This is used mainly with variable bit rate encoding. The first pass is used to decide what parts of the video have more detail and need more bits, and what parts cat spare bits they don't need. The second pass then does the actual encode more intelligently because it's already known where the bits need to be concentrated.

Since it does run through the entire project twice, rendering time is just about doubled.

Is there a particular type of video source that one would use this feature for? For example, I typically capture DV footage, and convert it to MPEG2 with the results looking absolutely fine. Actually, I've never noticed an area that could use improvement.

So, is the quality increase all that noticable and worth the double time need for rendering?

"For example, I typically capture DV footage, and convert it to MPEG2 with the results looking absolutely fine. Actually, I've never noticed an area that could use improvement."

Its not what your eye can see but more importantly, compression effenciy. While some things look great to your eye, they could be improved with better compression. For example, if you have a still shot using CBR, the same bits will be used for high action scenes. With 2 pass VBR, the bits that were overly allocated to the still shot in CBR mode, will be used in the high action scene to make it look better.

Plus one other huge benefit of 2 pass is better target file sizes. CBR doesnt always end up with the exact target file size you want. With 2 pass, the target size is more likely to end up what you are shooting for in file size.

So, in consideration of what you are saying, do you usually do "2 pass encoding"?

Or, do you sometimes not? What criteria do you use to decide?

Yes I always do 2 pass encodes. Making sure file size is correct and quality is intact.

The only time I ever do a one pass is for a rough draft for a client to sign off on.

It's not actually a new feature of 5b. You could do it in previous versions, but it was not enabled by default in the templates (don't know why). The improvement in 5b appears to be a speed increase (always welcome), although I have never considered the MainConcept encoder to be slow for what it does, especially in comparison to TMPGEnc.

Acidsex, I really liked your informative, " With 2 pass VBR, the bits that were overly allocated to the still shot in CBR mode, will be used in the high action scene to make it look better." - This is an excellent descrip[tion as to what is happenning; why one should think about using it; how you apply it - if you wish - and in what cricumstances you do it. . . . I have found this whole thread refreshing and just plain old helpful .. THANK you . . Now I understand 2-Pass and why I would want to use it.

Grazie

Glad I could help. I dont get much of a chance to offer anything informative to the forum but it looks like theres a first time for everything. :)

The high end encoders can run more than 2 passes as well, upto something like 400 passes and spend weeks optimising an encode. Main area where this seems to bring a benefit is for high quality and low bitrates.

There is more to this than I/we think . . eh?

Grazie

I tested a 1.5 hour avi file. I'm not sure if I did something wrong, but here are the results:

Took 4 hours to render. That seems as expected. To get this to work I used the DVD Architech template and went to custom and then checked the "2 pass" to ON.

Once the file was created, I dropped it into DVDA2. However, DVDA2 takes the file and re-renders it again! WHY?

Additionally, I don't know if it's just me, but the results did NOT seem as good.

Where have I gone wrong?

guns1inger,

You say that 2 pass is now selected by default? In other words, renders will take twice as long unless I un-select it? Where did you notice it defaulted to enabled? I don't see that because I had to enable it with a test that I performed. However, my test results were NOT pleasing.

A 1.5 hour avi is not going to fit on a DVD at the default bitrate using PCM audio. You need to either drop the bitrate to get it to fit or else goto ac3 audio.

farss,

This I do know. However, I'm still hoping someone can explain the problem noted above in previous message.

clearvu - I said I wasn't sure if that was one of the improvements or not, as I was surprised that it was not enabled in the VBR templates to begin with. One of the first things I did in Vegas 5 was create new templates with 2-pass enabled, as there is little to be gained by running a single pass VBR encode.

I am surprised that you believe the output to be inferior to CBR or single pass VBR. If I have less than an hour of footage to go on a disc, my default template sets high 8000, low 4200, average 6000. I generally end up with an average of 5990 - 6000 at the end of the encode.

A lot of what I do is dancing footage, with a lot of movement and action. To get the same results out of CBR I have to use stupidly high settings and waste a lot of room during the quite pieces. I have always used 2-pass out of Vegas and been more than happy with the results. Until Vegas allows for setting seperate encoding settings for different parts of the timeline, this is the best compromise for me.

When you say the results were 'not pleasing' what do you mean ?

I'm going to do an actual comparison of a render with and without the 2 pass enabled.

In the meantime, from what I could see, it LOOKED as though things were more "blocky" or "pixelated". Keep in mind that the source is from a VHS tape.

Ok, here's the verdict.

First of all, as far as the section I was watching, it did not contain that much movement. In any event, the differences are not drastic. The single pass (normal) render seems to provide a darker look. The blacks are richer looking. The 2 pass encode has what seems to be a soft white film over the picture. My wife says she thinks the single pass looks clearer. I tend to actually believe the 2 pass is clearer. Suprisingly, from what I can tell it looks a slight bit less pixelated. However, it might just be because of the "white film".

The conclusion is that, in this case, the differences are not enough to warrant double the rendering time. Maybe I'm doing something wrong?

Additionally, since I asked this question earlier with no answer to date, does DVDA2 not like the 2 pass encoding? It seems to re-render those clips. Any comments on this?

. .yup, me too . .I'd also like to hear the comments .. .

Grazie

1. 2-pass VBR definetely gives better quality than 1-pass VBR.
2. CBR at high bitrate would give you possibly the best quality and large files.
3. Encoding video with more than 2~3 passes is just wasting time & CPU cycles and not achieving better quality than the one obtained with the 2~3 passes.
Most of the HQ encoders like CCE,ProCoder,MainConcept do very good job with only 2-pass VBR, but you can tell which encoder does better job at lower bitrates such as min~1000,avg~3000,max~4000/5000.
CCE and ProCoder do excellent job at low bitrates!
One thing that I can't understand is who invented the ReRendering in DVDA.
If the encoded video is NOT DVD compliant then DVDA should NOT import it at the first place.
Of course custom settings in every encoder NOT properly configured may give non-compliant DVD video.
So, "using the presets will always give you DVD compliant video & DVDA will not recompress the video"

AudioIvan

Just got hold of 4 mpegs files from 2 video sequence projects of length 2 min and 6 min plus respectively from another video fella. Painstakingly, 2 renders were made for both projects resulting in 2 DVD compliant mpeg files for each project. The 1st render for both projects were rendered at VBR without 2-pass, while the 2nd render for both projects were rendered at VBR with 2-pass.

Looking and inspecting the mpeg files really carefully, I find that doing a 2-pass is very beneficial shall your source DV footages contains scenes of certain characteristics:

1) scenes with rapidly changing subject movements (subjects move about on-screen, sometimes fast, sometimes slow)
2) scenes with rapidly changing lighting (sometimes bright, sometimes dim)
3) scenes with lots of details

While definition (sharpness and level of details) is inherently restricted by the encoding bit rates so 2-pass really has nothing to do with it, doing a 2-pass effectively helps in combating 1 problem - compression artifacts e.g. macro-blocks/pixelation and mosquito effect. The positive result of doing a 2-pass is most apparent when your rendered mpeg is viewed on large-screen TV/projectors, low-end projectors trying to pass itself off as high-end one by utilising digital zooming, or projectors/computers with zooming turned on.

So, it is a matter of choice here: trade render time for artifact elimination or vice versa.

I guess doing multiple-pass and utilising better quatization algorithms with hardware encoders may probably be one of the main reasons why some of the VCDs widely available in Asia about 4, 5 years ago look tremendously amazing with little or no artifacts at all - unlike some of the craps video distributors churn out with their desktop workstations nowadays.

Hmm.. maybe upgrading to V5 from V4 may look like a nice idea after all.. Or maybe I should get the stand-alone MainConcept mpeg encoder instead..? ;)

A side note here: while deinterlacing interlaced DV footage when doing a final render to MPEG2 may result in mild quality hit, occasionally the quality loss seems acceptable shall you consider how interlaced video looks on progress display - pulsating background and flickles. The flickling is really obvious with generated media with straight horizontal lines and still 2D geometrical shapes. Unless you are certain your output will not be play back on progressive display, you may consider deinterlacing your final output once in a while (depending on the content of your project), even if Vegas's deinterlacing methods are not exactly that fabulous.

So, since it's possible that I've screwed things up considering, 1) my video results did not jump out as certainly better, and 2) DVDA2 re-renders the file, can someone please outline exactly how to set up a "2-pass" render?

What I did was I used the "DVD Architech NTSC Video Stream" template and went to custom and then checked the "2 pass" to ON. That's all I changed. Is this the correct way of doing it?

For what it's worth, here's a link to an 25 second clip. The upper left portion is the 2-pass version and the lower right is the normal 1-pass version.

http://members.rogers.com/webtravel/test.wmv