SonicEPM-

I think you'll have better luck communicating with PeterMac than with Elysium. You and PeterMac may be able to put your heads together and actually make an improvement in some areas of VV's encoder. Elysium sounds just really, really frustrated to the point where he can not set aside his emotions and help the situation. Only complain and insult. Sorry Elysium, but that's how your post comes across and it's not going to help the situation.

SonicEPM, I appreciate your willingness to communicate publicly and try to help. That's 10 times more than most software manufacturers.

I would just like to commend the sonicEPM for at least being involved in the forums regardless of the discussion. As for emotions translated in print, sometimes we come off reading wrong. I can only hope to begin to understand the jargon you people in the know use.
I am new to this computer editing game and have purchased the VV3. Sonic Foundry's acid program helped me with the decision after using it to make tracks for the old ways of VE . I have read quite a bit over the years on computer video editing and it seemed the expense and the time necessary to complete projects was unrealistic for just a hobby. I stayed with my used Videonics linear set up and used my SIX (you need them all you know!) hands to task and was somewhat satisfied with my work in Hi8 to SVHS format.
Then SoFo gave me an oppurtunity to purchase this VV3 at a really good price and I said what the hey, why not. I did just a few clips at first and trust me I'm never going back to linear editing again. My only hope is to get the best out of these forums and read between the lines when someone comes across a little smarter than the last guy. Of all the tools and not just in video editing, the forum is the best. Everyone's ideas working towards the end result is how we all become successful at our craft. So lets be civil in reaching these goals. The answers are there. All we have to do is pursue them!

Your probably right on that point :-)

But really, as I've said once before, I have had experience with SF in the past when I tried to be extremely helpful generating a number of renders that would have demonstrated a problem that exists with some widescreen PAL TV's and AVI DV-PAL output that would cause the mode of the TV to change from 4:3 to 16:9(or other widescreen modes) after each transition. I had expected the problem to be resolved in the VV3a point release but nothing.. I just feel I wasted a lot of my time and effort with SF who "listen" but actually can't be bothered to act. So, a loss of faith is quite understandable on my part.

Hi,

there can be no arguement that the MC codec in VV3.0a is an inferior mpeg2 convertor, especially when compared to TmpgENC.

What we all look towards SF are fast passages and "blockiness"...this is the biggest side effect problem, especially when we are trying to make SVCDS. An olympics avi snippet of speed ice skating shows it in all its glory...when the hands swing, lego blocks appear. TMPGENC does a FAR better job.

Try it yourself, or feel free to post a site and i will upload for you.

You should get together with TMPGENC, they are so proud of the products that now have their encoder, they list them on their site! Im sure theyd be up for joining your VV3. Could be a marriage made in heaven for you.

regards,
blackout
ps. i also get the error message half way thru a render that "could not be determined"....i do not get this error for the template settings, but make a few changes and its a gamble as to whether i make it past 50% render!

I guess I have to throw in my two cents worth. For my crudy source material, TMPEnc generally produces a better encode than the MC 30a version does. The amount of difference is dependent on the bitrate and other settings. The main difference across all settings though is that the colors are somewhat washed out by MC encoding. (sorry but I can't think of a better technical explanation)

I did not buy nor do I use VV becuase of the MPEG encoding. I use it becuase it is a great hardware independent NLE. I do not have high expectations of 'bundled' encoders. That said, let me add some more evaluations to the fire.

ULead DVD Workshop which is now available only as a trail and is not fully implemented has about the best MPEG2 encoding I have seen. It is fast and the colors are almost enhanced. They are certianly at least accurately reproduced. The encodes look great. That is at 8mb CBR. Now drop to a 6mb CBR encode and the ULead encode degrades very fast. TMPGenc does a better job at 6mb or when using a two pass VBR encode of the same file size.

All my testing has been done with poor video input. Either from a consumer DV camera or worse from VHS. That will produce different results than starting with high quality source material. I can firmly believe that the MC encoder might be superiour with high quality material. That is why SF has requested folks to send them the renders and the encodes for comparasion. Different source material will produce very different results. Even a minor setting will produce very differnet results.

I do not wish to be in SF's position. They have to produce an encoder that is somewhat idiot proof for a segment of their users. (hidden settings) But power users also use their great product so of course they want to play with I frames and Q settings. I don't know how you satify everyone. I for one am content to have VV produce great DV AVI's and use the encoder of my choice for the encode that I need.

At the same time, SF, if you would like someone to do specific tests and send you clips of results etc to assist you, I am happy to do so. I have a wide pipe although the clips would be obviously short for ease of handling. I have been testing a wide variety of stuff constantly. Besides, I find it fun to help. Just let me know.

K

My, my, SonicEPM - all those titles on your byline; you must be very proud ;-)

The settings I used are as follows:
Vegas 3a
SVCD: Select the SVCD PAL template and use 'as is' (VBR, Max=2376K, Avg=2000K, Min=192K)
DVD: Select the DVD PAL template and use 'as is' (VBR, Max=8000K, Avg=4200K, Min=192K)

Tmpgenc (Vers. 2.53)
SVCD: Use the wizard to select Super Video-CD PAL (CBR, 2520K) and change Motion Search Precision to Normal (default is Motion Estimate - Fast)
DVD: Use the wizard to select DVD PAL (CBR 6000K), leave Motion Search Precision at Motion Estimate - Fast.

Both encoders use same audio settings: 384 Kbits and 48KHz

VV3a took 19 minutes to encode the 2.5 minute footage into SVCD and 17 minutes to encode to DVD. Encoding took place from the timeline.
Tmpgenc took 11 minutes to encode to SVCD and 10 minutes for DVD. Encoding took place from an AVI exported from VV3a.

The differences were most marked in the SVCD renders. The DVD renders were less so, but still noticeable.

The test footage was severe. It had choppy or rippling water, vertical yacht masts and rigging which were panned across and shipyard buildings with many vertical structures guaranteed to cause flickering, if not migraine.

Tests on either encoder that allowed the bitrates to get to 8000K or beyond caused a little jumping or stuttering on playback. This occurred in both PowerDVD and on the Pioneer DVD player. There seemed no discernible quality advantage in doing so and so this setting was avoided.

I understand that the GOP structure, which is user settable in Tmpgenc, is optimised for NTSC. However, I left the settings at the default and did not change them to the PAL optimisation since I could see no way of doing the same thing in VV3a.

As I've mentioned before, the Tmpgenc renders appeared sharper and with more bite and contrast. In addition, flickering and blockiness were kept to a minimum. Nevertheless, I would expect that with less severe footage the quality of the two encoders would tend to converge on a generally acceptable standard, especially with DVD settings.

-Pete

Hi All,

Sorry I didn't jump in sooner...it took me several days to read the entire thread. ;-)

I'd like to offer a few comments:

1) As for quality, we feel that we offer the best MPEG encoder. However, we do understand that quality is measured in human terms, not scientific terms. So everyone is entitled to their own opinion. But we have had many people tell us that our encoder is the best.

2) It should be noted that each encoder handles things differently, and tweaking might be needed to fit the encoder to the type of material being encoded. If you think an ice-skating scene looks better with TMPGEnc, the same scene might look even better with the right settings in the MC encoder.

3) It should also be noted that DVD players seem to vary widely in VCD and SVCD playback. I was testing once and I noticed lots of blockiness, so I tweaked the settings and tried again, then again and again with the same results. Then I analyzed the same files on the computer in WinDVD and Media Player. There wasn't a bit of blocking, so I tried the same discs on different DVD players and the blockines wasn't there. Depending on the methods used to encode a different file, the problems you see might be in the DVD player, not the file.

4) I personally don't think it's fair to expect Sonic Foundry to devote the necessary resources to develop support for frame serving or other encoders. Sonic Foundry chose to offer users an excellent encoder, and that's what the MainConcept encoder is. MPEG has so many variables that people's results will differ and tweaking will be required, but our encoder is highly regarded for its speed and quality. SF would have a development and support nightmare if they supported multiple options in this arena. I firmly believe that the MainConcept encoder offers the best results, though tweaking might be required to get them.

That's my .08 worth! Thanks for reading it, especially considering the length of the thread already.

MCTech

Are the people claiming TMPGENC is the greatest using it for only SVCD? I've compared the two in DVD & I've never been overly impressed with TMPGENC. It is soooo sloooow. Quality is comparable to Ligos. Is everyone on the TMPGNC bandwagon because it was free? OK I did have a project for a client where they needed a vcd & the quality of MC wasn't as good as TMPGENC but I ended up doing it with Panasonic's MPEG encoder because for MPEG1 it was better than TMPEGNC. The only problem I've encountered lately with MC was on certain scenes the rendered file has a jitter effect for a couple of seconds but according to canopus that could be due to the source being inconsistant & the ADVC-100 doubling a frame on an occasion. I'm looking into that now. I'm picky about which encoder I'm using for best quality & alot of the finished product is watched on large screens & it seems overall MC has the sharpest & cleanest looking picture. Just my .02

Wow! I'm betting the flood gates open wide now. I've been reading this thread, and biting my tongue. The quality of TMPGenc encoded VCD's & SVCD's far surpasses the quality of those encode with MC.

Further, as I've posted previously, there are MC problems with audio when fading to black from static text...problems which have never been addressed...repeatable problems...problems which do not occur with tmpgenc.

But the statement that's really going to set everyone off.....

"2) ...the same scene might look even better with the right settings in the MC encoder."

Yikers!
Do we really want to start this up again?
Nearly everyone wants MC encoder INFO (see other L O N G threads on this topic).
Please provide the INFO on "tweaking" the settings with the MC encoder.
If it is in fact true, that MC encoded vcd's & svcd's can be made comparable to (or surpass???) tmpgenc files, please, please, please, provide the INFO needed to make those tweaks.

Well, the problem is that MPEG results vary depending on the nature of the material being encoded. So there isn't one particular setting that will work in all cases. What I meant was that tweaking might be required to find the settings that work best with the given material, and I think that's true with any MPEG encoder -- not just ours.

I might also point out that do really encode well, the settings will often have to change from scene to scene, and this is something that's just not possible with "home" encoders. No matter what material is being used or what the output destination is, no one set of settings is going to be perfect for the entire project.

I think what is being asked is just a brief explanation of the optional settings in the MC plug in. MC has used some different terminology than I am used to after learning to tweak the encode in TMPGenc. I adjust the DC Coefficient and the YUV settings as well as what standard of Q matrix is used. The YUV settings really help the color output. The DC helps me with the high action items by softening them instead of blocking them. The GOP structure is used less often by me.

All I think what is being asked is a two sentance explainations of each setting box on your plug in. That way I could apply my current understanding and results to your plugin. Then I can get apple to apple clips. I know that would allow me to get better results.

K

When I have to do that, I encode the scenes in chunks and then paste them together with TMPGEnc afterwards. I haven't found anything else that will paste like this. If anyone knows of something I would like to find it.

K

There is some info on the custom controls available now, in Vegas online help.

An encoding guide for the plug-in is in the works.

Also; you can use 'what's this' help on almost every control in the custom settings dialog. To do this either click on the question mark in the upper right corner of the dialog and then click on a control you are intersted in or select the control and press SHIFT+F1.

thanks for reminding us of the 'Whats this ?' tool. it took me a long time before I realized that was there :)

hate to drag this thread out any further, but to get back to the original topic, what about the SDK? is that available to anyone who wants to write a freeware plug-in? it seems like someone could use the Uncompressed output plugin as a starting point, and turn it into a frame server plug in.

at the point where the Uncompressed AVI writer would write a frame to disk, it could serve it instead. it seems like most of the work is already done, and it just needs to be re-tooled a bit. a good programmer would know what to do.

I don't think this is really about the MC plugin vs TMPG or whatever else... (obviously it is for some people) but I just think it would be a kick-ass feature to have (frame serving) and would probably only help VV thrive. the user community for VV is strong so there's no reason we can't put this together.

could someone from SF comment on the availability of:
1 - the SDK to anyone for free
2 - the Uncompressed output plug-in as a starting point
?

thanx
- ben (cheesehole!)

In case there is any question about this, you CAN, right now, create a DV or uncompressed .avi in Vegas. Just about any 3rd party encoder should be able to open and encode those files. Yes, it means an intermediate render.

A file format plug-in sdk is available free- a couple of legal docs would need to be signed before distribution (nothing nefarious, trademark and similar stuff). Interested parties may contact sonicepm@sonicfoundry.com.

As for feeding the plug-in uncompressed, that's the way it works now.

Q: If you were a skilled developer, would you be able to write a file format plug-in with this sdk?

A: Yes.

Q: Is this sdk some magical bunch of canned code that would allow anyone to hack in their own framserver?

A: No.

I have to have a large piece of humble pie and appologise to SF and MC. After using SHIFT F1 on all of the elements, I was able to add the tweaks to the MC encoder that make a difference in my test clips. The result is that the encode was clean and bright. An 8mb CBR encode stood tall with the best of them and was as good as I could find anywhere. Bravo.

I will do some testing with VBR and lower bitrates later, but the standard 8mb CBR was perfect. I will be surprized if it holds up in a VBR lower bitrate encode, but now understanding how you set up the settings and options I should be able to get good results. OK I want it real time and as good as a $5000 hardware encoder, but for now I will tip my hat and say good work with the update and appologise for having not explored the settings enough before.

K (I hate being wrong)

>>SpruceUp, I believe, required separate elementary streams. This is a render option in the MPEG>custom>systems page of Vegas. My old contacts at Spruce our no longer taking my calls...some guy named "Steve" keeps answering the phone and he always says the same thing: "Make it look like a Sunflower".
--------------------------------

LOL! I don't know what to say about that except it sounds like they're having way too much fun over there at Spruce.

btw - I figured out the elementary streams thing and it does work with SpruceUp, so I can render a file with the new MC MPEG plug in and then import it directly into SpruceUp with no intermediary steps.

the reason I screwed up before is because the new MC MPEG2 templates are named kind of weird. to me it looked like you could only render to separate streams using PAL, but with NTSC you had to render video, then audio in two passes, which is what I was doing. Why is the PAL template named 'separate streams' but the NTSC template named 'video stream'? oh well anyway it works.

kkolbo, don't feel too bad about missing the Whats this ? thing. I only found it after an MC tech pointed it out a while back and WOW it's like a whole new encoder once you find it! I think this is the first windows software I've encountered where that ? thing has been useful. maybe in the next update someone at SF can address that. Rollover Tooltips would probably work (like in TMPGEnc).

thanks for the quick response on the SDK issue. I'm going to be deep into the development side of things in the next couple months. by the end of it I might be willing to take a shot at coding a frameserve output plug if no one else takes on the task first (please? i'm just a hack! my code would be ugly and lame!)

- ben (cheesehole!)

Second report ..

I have been working with the MPEG-2 encoding from MC; a bunch of encodes side by side with TSMPG. Now that I understand the options and can tweak it the way I did TSMPG, I have to admit that I now like the MC encoder better for DVD level encodes. I have done some VBR's with lower bitrate settings and it stands up just fine against TSMPG. On top of that it is much faster. Hat's off, I am getting first rate encodes for a non-hardware encoder.

I have not done the same level of testing with MPEG-1 encodes yet but I do a lot fewer of them.

K

kkolbo, if you're out there, I revived this old thread just to ask you if you wouldn't mind sharing some of the tweaks you were able to make in order to get those great MPEG-2 results from the MC encoder. If there is another thread that you shared some info on this subject, please point me to it. I am about to do some MPEG-2 encoding with VV3a and want to get the best results possible. Your comments really sound encouraging. Thank you.

Dave,

> We are currently shipping with the MC encoder, as you all know. This is a replacement for the L**os encoder, which we were not happy with. We have been running many tests with MC, refining different functions along with MC engineering. We believe we have made significant progress with the version included with the Vegas 3.0a release. <

Back in March when you wrote the above, I was not yet an owner of Vegas Video, but by that time I had made up my mind that Vegas would be my video editor of choice. The bundled unknown MPEG encoder was not a deal breaker for me, because I knew I could render an AVI and encode it with the encoder of my choice, which was TMPGEnc. And there was always the chance that the unknown Main Concept encoder might turn out to be good.

With respect to this thread, I do think that it would be a good idea to eventually add frameserving to Vegas, but I don't think it should be done in a hurried manner, and some thought should be given to how it is done. Vegas' software architecture currently has many advantages, and those should be preserved. For the time being at least, I can do without frameserving. And I am aware that elsewhere frameserving has been used in the process of ripping commercial DVDs to strip them of their MacroVision protection and copy them to other media. But I believe that the supporters of frameserving here in this forum have good motives and wish merely to facilitate the use of third party MPEG encoders.

I think that MPEG encoders continue to be a concern here because the Main Concept encoder is something of a "dark horse." There are several important comparative reviews of MPEG encoders on Internet web sites, and Main Concept is conspicuous by it absence in those reviews. That absence does not inspire confidence in Main Concept as an MPEG encoder of choice. Let me give a couple of specific examples.

The popular VCDHelp site compared MPEG-2 encoding for SVCD with TMPGEnc 12j, Cinema Craft Encoder 2.50, Ligos LSX MPEG Encoder 3.5, and Nero 5.5 with its MPEG2/SVCD/DVD Plugin. Cinema Craft had the highest quality images, but TMPGEnc was nearly as good, and was the "winner" based on it being free. Main Concept was not rated or even mentioned. MPEG2 encoding for DVD was not tested.

http://www.vcdhelp.com/comparison.htm

That same comparison tested MPEG1 encoders for VCD, including TMPGEnc 12a, Cinema Craft Encoder 2.50, BBMPEG 1.2, Panasonic Mpeg Encoder 2.51, Ligos LSX MPEG Encoder 3.0, AVI2VCD 1.33, and Xing Mpeg Encoder 2.20. The "winner" again was TMPGEnc. The Ligos LSX MPEG Encoder 3.0 was rated "Bad" for Blockiness. Interestingly, for MPEG-1 TMPGEnc beat Cinema Craft for Blockiness (Great vs Good) and Sharpness (Excellent vs Great). I guess that is a kind of David vs Goliath thing.

TECO Ltd. tested Canopus' AMBER DV2MPEG converter, APOLLO DVD-er, BBMpeg 1.2, Cinemacraft Encoder Lite and SP 2.0, Pinnacle DV 500, Dazzle Digital VIdeoCreator II, Darvision DVMpeg 5.0, PixelTools ExpertDVD, FutureTel NS325, Ligos LSX Encoder 3.5, Digigami MEGAPeg, Terran Media Cleaner Pro 5.02, KDD Labs MPEG Conversion Studio 1.1, Vitec MPEG Profiler, Heuris MPEG Power Professional V2.0h, Darvision MPEGator 2, Sigma Designs Realmagic DVR, Canopus SoftMPEG, Tsunami Encoder Beta Version 1.2e, Matrox RT 2000, and Vitec MPEG2 tools. But in that long list, Main Concept did not show up.

http://www.tecoltd.com/enctest/enctest.htm

If I had been looking for a new encoder partner, I probably would have tried to work a deal with Cinema Craft, who have a proven track record. It would be interesting to see a comparative MPEG shootout between Cinema Craft and Main Concept and TMPGEnc. If Main Concept could stand up to that kind of competition, they could shake some of this "new kid on the block" image. Just my opinions. I am optimistic about Main Concept, but still a bit skeptical. At least Ulead also saw fit to make a deal with Main Concept. We don't have the benefit of seeing the results of the tests that Sonic Foundry did with the Main Concept encoder, however I do take your word that they were good. But if Main Concept wants to be a "player", they need to make an appearance on the Internet playing field. For a few more "playing field" sites, see:

http://www.homepages.hetnet.nl/~vcdcompetition/Results.html

http://www.puremotion.com/mpegxs/qualitycompared/dvd.htm

http://www.emediapro.net/EM1999/ozer9.html

http://www.cyberport.com/~tangent/video/reviews/

http://www.darvision.com/product/dvmpeg_compare.html

http://members.fortunecity.fr/gber2/rh_mpgcomp.htm

http://www.datacompression.info/CommercialProgs.shtml

http://www.sph.umich.edu/~aspinall/offsite/graphics.html

-- Burton --