts, in this thread (SD DVD best quality workflow) and stated: " ... you could switch back to the native HDV source material prior to final render, but the memory handling in Vegas is far better with the Cineform (or Sony DV or Sony YUV) codec, so watch out for that if you have a long timeline." Thus, the render from m2t will take longer. Since rendering the intermediate Cineform files is an "extra" step (compared to editing SD DV), it is nice to be able to get a little of that time back when doing the final render.

Wolfgang S. was extremely helpful in getting me up and running and wrote some very useful information in this post: HDV to SD Crop Procedures. He also recommends sticking with the Cineform intermediate.

The only things I've read that would make a case for going back to the m2t files involved problems with earlier versions of Vegas, and earlier versions of the Cineform codec. With Vegas 6.0c (which includes the latest Cineform codec) these problems (which included color shifts) have apparently been fixed.

Hope that helps!



fldave wrote on 1/13/2006, 9:08 PM
Again, if I understand the reasoning, it's a matter of time vs. a matter of quality.

The Gearshift utility, which I don't have, I do it manually, always switches back to the original m2t for final rendering. Spot could probably address this better.

I think this goes back to the 4:2:0 to 4:1:1 issue.

I would hope, and expect, after my editing with the intermediate to get the best color correction/effects, that Vegas can recreate my desired output best with the true original source file.
johnmeyer wrote on 1/13/2006, 9:24 PM
Color could be an issue.

As for my original post, I'm not sure the time difference between Best and Good when doing the Cineform render is significant. The original issue may be moot.
Jayster wrote on 5/10/2006, 2:30 PM
There are two different good/best settings in this case: 1) the one we all know, which is the Vegas setting, and 2) the one for the Cineform intermediate itself, which is only accessible if you go into "custom" dialog in the "render as" dialog.

I am doing a rather complex render with lots of effects, and it's a multicam with a mix of HDV and DV footage. I want to render the final product to a Cineform intermediate avi. From there it'll get converted to multiple formats (PAL, NTSC, WMVHD, etc.). By "rendering" to the intermediate, I'll only have to do the hard rendering of effects, transitions, etc. one time.

For sure I should use the "best" setting in Vegas, but what about the good/best setting for the Cineform codec? Anybody know how significant this is?

farss wrote on 5/10/2006, 3:26 PM
It does give better quality results however the file size is larger i.e. there's less compression. According to CF, for most footage Good is adequate, I guess if you're going to a film out you'd use Best.

Bob.
johnmeyer wrote on 5/10/2006, 4:52 PM
There are two different good/best settings in this case: 1) the one we all know, which is the Vegas setting, and 2) the one for the Cineform intermediate itself, which is only accessible if you go into "custom" dialog in the "render as" dialog. ...

I am not sure what "best" setting in Vegas you are referring to. The only one I know about is the preview window and it doesn't have anything to do with rendering. The only one I know about in the Render As dialog is when you click on Custom and then select the Project tab. As the posts earlier in this thread indicate, this can be left at Good because going from m2t to Cineform AVI does not involve up- or down-sampling the video (which is what requires "best"). I just looked on the Video tab and clicked on the Configure... tab next to "Video Format" to see if there are more option for the Cineform codec. However all it did was give me a copyright notice.

So, you've confused me about the "one we all know about" because I'm not sure what that is.


Jayster wrote on 5/10/2006, 6:11 PM
Sorry if my explanation wasn't good.

The "best" that I said is "the one we all know about" is the same one you just wrote about, in the Render As -> Custom dialog (in the project tab).

This same option appears in the project properties dialog, on the video tab.

The second, less obvious Good/Best is on the same Render As->Custom dialog, but it's in the "Video" tab. It says "Cineform HD Codec V2.5" and there is a button that says "Configure."

I see now I was wrong about this saying Good/Best. I was at work instead of home, and remembered it wrong. It says "Encoding Quality" and has options Low HD, Medium HD and High HD. (I thought a previous version of the Cineform had said good/best, but I probably got that mixed up).

In my project's case, I'm not using .m2t at all. I have several Cineform avi clips mixed with standard DV avi clips, some .mov clips, stills, and lots of effects. So I want to render the whole project to an intermediate format using the Cineform code. With stills and mixed resolution source files I figure I should use "best" in the rendering quality, and what I don't know is if the Cineform codec should be set for "High HD" or if "Medium HD" is good enough.

The intermediate I produce will be converted to a variety of formats for distribution, which explains why I want one big intermediate file (so I only have to render stills and effects once, then just a matter of converting/rendering it to PAL, NTSC, etc.).
johnmeyer wrote on 5/10/2006, 7:54 PM
Ah, you must have the full version of Cineform installed. The one that comes with 6.0d doesn't have that setting. You just get the copyright notice, as I discussed earlier.
mbryant wrote on 5/11/2006, 1:40 AM
The Gearshift utility, which I don't have, I do it manually, always switches back to the original m2t for final rendering. Spot could probably address this better.

I don't think this is quite correct. Firstly with Gearshift the user decides when to switch back and forth and what to render from.

You can use Gearshift in one of 2 modes:

1. DV proxy only: In this case you edit a DV proxy, and while you don't have to, it is best to swap back to the original m2t for final rendering.

2. DV proxy and Intermediate: In this mode the swapping is between the DV proxy and the intermediate.. you use the intermediate for the final render. (If you create an intermediate, Gearshift no longer swaps to the m2t, instead it swaps between the intermediate and the proxy).

I’ve been under the impression that if you’ve created an intermediate you might as well use that for the final render – that there is no advantage of swapping back – but I’m not 100% sure if this is correct or not. But given that I can’t distinguish the intermediates from the original in terms of quality this seems to make sense. In fact I’ve wondered if the quality could be better rendering from the intermediate. It seems possible to me that going m2t to cineform to m2t could be better than a straight m2t to m2t render – consider it like a special kind of 2 pass render, where the intermediate stores more information and thus the final render to m2t is better?

Mark
PeterWright wrote on 5/11/2006, 2:07 AM
> "I’ve been under the impression that if you’ve created an intermediate you might as well use that for the final render – that there is no advantage of swapping back – but I’m not 100% sure if this is correct or not."

Correct Mark - in fact there is really no point in creating Intermediates if you're going to use the original m2t files for rendering - better to use DV proxies, which take up a third of the space and are much easier to edit with, particularly on a slower computer.
Jayster wrote on 5/11/2006, 2:50 PM
Ah, you must have the full version of Cineform installed. The one that comes with 6.0d doesn't have that setting. You just get the copyright notice, as I discussed earlier.

Yes, I am using the full version. I haven't heard much commentary on what the impact is of using it's "HD High" setting instead of "HD Medium". Cineform says the files get bigger. Does anybody know what it buys you (for complex projects where you would need to use the best quality render)?
Jayster wrote on 5/11/2006, 3:28 PM
WIth some digging on the Cineform site, I found (I think) the answer to my questions and some interesting info about converting to the codec.

Cineform generally recommends you use the Mid range file size when doing m2t to CFHD conversions. This is for a few reasons:

That last part was talking about Aspect for Adobe PP, but would probably hold true for any NLE. And here's a comment they wrote about converting to CFHD with the full version vs. the codec that comes with Vegas.

For those who are interested and are technically minded, Vegas only uses the Video for Windows version of the CineForm codec while HDLink uses the DirectShow version. The DirectShow version is a bit more sophisticated about the manner in which we can code our files, so we get approximately a 30% reduction in file size with HDLink. These smaller files also perform "faster" on the Vegas timeline. There is no quality disadvantage with the VfW codec, and the bitstreams are 100% compatible.

If what they say is still true (that the full version makes intermediates about 30% smaller), you'd get faster previews than the stock version in Vegas. But that would only apply for something you converted with their HDLink utility, probably not for something you rendered from Vegas, regardless of whether or not you own the full version of Cineform. Because Vegas always renders to CFHD with the vfw codec.