Well joined stereo, when they say that, and I quote, "the human ear
cannot hear when the lower frequencies are in such and such a
channel," it is making one hell of an assumption, when, in fact, I
can hear which channel it is, particularly wearing headphones.
Speakers, now that's the rub in that case, they can get away with
phase distortions which that "joined stereo has to be since they are
changing the phase relationships from natural to an electronic
approximation of an algarithm which is, according to some, not
distinguishable from the "real" or "natural" sound. Since, the
sources are as variable as natural sounds, I'd want the natural
sounds to be not modified, since if I am wearing headphones, and the
bass is on the left, I do not want it in the center or the pseudo
center channel as modified by someone's protocol, thank you! :) And,
if given my drothers with speech, I'd wish it mono and not at a
wasted high bit rate unless a particular voice suffers from artifax
as the result of encoding by a particular protocol such as *.ogg,
*.mp3, *.m4a, etc. I'd want the choice to make the recording without
such artifax ladden protocols. :) Thanks guys for letting me vent a bit.
Curtis Delzer
Neal:
Joint Stereo is not necessary in files of 192 KBPS or higher (in my opinion.) If you pick <Default> for our Mode box, you will get Stereo if the bit rate is 192 or higher. Joint stereo can mess things up in some situations, but it really does help at the lower bit rates, like 128 KBPS.
I will probably leave VHQ in there because it is only used to instruct the encoder. Different versions of the LAME encoder may act differently, and some day, they really may have a VHQ mode that is useful. Perhaps they do now; I haven't felt like messing with the newer LAME versions lately. 3.96 was too much of a let-down.
By the way, it isn't always necessary to select high quality. In some situations, like for 48 KBPS spoken word files, i prefer low quality, because it is very fast and almost sounds better.
>>> neal.ewers@xxxxxxxxxxxxxx 06/16/06 02:08PM >>> Well, I have certainly learned some interesting things today.
Results.
Studio recorder on a 1 hour file with white noise and various other high quality sounds. No VBR high quality, and in simple stereo and not joint stereo. Converted all files at 192 KBPS. Time taken. 3 minutes 40 seconds.
Sound Forge with the same file, and the same settings except that I had stereo set to joint.
Are you ready for this? 20 minutes and 23 seconds. I then discovered the error of my ways and set Sound Forge to simple stereo. It still took 20 minutes and 23 seconds. The SF and SR files were within 3 KB of being the same size and I could tell no difference in quality between them.
I then read Rob's message about very high quality MP3 conversion. It was way down at the bottom of the list and I didn't see it the first time. So, I cheated. I used the same file in SR but only 10 minutes of it and multiplied it by 6 to get 60 minutes. When I got to 7 minutes on the 10 minute file, times 6 or 42 minutes, I canceled the operation. Rob, I think you're right. If there is no significant difference, I would remove it. The question is, is there a difference. Even if there is, it ain't worth that long a time to convert in my opinion.
One question though. Here is the definition of Joint stereo that Sound Forge uses. Rob, tell me if this is your understanding of this and if it is needed on a file that is converted at 192 KBPS. Sound forge does some other additional processing as you will see in their description.
if
Joint stereo
Select this radio button if you want the encoder to be able to
take advantage of the similarity between the two channels of
a stereo file. Joint stereo encoding is useful when you want
to create a stereo file but require a relatively low bit rate.
When the radio button is selected, select or clear the
Intensity stereo or MS (mid/side) stereo check boxes to
indicate which encoding methods will be used.
Intensity stereo
In this mode, the encoder can take advantage of the fact that
the human ear does not distinguish location information for
low frequencies as well as it does for higher frequencies.
For high frequencies, magnitude and spatial information is
encoded. For low frequencies, only magnitude and phase are
encoded.
These check boxes are not available in SF in normal stereo mode.
So, there you have it friends.
Neal Ewers Ravenswood Productions Local phone: 608-277-1995 Toll Free: 888-544-8332 Email: <mailto:Neal.ewers@xxxxxxxxxxxxxx> Neal.ewers@xxxxxxxxxxxxxx <http://www.ravenswood.org/> http://www.ravenswood.org
