Everyone told me (not here) to use CCE and that "MainConcept is horrible in comparison". I used CCE with the proper settings, and I can't for the life of me find a difference between the two. What should I be looking for? Is everyone else just incredibly anal?
I mean, how can one 8MBPS mpeg2 vary so much from another?
I don't have the answer for you, but I am extremely pleased with the results I've been getting from Main Concept. When I compare my DVDs to the master footage, the difference is negligible in my eyes.
I suspect the differences -- if there are any -- might be more apparent at lower encoding rates. Comparisons on this are very difficult because you need to set both encoders to exactly the same parameters, and many of these are difficult or impossible to set identically on both encoders because they don't provide the same controls, or they label them differently (like motion search parameters).
I have tried to do comparisons, but have never felt that I was able to get an apples to apples comparison, and finally gave up trying. I decided that the professional products (which includes both the encoders you are considering) are probably pretty similar, and that workflow considerations are more important. Having the encoder built into the editing program is a huge workflow advantage, so I am sticking with Vegas/Mainconcept.
One other thought: DCT compression (used in JPEG and MPEG) has been around for a LONG time, and the industry is WAY down the learning curve. Trying to find differences in such a mature technology is like trying to find performance differences between AMD and Intel processors. The differences definitely exist, but do they matter for 99% of the situations you encounter?
Not sure about this, but I think the proof of the MPEG encoder is in its low bitrate performance, rather than its high bitrate performance. For me, I always use high bitrate to master DVDs, and have no complaints with MainConcept. But I'm also not one of the professionals that populate this forum... we have a group of people with widely different levels of talent, pickiness, and BUDGET here. It's very good to have such a varied group of people to learn from, but you must always consider the source to determine whether a particular comment applies to you, and sometimes you just have to take what you've learned, try it yourself, and make your own decision.
I note that if mastering in DVD Architect, the manual says you MUST use MainConcept. Why is this? Also if one uses MainConcept and checks the "include audio stream" box in the audio tab, will this give the same result as encodeing the audio seperately?
As has been said WHAT's being encoded will have more effect than the encoder. Keep in the back of your mind "how many pixel change between this frame and the next" when shooting and editing and you'll make the job the encoder has to do easier and hence get better results.
For example a dissolve between two clips both with full frame motion means every pixel is changing between every frame during the dissolve. The encoder has to find a way to represent that numerically within the bit budget you've given it to work within. 2 pass encoding gives it a better chance but it's still a big ask. But if you'd shot both of those with a slow shutter or added motion blur less pixels may be changing so you can find lower quality images actually encode better. Noisy blacks are a real problem, I;ve encoded a shot of a country road at night with one street light and the encoder just couldn't cope. I ended up with noise that froze for every GOP, looked really nasty. As the shot was 100% static, motion blur averaged out the noise and then the encoder gave a pristine result.
Encoding is a mix of art and science, good commercial DVDs have large sums spent on them getting it right, sure they have high end kit but they also put a lot of effort into what goes into that to start with.
Don't know about you all, but I do like Main Concept and yes, I do use a lot TMPGENC Plus (which I do like better than CCE). You can render your timeline with TMPGENC through the PluginPack Framesaver. You can learn more about TMPGENC at www.videohelp.com ;)
Also if one uses MainConcept and checks the "include audio stream" box in the audio tab, will this give the same result as encoding the audio seperately?
No. If you encode separately to AC-3, you are compressing from the original audio. If you check the "include audio" box and let DVDA use that audio, it will be compressed in the MPEG2 file, decompressed, re-compressed back to AC-3. As you can see, there will be a lot of compressing and decompressing happening which is NOT good for your audio.
Of course, you can also "include audio" but render the AC-3 separately anyway. You'll have to tell DVDA to use the separate AC-3 file. This could be useful if you were going to use the MPEG2 file somewhere else where audio WAS needed.
I did an indepth test of CCE, Procoder, Mainconcept MPEG ecoder, TMPGenc and Sorrenson Squeeze a month ago.
I'm a cameraman first and also direct/produce, but my eyes are pretty good at judging overall picture quality and I've been burning DVDs since buying my Pioneer A03 a few years ago (I now own a Teac DVD burner and an A07 as well).
Here's my test:
I shot pristine DVCAM footage with my Sony DX-D30 CCD Dockable Betacam SP Camcorder using a DV back. The shots consisted off:
- Macbeth chip chart
- color bar
- well-lit talking heads (tripod)
- fast motion of dancers (handheld)
- a interior/few exterior shots (handheld)
I edited with Vegas and included a few basic transitions (fades/dissolves) and some fancy wipes. I exported an uncompressed 5-minute video clip and imported it into all the encoders and created MPEG2 video at various bitrates using CBR and VBR. I authored with DVD Workshiop, and carefully labelled each sequence.
After burning the DVD disc I played it back on a Panasonic DVD player and viewed it on my Toshiba 43" projection TV, my dad's high def Samsung DLP TV, my Sony broadcast monitor and a GAO 27" color TV.
Here's what I found:
Canopus Procoder, looked the best. It just looked a touch smoother than anyone elses. At low bitrates, using "mastering quality" it looked really, really good and produced the best MPEG2 video of the group. Be warned though, my test version 1.5 was really, really slow at mastering quality. GRADE - EXCELLENT
CCE was slightly under Procoder in overall picture quality, but quite good. I actually also created some MPEG2 with tweaks from guides I found online, but could never get it to match Procoder's quality (for VBR encoding, used VBR multipass, 3 passes) GRADE - VERY GOOD
Mainconcepts, the last of my big three, produced picture quality inferior to Procoder and CCE but it was a speed demon --- man it's fast. For fast jobs, I use it all the time. GRADE - GOOD
TMPGenc was close to the quality of Mainconcepts, not as good as CCE or Procoder and it was really slow as well, especially if you applied filters to it. There's a lot of nifty things you can do with TMPGenc and it seems to produce really nice SVCD and VCD video. GRADE - FAIR to GOOD.
Squeeze, forget it. Awesome Quicktime movies and believe it or not, Windows Media movies. Sucked for MPEG 2, but MPEG1 wasn't bad. GRADE - POOR to FAIR
At high bitrates, there was very little difference between Procoder, CCE and to a certain extent Mainconcepts, at average and low bitrates (4,000-5,000 kbps) the difference is clearly visible.
I don't have time to discuss what I was looking at, but I paid close attention to breakup in fast action scenes and for artifacting when transitions occured as well as overall picture smoothness (low bitrates or bad encoding, picture just looks grainier).
I also found that a good projection TVs and high def sets are great monitors, you see flaws in the picture that you don't catch on a 19" color set.
These are just my findings, using my own eyes and picture quality preferences, others might have varying results. Hope this helps, sorry this post got so big.
I did the same sort of tests as mbeli and found the same results.
Procoder is the only one that I could not tell the difference between the orginal DV and the DVD. However I do use Mainconcept when speed is required and not high quality.
squeeze and main concept are one and the same for mpeg2 .. squeeze uses the mainconcept encoder , same version more or less as vegas ..
CCE is better at progressive sources and takes a bit of knowledge for best results - but can be totally stunning ..
procoder is better than cce on interlaced sources .. version 2 is better than version 1.5 (and faster) ..... v2 and v1.5 are like completely different encoders in many respects ...
Take care with comparisons - what tests have shown is that the MC-Encoder delivers a much better quality in the version integrated in Vegas, compared to the Version that was integrated in Adobe Premiere 6.5 for example. Seems to be that the hand-over of the frames is superior, compared with other applications.
I have seen that in the past (before most people had DVD-burnder), when I generated SVCDs - for Premiere 6.5 you can foreget that, with Vegas it was fine.
The Vegas-4 encoder was compareable with the CCE, in the quality delivered. I have not tested the new 2-pass version in Vegas-5 so extensively, but I see no reason why it should have become worse.
So, beside speed there is no reason why you should use the CCE really. I agree that Procoder is a little bit better in picuture quality, but if you prepare DVDs the differenc in quality can be ignored today, for a typical video at high data rates (one hour or so).
However, was this version 2.0? And, what settings do you recommend? I am using mastering quality with 2 pass VBR at 7000KBPS average/7000KBPS max and 0KBPS minimum. I'd do a little higher than 7000 but I am using PCM audio and don't want to go over the DVD spec for bitrate.
there should really be about a 1000 split between max and average ..
you should really use ac3 for audio as that would give you a lot extra for video .. but if you dont , a setting of 7800 max , 6900 avg. would be good in PC2 .. min should be 0 , except in some types of video with dark or very bright scenes with some highly detailed portions .. then a setting of 1500 min is good .. you can use 1500 all the time really and check both dvd boxs ..
for cgi use 9 bit and high detail/fast motion/camera work use 10 bit .. cartoon even can use 8 bit for better quality ..
I do have a question, however, about what settings you used for each encoder. I did a lot of testing on one single encoder, the standalone Mainconcept encoder, and reported my results here:
What I found was that at low bitrate settings (where the differences between encoders really start to show), I was able to get a LOT more quality by changing certain settings. However, these changes also slowed down the encoding process. My conclusion at the time was that the awesome speed of the standalone Mainconcept encoder may largely be due to having the default settings for some of the search estimation parameters optimized for speed rather than quality.
I don't have some of the codecs that you used for your test, but I suspect that the quality and speed differences might be far less (and perhaps almost non-existent) if you were able to set all the parameters for each encoder to roughly the same setting. The problem, of course, is that each encoder gives you access to different sets of controls, and those controls are often not labeled the same. Also, the controls sometimes don't affect the same thing in the same way. This makes these comparisons very difficult to do.
My main point is this: The Mainconcept codec, if tweaked, might be able to provide similar quality to the Canopus, and would probably then be equally as slow.
I use standalone MainConcept and tweak it myself. Excellent 2-pass quality.
I also bought and tried Procoder Express, I think it's called, and did not really notice any better quality. It is very tweakable but I found the tweaks complicated. The interface is very different. Procoder is great when you need to control filesize and get max quality for minumum filesize. It seems to be able to encode just about every format. I need to get a manual and learn to use it right, though. For now, I turn to MainConcept for most all MPEG-2 encodes. I like it, I trust it.
Yes! DVDA will accept encoded files from procoder. I created a m2v file at a constant bitrate of 9.2Mbps with filters applied and was VERY surprised that DVDA accepted it without recompression. I have to agree - Procoder is far better than any encoder I used so far.