MPEG2 Encoder test

[Edmund Blackadder]

P.S.: Just to stay on topic, ProCoder 1.5 is still the best DV to MPEG2 encoder.

That's debatable. I think you're both full of yourselves. :P

Yes we are full of ourselves and proud of it! Though Fred, I think, has an edge over me. I gotta work on my fullness to catch him, as he's got a head so big he cannot walk through any doorway anymore . Perhaps there is a chance he can still walk through The Black Gate of Mordor when it's wide open .

Anyway, back to the test. For anybody to change my mind about ProCoder's 1.5 superiority over any other common cunsumer encoder in DV->MPEG2 area, I need to see some acual MPEG2 footage samples, not just JPEG stills, encoded with ProCoder 1.5 (as I said before - version 2.0 is a big step back quality-wise) and the competing encoder. I would love to find an even better encoder for DV source, but haven't found one yet (even PC 1.5 is not perfect). Besides, DV PAL to MPEG2 NTSC conversion feature along is priceless, especially when it's executed as superbly as ProCoder does it.

[State Of Mind]

Edmund Blackadder, I think you need to lay off the insults...it's only showing how immature you can be.
And what you said defies the "superiority" of ProCoder. According to you, it goes up half a version up the road in the wrong direction in regards to quality, well, that's no superior program at all. Also, you say your mind could be changed providing you had MPEG-2 samples to judge. Well, who knows, it is quite likely that other programs can do a better job than others. Each person's source video is different so its user-dependant. What suits your style may not suit anothers. *Sigh* Such a big bubble you have I think it would crash your whole system to have it popped.
Anyway, cheers.
PS: Oh, and you're quite hipocritical stating that others are full of themselves when clearly, you are way over the edge with the whole "Self-Righteous" act happening there. You're only means of justifying yourself is by stating that others are worse than you are. Such a pity. Wrong forum for psychological analysis, though. Sorry to mods but I had to make a point. Back to the testing, if there is yet to be anything proved by any further testing...

[Edmund Blackadder]

Edmund Blackadder, I think you need to lay off the insults...it's only showing how immature you can be.

Listen State Of Mind. I wasn't the one who started the pointless insults in this thread. Everything was nice and dandy in a good spirit of discussion until Fred T came here and started spilling bullshit all over what I had to say with absolutely no backing to his words. So if someone is immature in here is Fred T for hijacking a peaceful thread and turning it into a big argument. And you, State Of Mind, obviously are too serious, as you don't even get innocent jokes. You need to lighten up yourself. Half of the regular forum members do more screwing around here in On Topic threads than I could ever want to do, and they easily get away with it. For some reason it's always people like me (I guess I don't have enough post count) who get criticized for doing a similar thing.

And what you said defies the "superiority" of ProCoder. According to you, it goes up half a version up the road in the wrong direction in regards to quality, well, that's no superior program at all. Also, you say your mind could be changed providing you had MPEG-2 samples to judge. Well, who knows, it is quite likely that other programs can do a better job than others. Each person's source video is different so its user-dependant. What suits your style may not suit anothers. *Sigh* Such a big bubble you have I think it would crash your system to have it popped.
Anyway, cheers.

Listen, video is just another business for me (besides being a professional musician). And because of that I need the best possible software and hardware I can afford. ProCoder 1.5 is one of those programs that are unsurpassed for interlaced DV footage. Trust me, I captured both PAL and NTSC, out of about a dosen different DV cameras (mostly 3CCD), also though Canopus ADVC-100 and tested the conversion with numerous encoders. I guess what? ProCoder 1.5 is the only encoder I'm satistied with, be that viewed on CRT or LCD computer monitor, CRT or plasma TV or various other display devices. No other encoder was good enough for me to make a product with it and charge people money. My production values are high and I don't want to compromise those values by using an inferior MPEG2 encoder. And if you still think I'm full of myself, simply PM me your PO Box number and I'll gladly send you some of my original DVD's to prove my point. My motion menus that I design mosty in After Effects are completely original and second to none (and that is only because I'm always very critical to my work in preogress). I always give my clients more than what they pay for and I rely on ProCoder 1.5 to make the final encode because it's the best encoder around. I've seen a lot of work of my competitors and while sometimes I do see a really good production - it's the MPEG2 encoding quality that falls flat on its face.

Oh yes, regarding a downfall of PC 2.0's encoding quality. Like you've never in your life heard of any companies that made outstanding products and then started to slide down in quality. Gimme a break!

And the last. I joined this forum to give people some good advice from the things that I've learned and also would like to learn even more from other people's experiences. There are some people here like BJ_M that are very knowledgeable and so valuable to this furum, that we ought to pay them salary . But there's also people like Fred T and now you, State Of Mind, that just come here and either start stupid arguments or give worthless lessons. I don't need neither one of those. I'm here to deal wih video. So please, if you don't have anything valuable to say, please go to Off Topic forum which will suit you well.

[vhelp]

First ...

That sample that BJ_M provided us (thanks BJ) was not a good sample to
demo. It was riddled with macroblocks, and I just gave up on it. In
the end, I found my own simple method worked just as well as his did
..and then,

Second ..

Oh yes, regarding a downfall of PC 2.0's encoding quality. Like you've never in your life heard of any companies that made outstanding products and then started to slide down in quality. Gimme a break!

so true. So true. I mean.., look it TMPGenc. It went through so many
version (when it was 12) and was good; bad; buggy; bad; good; fair; good;
excellent, etc etc. And,

Third ...

You need to UPLoad some sample MPEGs for us to here !!
I've tride PC v1.5 to no avail. And, I'm no dummy. I very good 's
and good judgements and tweaking abilitys in the things of settings
with this encoder. And, in the end, I was not satisfied. I used both
DV and regular sources for this Encoder. In the end, I went running
back to TMPG.
.
But, the one thing that I did not like about PC, was the in the background
areas, I could see things sort of move or shift, and I didn't like that
one bit. I cant' quite put my finger on it, but its there. Anyways.

Moving on..

Actually, I think that if *HE* (you Edmund) wants something comparitable,
then *HE* needs to provide the sources for us to DEMO him

Why ??

Because it seems that *HE* is doing something wrong, or is not satisfied
with he own work, and have settled (comprimised) w/ his methods.. which
is ok.

But if he is going to judge any one of us ( be it *any* ) then he needs
to provide the source of *US* to demo, based on our process (vs. his)

@ Edmund

You need to provide us the sample source to demo you
Until then..

Cheers all
-vhelp 2985

[BJ_M]

i noticed later that it did not convert well to DV -- in the first post i mentioned i did not encode the samples from DV either ...

[vhelp]

oh.. then I'm a pretty stupid one here
cause I re and re read your posts several times. hmm..

Thanks for the clarification BJ_M
-vhelp

[BJ_M]

oh.. then I'm a pretty stupid one here
cause I re and re read your posts several times. hmm..

Thanks for the clarification BJ_M
-vhelp

im the stupid one -- it was from DV , I forgot .. The source was film to D-Beta to DV .. but the frame used was fine ....

BUT

for the purposes of that comparision it was fine also ..

[vhelp]

Ok. We're both

@ BJ_M ...

Hay, have you decided on a new test clip to try ??
.
I saw the one you did w/ your HDCAM (the man on the boat) that you
converted to divX for us to D/L

(
.. in another thread over here
.. * VideoHelp.com Video Benchmarking
)

I like either all true progressive frames, or Telecined frames
(3:2 pattern) to work with I mean, test with. I like the 3:2 pattern
scenario becuase I know how to handle these with reverse reproductions
without doubts

-vhelp 2988

[BJ_M]

yes - i was just going to do one ... should it be HD or NTSC ?

[vhelp]

I like NTSC (assuming 720 x 480) but I don't think that it matter how high
the resolution is (long as you bring it down to 720 x 480 for our "current
trend of the day" standards, which is DVD) Oh, but this time, please make
sure there are no macro blocks in them (unless you can't
help it)

Yeah, whenever you get the chance,
-vhelp

[vhelp]

Yea, I was playing aroudn w/ your other file to encode, and it for
some very strange reason, TMPGenc doesn't like it, and BOMBs out
really bad w/ a zillion error dialog pop-up ads (haha, just kidding
about the ads) but i'm only encoding it all in less than a minute,
if that matters any, on that thread. Anyways.
.
..point is, maybe that divX clip might not have ben a good encode,
unless that one was a bad make or something when you first made it.
.
I don't mind D/L'ing a bigger file over my dialup. a 20mb or is ok.
Its not like I'm going to be D/L'ing a bunch of these every day you
know.., and a little excitement is something I look forward
to, even in this D/L hehe.

Is it possible for a Telecine (29.970 fps) clip from your HD cam (if you
cam does this) ??
.
I'd love to see something like this from a cam device.

-vhelp 2990

[Edmund Blackadder]

But if he is going to judge any one of us ( be it *any* ) then he needs
to provide the source of *US* to demo, based on our process (vs. his)

@ Edmund

You need to provide us the sample source to demo you
Until then..

Cheers all
-vhelp 2985

Vhelp, I have no problem with that. However, first of all where do I upload those large files? I don't have my own server for that. Does Videohelp offer some sort of a temporary video testing storage space for current threads?

Second, the toughest test for an MPEG2 encoder is the outside shots, such as busy streets, in the daylight, with lots of movement, detail and some panning. All of my outside shots that I have on DV are in PAL format, which as I mentioned earlier makes even higher demands on MPEG2 encoder. So the best way to view those would be on a multi-system monitor. Converting them to NTSC will make testing much more difficult. But if any of you have PAL viewing capability I can make a RARed VIDEO_TS with different encoder test samples of the same footage subtitled to show which encoder it was encoded with. Or I can give you a DV file snippet for you to test. Sony DCR-TRV900E PAL camera with the street footage is the worst nightmare for MPEG2 encoders, and not because of the bad image quality, but on the contrary, because it produces incredibly detailed image, fully utilizing PAL resolution and therefore giving headaches to even DV codecs on certain spots!

On the other hand, my usual video productions that are in the theatre/music performing area are not as a good stress test for MPEG2 encoders as the street scenes. Besides, I cannot post those samples to a general public without first asking permission of actors/musicians, which I'm not about to do just for this test (though I can still send those DVD's out for anybody who wishes to see them). However, I'm always satisfied with ProCoder 1.5 for the performing arts stuff. It's the outside that sometimes doesn't quite cut it with PC 1.5, mostly because of the increased mosquito noise. And the other encoders I've tried look absolutely awful on those spots.

As for the moving backgrounds in ProCoder, yes I've noticed that too, but it only happens in VBR mode with lower average bitrates and fairly noisy footage. That happens because of the ProCoder's internal filtering that improves the overal perceivable quality by slightly softening the unimportant parts of the image and giving them very few encoding bits, while still retaining the sharpness of the good stuff. I think this is why it looks so much better than other encoders, by doing that kind of a compromise. It seems that mostly because of this technique PC 1.5 got less mosquito noise than Cinema Craft and less blockiness than TMPGEnc. And another thing I never liked about TMPGEnc is that its MPEG2 output was somehow never of the same brightness/contrast as the original DV file. Perhaps it got fixed in the newer version. I know you can tweak those values, but why didn't/doesn't it output normal values to begin with.

Anyway, please let me know what you wanna do. I'm all open for some experiments and eager to prove my point. .

P.S.: Of course in a year or so we'll probably be discussing such things as "What's the best encoder for 1080i HDV to Windows Media 12 for Blu-Ray burning" . And nowadays ProCoder 2.0 has at least one good use: I can take 1080i HDV captured M2T files and easily convert them to either standard MPEG2 for DVD, or better yet to Huffyuv and then to MPEG2 in ProCoder 1.5 for even better image quality. Though of course there are other ways to resize HDV down to SD, such as AfterEffects. Man, there are just too many possibilities.

[vhelp]

@ Edmund

Let me ask you an HONEST question ...

regarding your DV footaging.. do you use a tripod in your shots, or
are you counting upon the "fluid" experties of your arms motions
while guiding the CAM ??

I don't see how traffic would be an issue w/ respect to quality.
What I *DO* believe is that your cam MUST be able to shoot full frame
( 720 x 480 ) mode, Progressive. Like Cinema cameras do.., in order
to obtain that pro look. That is, non macro block appearance.
.
Course, the encoder helps

But, if your source has brushes, bushes, stone, grass (not the kind
that is smooked today) and a few other odds, these will be difficult
to handle in *some* situations.
.
Yes, panning is most difficult. But, if you pan in a fluid motion
with a tripod, there should be NO macro blocks in *that* scene as well !!

Posting a sample clip is not hard to do. But I don't think the vcdhelp
can help since they are strapped from resources. You have
to find a provider. I found one that offers an e-mail account and
20mb space. (I told them I only want to post web pages and that's all)
That's when they told me they could set me up w/ an email account and
20mb space, nothing more. From there, it was music to my ears. The
price is simiple. $5 bucks a month for 20mb space. I never look at
my e-mail from that provider. I instructed them to delete (if it
should ever fill up or whatever) and they said that'w what they would
do. So, if you or anyone have been sending me e-mail.. they are shredded
before I can even think about it, hehe.. Anyway.
.
I use an FTP app to upload my files to my account. That's about it.
.
Then, I post an update to my Sample Clips pages (here) ...

* VHELP's Samples..

..for everyone to D/L. I've ben doing this since 2002 I think. And it's
ben a fun experience.

-vhelp

[Edmund Blackadder]

@ Edmund

Let me ask you an HONEST question ...

regarding your DV footaging.. do you use a tripod in your shots, or
are you counting upon the "fluid" experties of your arms motions
while guiding the CAM ??

Vhelp, I have outside footages with both Bogen Manfrotto tripod and handheld with optical stabilization. That doesn't make much difference for the inferior encoders. As long as there's at least one of these elements: a lot of detail, pavement, movement/panning, sunny day - the lesser encoders give me a bunch of artifacts. Not to say that PC 1.5 doesn't give me some artifacts either, but a lot less than anything else I've tried.

Cloudy days always come out better on video and compress better too. And of course I always shoot with manual exposure, because automatic setting overexposes most of the time. So I'm always trying to make my shots as good as the particular camera can get.

What I *DO* believe is that your cam MUST be able to shoot full frame
( 720 x 480 ) mode, Progressive. Like Cinema cameras do.., in order
to obtain that pro look. That is, non macro block appearance.

Well, actually I'm not discussing here the advantages/disadvantages of progressive frames versus interlaced fields. My ProCoder 1.5 praises in this thread are strictly tied to interlaced DV footage. For example I actually quite like Cinema Craft SP 2.50 encoding results from progressive uncompressed (or Huffyuv) frames of the menu motion graphics that I do for DVD's. It actualy samples red color into MPEG2 better than ProCoder (CCE has less stairstepping jaggies on RGB 255 0 0 type bright red when viewed through DVD component output or even a computer monitor). However, put Cinema Craft to work on DV and it's simply unsatisfactory, no matter how high bitrate you put up or what kind of filtering you use - it just doesn't look good.

However, I still prefer interlaced look for video footage - it's more true to life. 24p, 25p and 30p I think are better only for fantasy world - in other words motion pictures. After all that's what 24p gives you - missing information so your brain can fill in the rest, creating a unique otherworldly feeling that's hard to describe (besides being simply jerkier than 50i or 60i). It's just a stylistic thing that originally was just a technical limitation of the possible projected film speed, I guess. Unfortunately it doesn't work very well for most other purposes, such as daily videography shot with CCD devices.

And sometimes things just don't make much sense. Man, my almost 10 year old DCR-VX1000E 3CCD PAL camera does not have the sharpest resolution, in fact it's far from it. It doesn't even have the best color rendition. But somehow when you see its footage on a high quality CRT TV, it's so true to life that it feels it will almost pop out off the screen into your livingroom. I've recently tested PD-150 and while it's got a higher resolution and brighter colors, the footage somehow looks flat. So, my point is, it's possible to achieve a really lifelike and almost 3-dimentional look with a quality video equipment, while the resolution and progressive capabilities aren't always the most important things for breathtaking video. After all, the impressionist paintings don't have that much detail at all, but often are more successful in giving you a great impression of a subject than some of the very detailed paintings of similar matter. It all varies . Some things defy logic and simply give you better perceivable results despite their inferior specifications. ProCoder 1.5 seems to be one of those things (along with DCR-VX1000E) .

My honest opinion is that progressive shooting should be left for 16/35/70mm film, as nothing, not even HD is good enough to replace the unique qualities of film. Every existing format has something unique about it and should be loved for that (or hated ). I'm really not looking forward to digital projectors replacing film in the movie theatres. Also I never understand those people that deinterlace their videos, put grain in them, or even shoot in progressive if it's available in their videocameras - it's not going to look like film simply because it's not a film - as simple as that. You will always see the horizontal (and in many cases vertical) lines, different depth of field and different light sensitivity with CCD-based images from what would be on film. I've never seen a progressive footage that was shot on video that would fool me into thinking that it was shot on film. All it ever does is just annoying me with that fake film-like look. Interlaced video has a lot of good benefits (smooth motion is one of them), so why not use it to the full?!

Posting a sample clip is not hard to do. But I don't think the vcdhelp
can help since they are strapped from resources. You have
to find a provider. I found one that offers an e-mail account and
20mb space.

I do have my own website on Yahoo with 15 MB storage (but most of it is used up) as well as Yahoo Briefcase with I think 30MB of space, but I would need to give an access permission to anybody with a Yahoo ID who wishes to download stuff from there. When I prepare the clips I could do that to you Vhelp, if you you wish (but you do need a Yahoo ID). Do you want me to post DV clips or would you prefer VIDEO_TS with subtitles of different encoder results? Let me know.

[vhelp]

FWIW.. I found the sample clip that BJ_M posted (man on a boat)
clip pretty good for an HD cam device.

.. The divX that was used in portraying it I didn't like though (neither
.. did my system - keeps crashing on me, and I loose stuff to coruption)

But I liked the quality (so far) and I believe it was Progressive.
I realize one can't replicate Cinema. But, we can strive for it w/ our
imagination, can't we ? anyways. I'm not much for full screen videos
that come from cams, unless you convert it to widescreen by cropping.
And I realize that means possibly loosing some real estate (pixels) but
as long as you don't go dumping your processes with everybody, then
there is no harm in the illusion of one's idea(s) right.. Anways.
.
There are some that don't believe one should crop a 4:3 video, but that
would have to depend on your direction (or, directors version) and what
you are visualizing, and how well you prepare it for your audiances to
believe.
.
When I shoot footage w/ my TRV22 ( a toy in comparison ) in 4:3
mode, I crop top/bottom to illustrate a widescreen illussion. Plus, it
gives my footage an edge in quality, and also help give the viewer a
sense of Cinema or something. Anyways.

I don't want to go off track of this topic here. And I respect your
choice on Procoder 1.5 as your encoder
.
Actually, I hadn't completed my Interlaced tests w/ PC 1.5 just yet.
It's ben on hold in lue of other things going on. (You know how it goes)
.
Still, I'd like to try another hand on a new source clip to test my
prefered encoder (TMPG) with, cause I like the excercise. Or, even a
good SuperBIT disk would do fine. My Fifth Element SBIT disk is one I
use as a gauge because there are no macro blocks to worry about. It's
higher bitrate, and better encoded (IMO) ..Anyways.

If BJ_M ever gets around to an new one, that'll be great. In the mean time,
I'll keep poking around.

-vhelp 3001

[Edmund Blackadder]

But, we can strive for it w/ our
imagination, can't we?

Surely we can. But it's just as you said - it cannot replace cinema. It will still look fake, no matter how close it comes to the desired film look.

I'm not much for full screen videos
that come from cams, unless you convert it to widescreen by cropping.
And I realize that means possibly loosing some real estate (pixels) but
as long as you don't go dumping your processes with everybody, then
there is no harm in the illusion of one's idea(s) right.. Anways.
.
There are some that don't believe one should crop a 4:3 video, but that
would have to depend on your direction (or, directors version) and what
you are visualizing, and how well you prepare it for your audiances to
believe.

There's nothing wrong with cropping 4:3 footage for 16:9 presentation. Unfortunately it does cut a lot of resolution. On 4:3 TV's the letterboxing will actually look better than anamorphic, because there's no need to scale down the anamorphic footage inside DVD player (or any other playback device) to display in letterbox, introducing artifacts. However on widescreen TV it's the other way around and simple cropping reduces the quality dramatically. So when you need to fill 16:9 screen with the letterbox footage you need to zoom in and then see all the unpleasantries of zooming in.

When I shoot footage w/ my TRV22 ( a toy in comparison ) in 4:3
mode, I crop top/bottom to illustrate a widescreen illussion. Plus, it
gives my footage an edge in quality, and also help give the viewer a
sense of Cinema or something. Anyways.

Since I also like shooting in widescreen, but the resolution is important to me, I decided to go broke and bought a used OpTex anamorphic adapter. It produces outstanding widescreen images. But there's one drawback - you cannot do a full zoom, especially with wide open iris. That's why I cannot use it for my indoor video work that requires a lot of close ups of distant objects .

Still, I'd like to try another hand on a new source clip to test my
prefered encoder (TMPG) with, cause I like the excercise. Or, even a
good SuperBIT disk would do fine. My Fifth Element SBIT disk is one I
use as a gauge because there are no macro blocks to worry about. It's
higher bitrate, and better encoded (IMO) ..Anyways.

That's all progressive footage and therefore is much easier on encoders than interlaced. So, I'd recommend you trying on high quality interlaced footage to see the real encoder issues.

If BJ_M ever gets around to an new one, that'll be great. In the mean time,
I'll keep poking around.

You mean you don't trust me to get you some quality footage?
As I write this I'm encoding a 7 seconds sample with some slightly moving pretty greenery (close up shot of big leaves with some unopened buds in cloudy weather) that I shot using a tripod with DCR-TRV900E (PAL). From what I've seen so far rates like this (at a bitrate of VBR average 4500kbps/maximum 9300kbps):

Best: ProCoder 1.5
OK: Cinema Craft 2.50
Worst: TMPGEnc 2.511.51.160 (was that a bad version?)

I will make a VIDEO_TS out of it and I can still try to post it. Do you wish to see it, Vhelp? Though it's still best if you have a PAL capable TV.

P.S.: Just now I've extracted some stills from the results of those three encoders, first of the still greenery and then of the same stuff, but with a slight wind. I will post them in a short while...

P.P.S.: I chose VBR average 4500kbps/maximum 9300kbps for this test because at the highest CBR they all looked decent (ProCoder still having an edge), but with lower VBR the obvious artifacts come out like vampires at night in Transylvania -> at 24p of course, and only on film .

[burnman99]

Has anyone tried either the free AVIDemux or Quenc or even Nuenc or Freenc? It'd be interesting to see how they'd stack up against the pay versions. I know that bbmpeg & the old free tmpgenc are much slower than these so they should be discounted. Just a thought.

Rog

[Edmund Blackadder]

OK, here are the results of my interlaced DV to MPEG2 short test.

This was a 7 seconds 25fps 720x576 PAL DV clip shot awhile back using Sony DCR-TRV900E 3CCD camcorder. It was mounted on a high quality Bogen tripod. No deinterlacing was used, so you will see interlacing horizontal teeth on a dynamic shot - and you're supposed to see that. This test is for interlaced DV footage and how these 3 popular encoders cope with it.

The 00s 24f set of shots is when everything was relatively still.

The 03s 20f set is when the wind was slightly blowing the leaves. This is where you'll see the weaknesses of Cinema Craft and TMPGEnc.

The MPEG2 files look a little darker than DV probably due to whatever MPEG2 decoder Premiere 6.5 is using (it might be Ligos). But all of the MPEG2 clips were decoded using the same decoder and from the same timeline. You might want to pump up the brightness of your monitor slightly. I didn't want to alter the brightness/gamma of these files, so they would stay as close to the original as possible.

All of the MPEG2 clips were encoded with VBR average 4500kbps/maximum 9300kbps.

ProCoder 1.50:
2 pass VBR, Mastering Quality, DC Precision 10.

Cinema Craft SP 2.50:
3 pass VBR, Luminance 0 to 255, DC Precision Auto, Image Quality Priority 25, No Anti Noise Filter.

TMPGEnc 2.511:
2 pass VBR (VBR), DC Precision 10, Highest Quality (very slow), No filters, Quantize Matrix Soften Block Noise 35-35.

All of the JPEG's were made using Adobe ImageReady with Maximum Quality 80 setting.


00s 24f (quiet, no wind):













03s 20f (windy):















I think it's pretty obvious that ProCoder 1.5 won this test. Cinema Craft produced way too much noise, while TMPGEnc created blocky artifacts.

If any of you would like to conduct your own tests on this clip, I can chop it up, RAR it and send you an e-mail. PM me for that. The original DV clip is 27MB, but many free e-mail providers (such as Yahoo and Hotmail) offer 100MB to 250MB of free storage, so it will easily fit in your mailbox. You can even create an alternative/disposable/alias address nowadays if you for some reason think that Edmund Blackadder will spam your inbox .

So, until I see one of your results with this clip and proven wrong - I declare ProCoder 1.5 a clear winner for the interlaced DV footage, especially on lower bitrates. ProCoder 2.0 though, even while being inferior to 1.5, is a close second. I didn't feel the need to include PC 2.0 tests for that reason.

[State Of Mind]

I admit, Procoder does look better than TMPGEnc does. I wonder how much better it could look if the video was enhanced with AviSynth desired filters which I certainly intend to do with my DV footage.

[adam]

Edmund Blackadder alot of that noise you see in CCE will disappear if you raise your image quality priority. 25 is far too low for that bitrate. Try it at about 45.

I do prefer Procoder 1.5 for interlaced footage (generally, not just in your test), but it does produce a much softer image. This is fine for consumer level DV because you want to flatten out some of that noise, but for something like a DVD-Rip this is very bad, IMO.

[Edmund Blackadder]

Edmund Blackadder alot of that noise you see in CCE will disappear if you raise your image quality priority. 25 is far too low for that bitrate. Try it at about 45.

Adam, OK.

ADDED:

Cinema Craft SP 2.50:
3 pass VBR, Luminance 0 to 255, DC Precision Auto, Image Quality Priority 45, No Anti Noise Filter. Only the windy clip (03s 20f) was updated with this new CCE setting.

I don't think it changed much of anything. It still looks like Cinema Craft adds excessive amount of noise for no reason. Still no match to ProCoder's result. You may ask why I didn't use CCE's noise reduction? Because the higher NR value you put in CCE the jerkier the motion looks in in the MPEG2 result - almost like interlaced look and feel with a hint of progressive jerky movements.

I do prefer Procoder 1.5 for interlaced footage (generally, not just in your test), but it does produce a much softer image.

I don't disagree, but don't completely agree either. While ProCoder makes images appear softer, it still retains the original resolution on important parts of the image. The A/B comparisons between video camera and MPEG2 files played on TV show that ProCoder clips are the closest to the look and feel of the original DV tape than the other two encoders. And that's what matters the most, in my opinion.

This is fine for consumer level DV because you want to flatten out some of that noise, but for something like a DVD-Rip this is very bad, IMO.

Yes, for the progressive film based DVD-Rip that's true, but how about interlaced documentaries and other similar material? I think ProCoder will win there too.

[BJ_M]

The noise redcution and dithering in the newest version of CCE has changed quite a bit -- 2.50 is a pretty old version (but very very good quality)

[cosmin]

Thanks, Edmund Blackadder, for the tip about ProCoder 1.5 being better than 2.0. I find this info very useful.

There is a problem, though: ProCoder Express uses the ProCoder 2.0 engine. There is no "affordable" version of ProCoder 1.5, right? Or, is ProCoder 1.5 still being sold, at all?

[BJ_M]

not being sold - on ebay sometimes .. has a usb dongle

[monzie]

Can I just comment?

Are we mixing up sharpness and smoothness? Lets look at the veins on the underside of the leaves on the first set of pics.......which wins here? CCE in my book all the others SMOOTH out the detail....but as already noted some artifacts (it aint 'noise') are added with CCE around the buds and a slightly 'pixelated' (or mosquito net) effect.....this is where CCE's IQP comes into play..IE..what are you trying to achieve? Smoothness or Sharpness? Go left (less IQP) for sharper images or right (higher IQP) for smoother images. Why did you choose 0-255 in CCE? Plus you havent stated the other setting used in CCE (zigzag etc)

Also Procoder over saturates (higher contrast level) the colours giving a 'percieved' sharpness whilst smoothing the image....commonly known as CHEATING.

Tmpg gives good results yet also smooths out the picture.

But the end result is.......forget any still frame/field test but does it look good to your eyes on the TV ? My answer is they will ALL look good BUT none are 100% replications of the original.

[Edmund Blackadder]

Can I just comment?

Are we mixing up sharpness and smoothness? Lets look at the veins on the underside of the leaves on the first set of pics.......which wins here? CCE in my book all the others SMOOTH out the detail....but as already noted some artifacts (it aint 'noise') are added with CCE around the buds and a slightly 'pixelated' (or mosquito net) effect.....this is where CCE's IQP comes into play..IE..what are you trying to achieve? Smoothness or Sharpness? Go left (less IQP) for sharper images or right (higher IQP) for smoother images. Why did you choose 0-255 in CCE? Plus you havent stated the other setting used in CCE (zigzag etc)

Monzie, in the old days when I had no ProCoder, I played with every possible setting in CCE SP 2.50 and no matter what I did the DV-originated footage never looked good to me. Or at least not as good as the original DV tape.

Regarding IQP setting, I've tried everything starting with 0 and ending with 100. With values near 0 there was not much mosquito noise, but the backgrounds looked simply awful, especially on gradient type surfaces. But as you move the slider towards 100 the backgrounds start to look better, but the edges become plagued by the mosquito noise. No matter what I did with it I wasn't happy at all.

This leads me to the ProCoder 2.0 issues. If let's say, let's just assume for the heck of it, that ProCoder 1.5 had pre-programmed the equivalent of CCE's IQP 45 setting (near perfect combination between background and main subject priorities), then ProCoder 2.0 would be more like the equivalent of CCE's IQP 10. There's hardly any mosquito noise at all, but the backgrounds don't look as good as in ProCoder 1.5. I know it's a silly comparison, but I think that something similar happened when Canopus was designing PC 2.0.

Also Procoder over saturates (higher contrast level) the colours giving a 'percieved' sharpness whilst smoothing the image....commonly known as CHEATING.

But isn't any MPEG2 encoder's task is to cheat our eyes into seeing a nice picture, while chopping off about 3/4 of the original DV information (or even more so for higher quality digital formats)? They all must do quite a bit of cheating and they have to do it well. I haven't noticed virtually any change between the brightness/contrast/gamma/saturation of the DV tape played from the video camera and the ProCoder encoded DVD, played back on quality Pioneer and Toshiba DVD players. On the other hand TMPGEnc produces a darker picture and CCE is just too noisy. So, ProCoder tricks me really well because whatever cheats it does, its encodes are virtually indistinquishable from the original DV. This is the kind of encoder I've always been looking for and until ProCoder came out I thought I would never be able to achieve DV quality on DVD. Well, now I can, as anybody else with ProCoder.

Tmpg gives good results yet also smooths out the picture.

... and creates pretty nasty blocky artifacts.

But the end result is.......forget any still frame/field test but does it look good to your eyes on the TV ?

That's right. MPEG2 can be just like JPEG pictures that are compressed using a good algorhythm. They are highly compressed and very lossy, but can look like 1:1 copies with original BMP's (or whatever lossless format), because of using smart techniques to trick our eye, with some other math tricks on top of that. So you're right, still images don't do much justice to MPEG2 compression - that was my point from the very beginning. But I felt there was a need to have at least some kind of proof to my words without going out and getting a server to host those videoclips.

My answer is they will ALL look good BUT none are 100% replications of the original.

They are not 100% replications of the original, and unfortunately they don't ALL look good. So, praise to Canopus!

[troyvcd1]

I always use convolution3d on my videos. I use the lowest setting mostly for dv material. I wonder if this would eliminate the noise of the CCE encoded material. It seems to me that this is essentially what Procoder is doing here anyway is just eliminating noise by bluring some of the details.

[Edmund Blackadder]

Speaking of loss of detail. Many of you have said that ProCoder is removing too much detail and blurring everything. Well, I just noticed something. Take a really good look at this leaf in the 03s 20f windy shots:





Unless my eyes are going completely mad, but I think it's the ProCoder that preserved the most of detail on that leaf, comparing to the DV file! Please pay close attention to the leaf's veins (I don't know if that's what they're called). You'll see that CCE and TMPGEnc removed quite a bit of valuable veins information there. Oh man, that ProCoder just keeps piling up all kinds of trick-or-treat awards! Or more like trick-and-treat.

[BSR]

Could you also do the test with Procoder 2.0 since 1.5 isn't available anymore. Just to see the difference.

Yea 1.5 is great for anyone that already has it, but for anyone going to buy Procoder now, they are limited to 2.0, correct?

[Edmund Blackadder]

Could you also do the test with Procoder 2.0 since 1.5 isn't available anymore. Just to see the difference.

Yea 1.5 is great for anyone that already has it, but for anyone going to buy Procoder now, they are limited to 2.0, correct?

BSR, OK I'll add ProCoder 2.0 results later today. Cheers.

P.S to all: ProCoder 1.25 Demo is still available at the Canopus site. You need to register to get access to it - I believe you don't actually have to be a Canopus customer to be able to register, so there shouldn't be any problems. Here's the link:

http://www.canopus.com/cc/demos/demo_downloads.php

ProCoder 1.25 is pretty much the same in quality as 1.50, so that should give you a good idea of what you'll get if you first of all find it and then decide to purchase a used ProCoder 1.5. The only difference I've noticed is that ProCoder 1.25 behaves better on Windows98SE, but ProCoder 1.5 has a much better PAL<->NTSC conversion engine. Otherwise the MPEG2 quality is exactly the same, at least to what I was able to see. The 1.0 version though was not as good in the MPEG2 department. Also, the new ProCoder 2.0 doesn't work on Windows98SE, if that matters to anybody.

If you need to create a new account with Canopus, the "New users - Create a new account here!" section in the above link doesn't work. To register go here:

http://register.canopus.com/US/Support/Tech_support/techsupport_new_account.asp