V.I/VI.I - Cuestions

[Elektra999]

Hello
Steve, V.I/VI.I,
- ¿Which is the maximum level of (decibels) the channels SL, SR, SC?
- ¿Every line is a decibel of Rear Level?

http://img227.imageshack.us/img227/7012/03ch0.jpg

Thanks
Frank

[ursamtl]

Hi Frank,

No these lines do not represent decibals. They are simply an arbitrary reference scale for settings values. It's difficult to assign a dB level to these as the process of generating the surrounds can tend to unmask transients, etc., that phase-cancelled each other out in the regular stereo image. Just be sure to always use a limiter on the outputs and also to keep your inputs from getting too loud. I strongly advocate the use of Replaygain on your source files. This is especially so when using any source that comes from CDs mastered after the early to mid 1990's and definitely anything from the late 1990's on. The "loudness war" has boosted the average level of CDs by about 10dB!

[Elektra999]

Thanks for the additional info ;)

[SealTooGreat]

@ursamtl

So chain filter sould be like:
Remove DC offset-> convert to 24bit (or 32bit) -> 44.1kHz to 48kHz (for DVD surrond project) or skip this step for CD surround Project -> Replaygain (or Normalization) -> V.I VST suit-> limter on each channel.

I see in some guide reverb effect on SL&SR - Is it necessary?

[Elektra999]

Hello

01. Apply 32bits float
02. Downsampling. Example 48.000 to 44.100.
03. Remove DC Offset.
04. Apply Dithering of 48.000/96.000,... to 44.100 (Downsampling), or 96.000 to 48.000.
Upsampling not necessary dithering.
05. Limiter to 0dB and it will eliminate the Clips.
05. Reverb effects. That explains it to you, Ursamlt.

[ursamtl]

@ursamtl

So chain filter sould be like:
Remove DC offset-> convert to 24bit (or 32bit) -> 44.1kHz to 48kHz (for DVD surrond project) or skip this step for CD surround Project -> Replaygain (or Normalization) -> V.I VST suit-> limter on each channel.

I see in some guide reverb effect on SL&SR - Is it necessary?

1. ALWAYS DO THIS FIRST!! Convert to 32 bit. Before you do anything else (even simple level changes), always convert to 32-bits. If not, you will lose data.

2. Remove DCOffset. People sometimes think this has to do with ultra low frequencies, but actually it's an alignment of the waveform that ensures it's using the maximum dynamic range. The step is important because otherwise, you might not be using the maximum resolution or dynamic range, especially if you convert back to 16-bit audio at the end. Low-level signals in 16-bit audo such as cymbal fades and reverb tails, can end up being rendered with only a few bits (like an early-80's game computer) so you want to use as many bits as possible.
[QUOTE=SealTooGreat;1006895]@ursamtl

3. If source is not at the same sample rate as the destination (44.1kHz for surround CD; 48kHz for surround DVD; either is fine for DVD-A), convert the sample rate to the destination rate using the best sample rate converter you can find. After reading some test results, I recommend Adobe Audition or Cool Edit Pro for commerical software packages, and SSRC or r8brain for free ones.

4. Apply replaygain or group waveform normalization.

5. V.I with limiter on outputs.

If you want reverb on the Sl & SR, go for it. It's not necessary, but depending on the source material, it may give you a better sense of envelopment in your final surround file.

6. Reduce the bit depth (also known as bit resolution or word length) of your output file(s). If converting to 24-bit file, dither is not necessary. IMPORTANT: If converting output files to 16-bits, make sure you ALWAYS use dithering. The consensus from my readings and discussions with audio pros seems to be that triangular dither is best. Noise shaping is also preferred.

Regards,
Steve.

[SealTooGreat]

4. Apply replaygain or group waveform normalization.


Could you explain group waveform normalization?
I'm asking this 'cause what if I don't have replaygain in my app and have only normalization option.
Does group waveform normalization differ from normalization and how to achieve the same with usual normalization?

Well I know, you've explained what is replaygain here (http://forum.doom9.org/showthread.php?p=1003996#post1003996), but I was wondering how to get the same effect without using to many apps only because of specify command if I can do the same without them.

btw thanks for the tips! :)

[ursamtl]

Group Waveform Normalization is an option in Cool Edit Pro and Adobe Audition. If you don't have either of these, I recommend a program called Wavegain, which does Replaygain. You can also use Foobar2000 but it does not apply replaygain to wave files, you have to first compress them to a lossless format such as FLAC, APE or WV. As long as you use one of these formats, you won't lose any data or sound quality. Regular normalization isn't the best approach because it only adjusts the peaks of a file to a certain level. If the file has no big peaks, the result will be too loud. If the file's peaks are very high, the result will not be loud enough.

[Elektra999]

Hello
A question:

audio 5.1 for film, channels SL - SR - LFE ¿adjust to -3dB?

[ursamtl]

If I remember correctly, the -3dB adjustment is not necessary unless you are taking audio from a professional movie soundtrack in the studio. For home use, you don't need it. I believe this is because the surrounds were boosted by 3dB on movie soundtracks to account for the large size of movie theaters. However, if you're converting stereo to surround, the 3dB boost didn't happen in the first place so you don't need to cut by 3dB. My advice is always to adjust the surrounds to sound good (if you have a multichannel setup to listen while adjusting). Ideally, you raise the level of the surrounds until they are apparent. Then reduce the level a bit. The surrounds should reinforce the sound coming from the front and make everything sound more three-dimensional (or actually 2D since no height is involved). You should hear the space around the instruments or voices, as if they were real in front of you.

[Elektra999]

Thanks

I'm sorry. Is that the levels, of the channels SL SR (of V.I/VI.I), seems me that they listen themselves high(for audio of film). And for not adjust Rear Level of V.I/VI.I, thought about adjusting to -3dB from Hardware Meters of Sony Sound Forge 9.

[ursamtl]

Elektra,

Since every source is different, it's always best to set the levels to whatever sounds the best. After all, what really matters in the end is how the final thing sounds. Let your ears be the judge.

[Elektra999]

Hi Steve.
¿of how much it is the degree of width, of the channels created by V.I?

It is that, in Sony Sound Forge 9, (Process / Channel Convert), in some of the Presets, [Sys] Stereo to Stereo - Vocal Cut (remove center material) remove the vocal, but I believe, that, of how much it is the degree of width for the channels FL and FR.

http://img517.imageshack.us/img517/2034/dibujorw0.jpg

[ursamtl]

Hi Elektra,

The front width correction controls on V.I are roughly 50% to 150%

[Elektra999]

Thanks for the reply :)

I believe that these channels, selecting the Preset; Stereo to Stereo - Vocal Cut (remove center material), can leave 50% or 60% and -50% or -60% for the channels Surround, with some effect Reverb and with a delay of 10 or 16ms in the channel Right.

Extent seems to 100 % and -100 % to be too much for the channels Surround.

[Elektra999]

Now I am going to practise with the channels FL and FR 50 % to 150 %, to see...

Thanks

[Elektra999]

Ursamlt
¿How I can direct the channels fL fR, towards the center of the central channel?

I am proving my new method 5.1 Surround, with Sony Sound Forge 9 (http://forum.doom9.org/showthread.php?p=1075935#post1075935) , and try, to create an environment Surround.

Thanks

[ursamtl]

A simple way would be to combine about 25-50% of fL plus 25-50% of fR. You'll have to adjust this to give you the most balanced sound.

To duplicate what's done in VI.I you'd need to read up on Ambisonic theory then plot the positions of a dodecahedron II layout on a horizontal plane superimposed over an ITU 5.1 layout then calculate all the virtual connections required to derive the rear channels, taking into account the speed of sound. There were over 300 connections inside the original V.I layout, so about 150 of these go into the rears, so about 50 for one channel. That might sound like a bunch of stuff, but it's an actual description of what I did when I developed V.I! For VI.I, I simply tapped the center back connections required.

[Elektra999]

Thanks Ursamlt

I go years, trying to do (for my same one) 5.1 Surround, but the results that I obtain, are unsatisfactory.

And even more, when I try, to be employed at the discreet channels, buffff.
But, my addiction multichannel, it breaks barriers ... je je ;)