Confused.. What about the old encoders?

[Kamen]

Hello,

I would like to ask you about the old versions... As far as I know, mppenc 1.1 is the only one release version, right? Is this mean it is a stable one? I also have the version 1.7.9c (year 2001) of Andree Buschmann... What about them and their quality, their bugs? Are they still good enough to be used? I see you making new and new versions, fixing bugs and changing features. Do some of the bugs fixed in 1.14 - 1.15v occur only in these versions and are new to the old ones, or they also have them? And the most important: Is the sound quality the same? It is difficult for me to say that, because currently I don't have quality equipment. I read Andree Buschmann have written that there is almost nothing more to do about the quality and it is already developed? Do these versions, especially 1.1, really produce the same sound quality?

Thank you very much!

--
Best regards,
Whistler

[Shy]

Much has changed since version 1.1. Bugs were fixed and especially quality was improved. 1.1, although it's a "final" number, and 1.14, although it's a "beta" number, are less stable and are of lower quality than 1.15v, which is the latest "alpha" release. Bugs that existed as far back as in Andree's old versions are fixed in 1.15v. We recommend only 1.15v due to what I mentioned.
Andree wrote a long time ago "In terms of quality the encoder has almost reached the final state." That was a wrong assumption. As was proven later, the old sayings stick to this as well. If you check hard enough, you'll keep finding something to improve, and no matter how good something is, it can always be improved further. That was definitely shown in later versions after Frank Klemm took over the project. Also, decoding and encoding were immensely sped up.
To sum things up, in the words of Frank, "1.14 is uninteresting".

Edit: Keep in mind that 1.14 which is the last "beta" version is from when the Musepack encoder was still not widely made open source. Since it was made open source, we have been able to fix several important things. Musepack's recent quality assurance is stricter than what it used to be, thus the current "alpha" status isn't related to previous encoder status, and that's why it doesn't mean it's of less quality than previous releases.

[Kamen]

Thanks, that's a perfect answer!
And one more question...

I am happy of the fact that almost all MPC's options can be manually configured, it is very useful sometimes depending on the music, for very fine tune, if one knows for what they are (especially if is interested in acoustics and psychoacoustics)... But I did a test: I entered manually all options of --quality 10, of --xtreme, etc., but the encoder always sees them as "Normal Profile". For example:

===
D:\MPC>mppenc --bw 22050 --ans 6 --cvd 2 --ms 15 --ltq 599 --nmt 14 --tmn 33 --tmpMask 1 --minSMR 6 --ltq_max 46 --ltq_gain -30 --verbose --verbose jim.wav
MPC Encoder 1.15v --Alpha-- (C) 1999-2005 Buschmann/Klemm/Piecha/MDT

mppenc: Output file 'jim.mpc' already exists, overwrite (Y/n)? YES

encoding file 'jim.wav'
to file 'jim.mpc'

SV 7.0 + XLevel coding, Profile 'Standard'

PCM fader : fade-in: 0.00 s, fade-out: 0.00 s, shape: 1
Scaling input by : left 1.00000, right: 1.00000
Maximum encoded bandwidth: 22.0 kHz
Adaptive Noise Shaping : max. 6th order
Clear Voice Detection : Dual
Mid/Side Stereo : Mid/Side Stereo when superior + enhanced (3/oo dB)
Threshold of Hearing : Model 599: min(Klemm,Busch), Max ATH: 46 dB, Offset:
-30 dB, +Offset@20 kHz:-29 dB
Noise masks Tone Ratio : 14.0 dB
Tone masks Noise Ratio : 33.0 dB
Exploitation of temporal post masking
Minimum Signal-to-Mask : 6.0 dB
No deleting of input file after encoding

%|avg.bitrate| speed|play time (proc/tot)| CPU time (proc/tot)| ETA
16.7 335.6 kbps 1.84x 0:38.0 3:47.7 0:20.7 2:03.8 1:43.1
===

Am I missing some option, maybe some of the "Stuff Settings"? Or some values are just rounded...?

[Shy]

The encoder sees a normal profile whether you use custom values or not. (in the above case --standard as it's the default when no profile is entered)
That's not much of a problem because as you probably know, we definitely don't recommend manual "tweaking" for normal use. The various values are available for completeness of the encoder, but are definitely not meant for regular use/users. Almost no cases exist where you're better off inserting manual values. In almost any case the result will be no change of perceived quality with the price of (often much) higher bitrate than needed or reduced quality, which will usually also come with a higher bitrate extra .
And as I can see, that horrible line gave you quite a bitrate..

[Kamen]

Yep, it is logical and both audible and visible that the psychoacoustical algorithm is perfectly adjusted and computed, so if one wants to enter values manually, can compete with the precise presets maybe only if he/she has enough time for own computations... and won't be better.
BTW, the --quality 10 switch gives the same bitrate: 338 to 348 kbps.

Well, I hope everything is perfect, 'cause I plan to rip and archive several CDs in .mpc with this encoder.

Thanks for the support!

[Kamen]

Hmmm, btw it looks like the encoder can see the manual settings correctly - it sees them as "insane", "xtreme", "radio", "telephone", etc., when they are around the margins of the corresponding presets. I have forgotten and missed some manual switches before, I noticed it when I began thinking over some desired model and the full exploitation of them to achieve it, according to the desired quality. And then, without wanting it, the encoder saw them correctly. Just to be accurate...
And... I don't want to be pedant, but maybe it is good idea to correct the cycling if one enters stereo mode which doesn't exist - the encoder begins cycling in the window with tons of messages "Unknown stereo mode". Well, that's a cosmetic fix only and isn't important. I forgot to write about this before.

Sincerely,
K. Sabev

[Shy]

It's a better idea to not enter wrong settings but yes, it's something to to fix.