Music bitrate

[Bubba]

When I convert a music file from WAV to MP3, I noticed the file size is considerably bigger the higher the bitrate. Any suggestions on the best bitrate/quality setting? I assume the larger music file sizes cause slower computers to have problems processing the slideshows. I did a quick test and didn't hear any difference when converting at a lower quality setting, even as low as 32. I understand 128 is 'CD' quality.

[Ken Cox]

< 128 you start to lose richness -- you can see the difference using a program that displays a wave pattern -- windows media player or audacity to name a couple.

from

http://www.google.ca/search?hl=en&ie=UTF-8...le+Search&meta=

see

this site will give you some hint

http://www.iis.fraunhofer.de/amm/techinf/layer3/

ken

[OGL]

Hello, Bubba!

I did a quick test and didn't hear any difference when converting at a lower quality setting,

This is incredible! Quality 32 kbps music is very bad...

What acoustic system you use: home stereo speakers, high quality headerphones,

ordinar headerphones or PC speaker?

Oleg

[Alan Lyons]

Hi Bubba,

The difference will be noticable when you try to play the file over speakers in larger rooms then where you make your shows. as Ken says the richness goes first this difference is like comparing tape to CD. If you use voice it is importent to use the best quality possible as when the voice becomes less clear so does the understanding of what was said. What I do to check my tracks is to burn them toCD or onto tape and play them over the home hi fi. Try listining from another room or part of the house, but remember if you have worked on the soundtrack then you already now how it should sound. Try to have someone else listen to it.

Alan

[Merlin]

It depends on the soundtrack and type of music.

CDs sample at about 41 (thousand bits/second) 16 bit

AM radio samples at 10 (thousand bits/second) 8 bit at best

Telephones sample at 8 (thousand bits/second) 7 bit

What quality of sound do you want for the show?

How will the sound be played?

What does your audience expect?

Merlin

[Bubba]

I have a reasonably good speaker system on my pc with a good sound card. I haven't tested voice yet. I generally will be making slideshows for relatives and friends personal use so big sound isn't much of a factor. Occassionally a slideshow I make may be played for a group in a big room with a projector.

For someone's personal use I wonder if a 96 bitrate (approx. FM stereo radio quality) is about right. I'm curious what others use for most of their shows?

[LumenLux]

Bubba, I choose typically 128, 96, or 64. I will usually use 128 unless there is a compelling reason to keep the final exe file as small as possible. Sometimes when I use lower than 128 I hear the difference - other times I can not. But there may be listeners with better hearing who may always hear the difference. I think also the complexity of the music can make a difference. Finally there is the lazy factor to consider. My latest posting on Beechbrook is That's What Makes Us Sing. It has a sound glich or two that bothered me but I was too lazy to go redo the mp3 file. Give it a listen and see how it seems to you. I suspect I could have eliminated the rough spots even without upping the bitrate. I guess we could expand the game to see if you can guess the bitrate of the subject sound track if you or others care. I think the answer finally is for the show creator to listen and in each case decide what is good enough for the intent of each particular presentation.

[OGL]

I'm curious what others use for most of their shows?

OK, Bubba.

In my shows (beechbrook.com) I use sound files:

"Olga Boritsch..." - bitrate 64kbps, sample rate 22.05kHz (4.4Mb/9min14sec);

"Life is life" - bitrate 56kbps, sample rate 22.05kHz (1.6Mb/3min50sec) - with processing;

"Malajsia.Sipadan..." - bitrate 48kbps, sample rate 16kHz (3.5Mb/10min) - with processing;

Method of the processing sound files you can see in my show "Supercompression soundfiles".

For CD version I use 160kbps 44.1kHz music.

Regards,

Oleg

[LumenLux]

Oleg, I appreciate hearing that from a master sound man. Thank you. Would you have any advice for getting the best possible voice recordings with inexpensive microphones to our pc sound cards?

[Alan Lyons]

Hi All,

The trick with voice is to record with the peaks at about +3 on your meter you then need to adjust the trebble / bass to clean up the voice dont try to do this on headphones as these can sometimes distort the final sound. For voice over you need to have the misic set at about 20db below the voice track,

Alan

[OGL]

To Alan Lyons

For voice over you need to have the misic set at about 20db below the voice track

Your idea is very cooooool! What you think about "Normalise" procedure?

To LumenLux

Hi, Robert.

For voice recordings I use headerphones with microphone HSQ-204p ($16 in Russia)

For record you can use soundcard recorder:

Creatiwe Live!5 Main Panel -> Recorder -> Input: Microphone then press button Rec and speak. :-)

or any sound editor (Audacity, CoolEdit2, SoundForge)

Voice file you can save as WAV or MP3 bit rate 96 - 128 kbps (this is good for editing).

For soundediting I use SoundForge6 or CoolEdit2 (audiomontage in Audacity is mazohism, hi-hi)

Step1- Equalizing

Point 1 20Hz > Inf

Point 2 80Hz > -24db

Point 3 160Hz > -12db

Point 4 200Hz > -6db

Point 5 320Hz > -3db

Point 6 640Hz > -3db

Point 7 1.3kHz > +3db

Point 8 2.5kHz > +3db

Point 9 3.5kHz > -6db

Point10 5kHz > -12db

Point11 10kHz > Inf

This is not universal curve, for my voice only!

Step2 - Denoise ( 2 methods in Forge)

1. Effects --> Noise Gate

Use Presets: NoiseGate1 or NoiseGate2

(adjust "Treshold level")

2. Effects --> Dynamics --> Graphic

Preset: Soft noise gate below -36db

(adjust "Treshold" /-16db to -46db/)

If you use Audacity or Nero WaveEdit, you can use standart procedure NoiseReducte.

Step3 - Postcorrection

This is very individual procedure, see my "Supercompression Soundfile" show and make experiments.

Step4 - Normalise

Without comments

Step5 - Saving.

Voice files for slide show you can save as MP3 32kbps, 48kbps or 56kbps. For compression use Lame codec.

Very good for it is low rite WMA formats.

If my English is bad, I'm sorry!

Regards,

Oleg

[Ken Cox]

Bart has done some digging re optimum bit rate

see

http://community.barcin.net/forums/viewtopic.php?t=34

ken

[LumenLux]

Thank you Oleg for the detailed route. Thank you everyone. There are such great people inn this forum.

[Alan Lyons]

To Oleg,

I think you need to be careful with normalise when you have a mixed track as it can try to find a general level for the whole track. Abit like when you try to darken or brighten an entire image, you've got to compromise somewhere. Sound mixing is like light exposure you need to read what the meter is telling you and use the info to get the right result,

Alan

[OGL]

Hi, Alan.

Sound mixing is like light exposure you need to read what the meter is telling you

Thank you very much this information!

Situation "background music - foregraund voice" is very interesting and ambiguous.

Voice signal have very loose structure and as consequence, have low (or very low) middle

(integral) level, about( -16 to -26)db

And (-26db) + (-20db)=-46db. This music is very-very quiet.

If (-16db)+(-20db)=-36db, it's may be so-so for background.

For (background music+voice) mixing I use next method.

-Normalize voice signal: RMS -10db, Scan setting -30db, Attack time abt 50ms, Release time 200ms.

-Normalize music signal: standard RMS -16db preset.

-For mixing I use AGC (Automatic Gain Control), reference channel is "voice" (treshold abt -26db, attack 1ms,

release 200ms, level of Gain control abt 8-12db).

Or mixing with "handmade" leveling. But it is cooool work!

Regargs,

Oleg

[Alan Lyons]

Hi Oleg,

I think you may be going in too deep on this. What I do is record voice peaking at +3 db on the out out meter ( just in the red) as the old guys would say. My music would be recorded the same on sepreate tracks i.e. l+r. I then make the seperation at mix down keeping the music peak around -17db (+3 -20). This will give a good sound seperation and means you dont need to power up the sound on long non voice sections. If the voice is been taken over by the music try taking some bass off the voice it may sound thin on its own, but when the music is played with it the top end of the voice will be hidden behind the music and it will sound more natural,

are you with me?

Alan