Quote from: Garf on 2006-05-08 12:03:56

So please, just remember, 2 pass is for ABR.

Why don't you just disable 2pass in VBR Mode?

I'm wondering if we'll now get people who refuse to upgrade "because the new encoder doesn't support VBR 2pass".

Another graph.
This time I split each 2-pass encoding with YAMB (MP4Box) into shorter parts. The length of each part correspond to the 2-pass averaging period. I works pretty well (each small part in exactly cut at the right sample!).
There are:
- 77 segments of 2 seconds (input: -br 128000 -2pass -2passperiod 2000.mp4)
- 16 segments of 10 seconds (input: -br 128000 -2pass -2passperiod 10000.mp4)
- 6 segments of 30 seconds (input: -br 128000 -2pass -2passperiod 30000.mp4)

The last 10 sec & 30 sec segments are not present in the graph (they're too shorts because reference file is not a multiple of 10 and 30).





As you can see, the bitrate of each 2-pass segment is by far not constant.

-2 seconds: MIN=122 kbps  MAX=153 kbps  AVG=130kbps
-10 seconds: MIN=125 kbps  MAX=179 kbps  AVG=130kbps
-30 seconds: MIN=127 kbps  MAX=149 kbps  AVG=119/137kbps [complete serie: 127-129-133-149-149-27: the series is too short, hence the inaccurate averaging]

It also illustrate why short samples shouldn't be used for listening evaluation of "2-pass mode". If you encode each segment separately, all of them should end at 128 kbps and the plots would be perfectly linear; if you extract them from a long encoding, the bitrate would vary as shown in the plot's variations. Here: the variation goes from 122 to 179 kbps... no need to say that it should affects the output quality.

Thanks. I was already using wine for this, but wine is way too slow for executing things, at least on my Duron computer (which is the fastest that I have at my disposal).

Really good work. Conclusion is 2passperiod don't work correcty or perhaps that we don't understand correctly 2passperiod fonction ... ...

by Garf
I'm wondering if we'll now get people who refuse to upgrade "because the new encoder doesn't support VBR 2pass".

This is not a product for novices, this is a product for people that know exactly what they want to accomplish.  Of course, our quality settings are optimized to give best quality for a given bit-rate / quality levels,  but it is up to the user to decide what size he wants, and what additional methods he would wish to use in the process (e.g. 2-pass, optimizing for streaming, etc...)

... the removal of the 2-pass VBR option--[...] --would undermine the control that the expert user has over the encoding process, a process in which the user is expected to know full well what he's doing and what to expect of the encoder.

<hat and cape on>
I'm in a position where I can steer history so we do not end up downloading a garbled 192kbps SBR+PS L/R stereo tune from an artists website, and I will take my responsibility to stop it from happening, at least with Nero AAC 
</hat and cape off>

<hat and cape on>
I'm in a position where I can steer history so we do not end up downloading a garbled 192kbps SBR+PS L/R stereo tune from an artists website, and I will take my responsibility to stop it from happening, at least with Nero AAC smile.gif
</hat and cape off>

The Vorbis encoder has seperately documented, nearly hidden commandlines which all start out like --advanced-encode-option X...

Quote from: The Sheep of DEATH on 2006-05-08 18:52:00

The Vorbis encoder has seperately documented, nearly hidden commandlines which all start out like --advanced-encode-option X...

Boo for the fake link.

<hat and cape on>
I'm in a position where I can steer history so we do not end up downloading a garbled 192kbps SBR+PS L/R stereo tune from an artists website, and I will take my responsibility to stop it from happening, at least with Nero AAC 
</hat and cape off>

You now know a second person, i agree with dark sheep, becouse is right for advanced settings, but for me M/S stereo not sounds good, you can me prove that is not true, but for me is not same, nevermind that be mp3/aac other codec which uses M/S, i can prove to you what i hear becouse the differences are very small and not noticable. Dual Channel are not god and not same as real stereo L/R, i am not voting for Dual channel, than for L/R.

for me M/S stereo not sounds good, you can me prove that is not true

I don't trust 99% of the "advanced" users to use these settings correctly, and the past with lamelines and everything (and even this thread) clearly show that it is a bad idea to expose them.

I don't trust 99% of the "advanced" users to use these settings correctly, and the past with lamelines and everything (and even this thread) clearly show that it is a bad idea to expose them.

<hat and cape on>
I'm in a position where I can steer history so we do not end up downloading a garbled 192kbps SBR+PS L/R stereo tune from an artists website, and I will take my responsibility to stop it from happening, at least with Nero AAC 
</hat and cape off>

Quote from: Garf on 2006-05-08 17:29:04

I don't trust 99% of the "advanced" users to use these settings correctly, and the past with lamelines and everything (and even this thread) clearly show that it is a bad idea to expose them.

Although I agree that access to advanced options could lead to garbled encodings by some who think that they know how to use them, there is that 1% of users (probably 0,1%) that could actually help with the improvement of the encoder, by finding settings for example which could provide better quality with certain files. We’ve seen this in the past with lame for example, where certain switches, like “--athaa-sensitivity x” or “--vbr-new” just to name a few, where proven (by that aforementioned 1% of people and verified by others) to offer better quality than the default settings, helping the developers improve the codec.

I’m just saying that limiting the access to the internals could possibly inhibit the faster evolution of the codec and I don’t think that lame’s or Vorbis’ good reputation was ruined by the offer of such advanced options or that there was a big damage done by those very few, who despite all the warnings, use dubious custom command lines.

As a last note, I think that at least it would nice to be able to choose, when to use SBR or PS within certain bit rate limitations of course. Something similar to what the old encoder did.

-artwork      : Adds cover artwork from JPG, PNG or GIF files. If multiple
  <image>     : images are being added to a single album, the assumed sequence
              : is FRONT COVER, INSIDE 1, INSIDE 2, ... BACK COVER

-cdtextcue    : Parses an extended CD-TEXT compliant CUE sheet to determine
  <cue>       : MPEG-4 chapters, album information and song tags.

-channels     : Specifies the desired number of audio channels in the encoded
  <number>    : stream, if this differs from the source.
              : "1" will downmix to mono, by averaging channels.
              : "2" will downmix surround to stereo, or duplicate a single
              : channel as dual-channel mono.

-invert       : Specifies channels to invert. For stereo, can be used
  <chan,chan> : without any additional parameters to switch L and R.
              : For surround, channel pair to be inverted must be specified
              : (LF, RF, LR, RR, FC, P1...).
              : If inverting multiple sets in a surround signal, sets may
              : be given in sequence.
              : EX -> "-invert LF,LR,RR,RF" would switch the LF with the LR,
              : and the RR with the RF.
              : WARNING: It is pointless to use this with MONO! (Lest you ask...)