An explanation of TMPG's CQ please.

[markrb]

I have searched high and low for this answer, but I just can't seem to find it. I thought for sure it would have been easy to find, but it just doesn't seem to be. It doesn't help that the search here requires at least 3 letters.


What is CQ in TMPG and what makes it better then CBR? It seems it's what people like to use, but I can't find out why. In my case it relates to mpeg-1 as I am only using TMPG for VCD's and not SVCD's.

Any help or links would be appreciated.

Thanks,
Mark

[nyplayer]

Try this link
http://216.239.33.100/search?q=cache:w-jMuvNuVz4C:tangentsoft.net/video/mpeg/enc-modes.html

[markrb]

Thanks that link does help, but it still leaves me wondering is CBR better then CQ encoding if space is of no issue? It seems to me that CBR should be at least as good, if not better, since CQ is still a form of VBR. If space is not an issue should CBR be used instead of CQ to assure you of the best possible quality?

I am going to re-encode the video I just did using CQ with CBR and see what I think, but I would like to hear more from the people that use TMPG more often then I do.

Thanks,
Mark

[Boulder]

Based on my own experiences, CBR looks better. I've encoded some XVCDs with both CBR and CQ and I like CBR better. The difference is not big but it's there anyway. I'm sure the opinions will differ since people have different TV sets etc.

[Abond]

Well, it is a bit dificult to compare, as CBR quality depend solely of bitrate - higher=better. CQ is VBR and has also level of quality. IMO if you put the level of quality of CQ so, that the average bitrate match the bitrate for CBR, then CQ is better.
Greetings.

[hitch]

vcdhelp.com has a lot of threads about this. Try this

http://www.vcdhelp.com/forum/viewtopic.php?t=114562&highlight=2pass

[dvd2svcd]

Don't take whatever banjazzer is saying at face value, I've had a long and exhausting discussion with him about TMPGenc CQ and CCE's multipass and he is extremely biased upon CQ, but in the end he admitted that he actually never had tried CCE but still insisted that CQ was better, how's that for argumentation.

[hitch]

I get a kick out of reading banjazzer's posts, seems one sided most time but he's earnest. I was just giving markrb a clue where to find more posts on the subject of CQ.

CBR vs CQ is a strange comparisson to me. With a high enough bitrate, CBR will always be bette than CQ by definition since CQ will adjust bitrate as necessary. Am I wrong on this? :confused:

[dvd2svcd]

My first thought was "your totally right" but then I thought do we really know what TMPGEnc does "behind the scenes" when encoding in CQ? Maybe there's a soften filter or sharpen filter or deblocking filter or something that is used in TMPGEnc's CQ mode which "pleases the eye" more than the raw CBR encoding, do we really know?

[hitch]

You're right, the real test is what the user thinks is good.

BTW , have you asked the author for TMPGEnc this question? :D

I would think though that in most practical applications the sweet spot is the trade off between file size (or even how automated the process is) vs perceived quality. TMPGEnc CQ may give a smaller file size than CBR for a perfectly pleasing quality, but I think requires more trial and error to get the perfect setting. Thus it's a usefull option if you have the time and can't afford CCE ... and most can't.

The difficult thing for me personally in the past with CQ was to find the right setting since for each movie the sweet spot is slightly different. I'm a sort of an imperfect perfectionist, which is to say I like it perfect but I'm practical when it comes to time and cost. Thsi is why DVD2SVCD is great for me, but I still wind up spending addiitonal time converting DVD menus...:)

I have noticed that for old B&W films 2 pass VBR does a very nice job for relatively low bitrate (1800 avg) on most old movies since there aren't a lot of action and high motion scenes. Also it's a more automated process than searching for the right CQ setting. I would imagine CCE would be even nicer.

[Abond]

I would think though that in most practical applications the sweet spot is the trade off between file size (or even how automated the process is) vs perceived quality. TMPGEnc CQ may give a smaller file size than CBR for a perfectly pleasing quality, but I think requires more trial and error to get the perfect setting. Thus it's a usefull option if you have the time and can't afford CCE ... and most can't.
Well, exactly this is why the CQ is there;) . Hmm, imagine that we are going to the upper level of bitrate, let say 9000 kbps (I don't know of course if this is the level, only for example) and make CBR. To achieve the same quality with CQ one should put the max bitrate to 9000 and the quality level to max (100, I think this is %) and the result will be... CBR with 9000kbps:D .
IMO CQ as function is the same as 1 pass VBR in CCE. Hmm, I don't know if it is posible to be used for roba method with TMPGEnc...
Greetings.