rendering time
 

Just out of curiosity and I want to know what others are getting out of their vegas 6c. I had converted m2t files on the time line to the supplied cineform codec. Time ratio I've been getting is 1:3. Is that practical??

Second question is when I am done editing and have replace files to the original m2t will it do another render before printing back to tape?

What are your specs? I am using a P4 dual core 3.2, with 4gb of ram, 10k sata os drives and my rendering time is not as good.

1 min = 3:45

This is a m2t file captured through vegas.

Second test was

Connect HD avi's to avi intermediate = 4 to 1

DSE, if you read this. Will gearshift help to switch out avi's for m2t? When I render avi to m2t it goes to 8 to 1.

Thanks,

Jon

You got a more beefy computer. My specs are P4 2.8 HT 2GB RAM and a 500GB Harddrive.

Wow, twice the processor with twice the ram and yours runs faster? I am going to do a fresh load on my pc with just os and vegas. That will be a true test, considering I have tons of apps on my pc. I will post what happens.

I just printed to tape for the first time last night, works perfectly.

Jon

halo, i don't quite know what you're asking about but here are my spex&times:

spex:
-dual mp2800
-3gb ram
-558gb raid0
-etc.

using gearshift i converted 6.5 hrs of m2t's to DV proxy overnight (somewhere around 10 hrs).

after editing i swap back the DV proxies for m2t files and render to DVD or HD output. a 20min DVD takes about 3 hours. i'm currently rendering a 3.5 hour DVD from m2t's and that's taking about 16 hours so far and it's still got 3 hours to go.

thus, rendering times.

Yi Fong Yu,

Would you test your avi proxies with a one minute section, render it to m2t and see how long it takes. I am experiencing 1 minute takes 4 minutes to render. This doesn't change if I render it has an m2t or avi intermidate.

Thanks,

Jon

I’ve had the results of my tests posted to my web site for while now. 3x to 4x is about right for rendering m2t to CineForm on a single core P4 3.0Ghz. It drops down to 2.5x for my AMD Athlon64 X2 4600+ dual core with 2GB of memory.

~jr

Thanks! I need to run a test with it at 720p versus 1080i and see what happens.

Jon

I think is gonna be longer. Fx1/z1 are shot in 1080 so now your computer need to recalculate to 720p.

But yes I am getting a 1:3 ratio when I render m2t to the supplied cineform codec in vegas 6 again my specs are:
P4 2.8 HT
2GB of RAM
and a dedicated HD of 500gb using firewire400

I must be doinging something wrong then. Considering my specs are much higher than your own.

p4 3.2ghz Dual core, 4gb of ram, 10k sata OS drive and more internal sata drives, no external drives(ie firewire).

I think it's time for a reload, with just windows xp pro, hd link and vegas, nothing else.

Jon

I just realized there are a few unanswered questions here:

Yes, it will do another render which is why you should NOT replace the CineForm AVI files with the original M2T files. There is no point in it. Vegas will re-render either way and some say you will get a better re-render from the CineForm files. At a minimum it is guaranteed not to be worse so why bother swapping. (you gain nothing)

No it will not for the same reason as given above. There should be no need to swap out the CineForm files. They are the same quality as the original M2T files. I just did a 1 minute test render from M2T to M2T and from CineForm to M2T and both took about 2:20.

And to address Jon’s new question:

You are getting about the same time with your P4 3.2 dual core as my P4 3.0 single core. Fred was not as accurate. He simply said 1:3 and he has a P4 2.8. I don’t think you will see much of a difference between 2.8 and 3.0 P4 so I assume his actual times are closer to mine and yours. Until he posts a more accurate timing you can’t be sure that his computer is really faster.

Also, I hate to say this, but the P4 dual cores are a lot slower than the AMD dual cores and some have even reported that they are slower than the P4 single cores! Your PC seems to bear that out. So it might not be anything you are doing.

~jr

When I render out DV it is so much faster than my P4 3.2ghz. It's definately much faster. I just can't bring myself to buy AMD. I am going to do a fresh reload, since I have a million things loaded on the pc.

Thanks,

Jon

I understand how you feel cuz’ I gotta tell ya’. I have never owned an AMD before. I only bought Intel and when I started building PC’s I only built Intel. My thinking was, “why buy a wanna-be when you can have the real thing”. But when I was researching parts for this latest PC that I built and the facts were as plane as day. AMD trounced Intel by a wide margin in every performance test. If it was a small amount I would have stayed loyal to Intel but it was no contest. The AMD dual cores were significantly faster and reports said that Intel wasn’t going to catch up until 2008.

As I started to research more, I found out that Hyper Threading was not a new breakthrough to improve performance. It was a patch to fix a design flaw in the Intel architecture. Apparently the Intel pipeline is so long that when they get a cache miss it takes a lot of cycles to recover. By using two threads, any cache miss would switch to the next thread which would already have instructions in the pipeline ready to go. The reason AMD didn’t implement HT is because they have a short pipeline and don't need this "insurance". (i.e., they don’t have this design problem).

I also found out that AMD chips process more instructions per clock cycle than Intel which is why their chips can run at a slower clock speed yet get just as much work done as an Intel at a higher clock speed. That’s why they went to the naming convention instead of reporting their clock speed. (so that customers could compare “relative” speeds). That say’s that the AMD X2 4800+ is as fast as an Intel dual core 4.8Ghz (if such a chip could be built)

Slower clock speeds also mean a cooler running chip. My AMD dual core doesn’t get anywhere near as hot as my old Intel P4 on long renders. In fact, it never goes above 55 Centigrade even after hours of rendering.

After reading all this, there was no denying that Intel may have been first, but AMD had a superior architecture which got more work done and ran cooler. These are exactly the attributes needed for video work and especially rendering (as you are finding out) so… I jumped ship and got an AMD. I am so impressed with my AMD dual core, I don’t think I would ever buy Intel again.

I’m not trying to convert you or anything, but I’m just laying out the facts as I found them to give you an idea of why this long-time Intel user finally bought an AMD. You might want to rethink your strategy and get the best hardware for the job. Today, that would be AMD X2.

~jr

Don't need to swap back to m2t when I am printing back to hdv? Isn't there be a loss in quality?

No, there is no loss in quality if you are using the CineForm codec. The intermediary is not a proxy. It’s a high-quality render-ready copy of the original m2t file. You can read about the quality on CineForm’s web site here.

The reason is because the mpeg transport stream (m2t) is extremely lossy. It only records one full frame of information for every 15 frames. The other 14 frames are predictive and delta information. The CineForm codec is totally full frames. Think of it this way, you are encoding a high compression stream (m2t) with a low compression codec (CineForm). There is no loss because the lower compression codec is totally capably of accurately representing the highly compressed stream and then some!

Just to summarize, if you use a CineForm intermediary or Sony YUV intermediary, you can throw the m2t file away. It is no longer needed and you should just render from the intermediary file.

When I capture with Connect HD, I capture direct to a CineForm AVI and there isn’t even an m2t file on my hard drive. (no need for ‘em)

~jr

Does this applies to the supplied Cineform codec on vegas 6c. I am still debating weather or not to get the full blown Connect HD. Again thanks for everyone's contributions!!

Does this also applies to the supplied cineform codec that comes with veags 6c? I am still debating if I still need the full blown Connect HD.

Thanks everyone!

After a bunch of research, you are correct the AMD X2 definately outpreforms intels p4 dual core 3.2ghz. I priced the motherboard and processor and it comes to about $800. I can use the rest from my other system, meaning graphics, hard drives ane memory. I might just do it :)
Thanks for the heads up!

Yes, you can render your M2T using the CineForm codec in Vegas 6 and render right from that codec to your final format (i.e,. back to tape or to DVD). The whole idea behind GearShift is to take advantage of this fact and automate the process for you if you have a lot of M2T files to process.

~jr

Sad but true, I know. I have the specs for my system on the PC Equipment page of my web site. The ASUS A8N-SLI Premium ($167) and AMD X2 4600+ ($547) are $714 at newegg.com. They have been solid as a rock for me. If you got the top-of-the-line AMD X2 4800+ ($630) it would be $797. Something to consider.

~jr

and for those interested in building new amd dualcore systems, there are socket 939-based Opterons that can OC'd to kingdom come (4800+ levels) for a price closer to x2, 3800+. FYI.

re: amd. they were kicking ass waaay back in the late 90s when the 'k7' architecture came out and performed faster clock for clock besting equivalent p3's. k6 was horrible, but k7 was when amd became a serious contender/alternative to a lazy intel. ah those were the days =).

Connect HD has it advantages, first is encoding performance. Connect HD will convert M2T to CFHD 3+ times faster than Vegas/Gearshift. Then there is the quality controls, and the huge range of capture features supported. All of this info is available from www.cineform.com. The free trial will help you decide.

From what's been said here, it sounds like CineForm's codec might possibly get confused for a lossless codec. It isn't.

As a practical matter, rendering final output, from a CineForm copy of the source M2T original, will work just fine for almost anything, but if you want the absolute best possible quality render, the original MT2 source is the only choice between the two (simply because it is the original, unaltered version). It is unlikely anyone will visually perceive a difference though.

I only mention this, because not making the seemingly subtle distinction, between absolute and practical, can potentially come back to haunt someday (so to speak). Most humans would not be able to tell the difference between a 320kbps MP3 audio encoding and it's source either, but I doubt anyone in the audio profession would want to confuse the MP3 copy for a "perfect" copy.

For editing purposes, CineForm's codec is, for HD, much like using a DV codec for SD, in that it has most of the advantages of working with uncompressed data, yet with much smaller file sizes, with only extremely minimal loss in quality.

This is a good distinction to make. I don’t believe anyone said the CineForm codec was lossless and I would not want to mislead someone into thinking it was. What I was trying to point out was that it is less lossy than M2T. So I don’t agree with your MP3 analogy. If I read it correctly, you are comparing CineForm to a 320kbps MP3 and M2T to an original Uncompressed wave file. M2T is no where near a “perfect” uncompressed copy. It is highly compressed. More so than CineForm.

CineForm uses variable wavelet compression at a rate of 5:1 to 10:1. An MPEG2 Transport Stream (M2T) uses compression of about 20:1! When you convert an M2T stream to CineForm you are going from 20:1 compression to only 10:1 compression (or even as low as 5:1 compression in some cases). You are going from a lower quality codec to a higher quality codec.

To more accurately state your MP3 example I would say that M2T is a 128kbps MP3 and CineForm is a 320kbps MP3. What you are doing then, is recording the original at 128kbps MP3 and encoding it as 320kbps MP3 later to work on it. You are right, you will probably not perceive any loss in quality when rendering a 128kbps MP3 (M2T) to a 320kbps MP3 (CineForm)

I do agree with you that all encoding suffers some loss. So even going from the lower quality M2T to the higher quality CineForm may lose something in the translation. Like I said, I don’t even capture M2T files half the time. I capture straight to CineForm.

~jr

stepping back a bit. are we really going to be stuck with .m2t as "the" sole source codec for HD-based content we want to record? are there future codecs from other competing HD taped based transports that yield better/high quality results than .m2t's?

There already are but not from tape based transports. The Panasonic HVX200 does not use M2T (which means it doesn’t record HDV since M2T is the HDV standard). In fact, it doesn’t even record HD to tape. It only records it to P2 memory cards. They call their version DVCPRO-HD and it records with less compression but less compression means a higher data rate (up to 100Mbps), so it can only record to solid state memory cards or hard drives.

BTW, we are not “stuck”. HDV M2T is very high quality. The “quality” of my HDV footage far surpasses ANYTHING I’ve shot on SD. I apologize if this thread is giving you the impression that HDV M2T is low quality. It is not. It’s just that CineForm is higher which is why it is recommended that it is safe to convert to CineForm and render from there. You should not read anything else into it.

Perhaps my analogy of M2T equating to 128kbps MP3 is what gave you that impression. I picked that rate because that is said to be CD quality which most people feel is high quality. Perhaps it was a bad analogy.

~jr

Too much technical stuffs here. So heres what I did and would like to share. I did it as simple as can be. Downloaded the triall version of CF and did a 15 sec clip. Edit clip just by doing simple crossfade transitions. Then printed back to HDV tape. Not bad, I must say the quailty on my HD TV I can even tell, there was no difference from the original footage. But the ratio I got is another story 1:8 ratio when rendering. I need a faster computer!!! Atleast Cineform HD converted the m2t file on the fly and it has scene detection!!! I am definelty picking one up after the trial period is over.

Too much technical stuffs here. So heres what I did and would like to share. I did it as simple as can be. Downloaded the trail version of CF and did a 15 sec clip. Edit clip just by doing simple crossfade transitions. Then printed back to HDV tape. Not bad, I must say the quailty on my HD TV I can't even tell, there was no difference from the original footage. But the ratio I got is another story 1:8 ratio when rendering. I need a faster computer!!! Cineform HD converted the m2t file on the fly and it has scene detection!!! I am definelty picking one up after the trial period is over.

So in summary:

- re-encoding will always carry some loss, even if is to a less compressed format.
- But that in practical terms, one is unlikely to notice this loss when using Cineform as an intermediate render.

But what I’m still not sure about is… it’s not just re-encoding we are dealing with, it is also editing.

So, comparing the 2 end-to-end scenarios. And let’s assume the output format is m2t, printed back to tape:

1. Edit the M2t (either natively with fast PC, or using a proxy). Single render from m2t back out to new m2t.
2. m2t to Cineform (assume I do this in Vegas). Edit Cineform, render back to m2t for output to tape.

I’ve read that there are concerns with having Vegas edit m2t directly, specifically with color correction. So it is possible (or correct) that flow 2 could actually be better?


Mark

hi john,

i realize m2t's are nice. having just done a wedding shot on XL H1 =).

i just wanted to know if there are alternative tape transport-based codecs that make low-end machines easier, taht's all =).

I use Connect HD(avi intermediate) with Vegas and it makes for a great editing experience. Trying to edit m2t stream would make me switch back to DV.

Try it out and see what you think. They offer a 15 day trail.

Jon