Subject:audio for tv -- bulletproofing the mix

Posted by: BobCole

Date:5/26/2000 7:37:00 PM

How do I use SF to prepare an audio track for television?

I'm a long-time video editor who has had to do more and
more sound work, particularly for low-budget tv spots.
They have sometimes gotten mangled by the stations'
processing and I'd like to learn how to bulletproof my
mixes.

New to this forum, but I already found the discussion of
normalizing particularly helpful, because it didn't assume
ANY knowledge! If you can reply to any of my questions,
please don't assume any knowledge on my part.

1 - Which Sound Forge tools, in which order, and which
settings within them, can help me to bulletproof a mix?
Here is the best procedure I've come up with so far.
Comments/criticisms are welcome.

(a) Paragraphic EQ: shelf off <50 ?
Also, do you recommend routinely boosting gain at certain
frequencies here? If so, which ones?

(b) Normalize - Peak, Average or both? Peak: Is Peak
somewhat irrelevant, given that I have a good basic level
already, and the audio will be mastered onto a master
videotape with its own level settings? Average RMS: Is
there a good general average setting, or do I set this to
compensate for the scan? (In a video, unlike a musical
piece, the volume can change significantly. So is the
average RMS scan that relevant for a video?)

(c) Dynamics - I've been using Graphic -- Is there a
specific preset which works best for tv? Ditto with output
gain/reduction setting?

Brace yourself for a really dumb question -- Why is there
an output gain/reduction slider in Graphic Dynamics? I can
hear the bigger presence after compression, but aside from
the difference in volume... I don't hear the difference.
When I implement a hard 6 dB limit in Graphic Dynamics,
sometimes it looks as if almost every moment is being
limited. I assume this is a good sign that my volume
setting in Graphic Dynamics should be set to decrease
rather than boost.

General guidelines would be much appreciated. I've been
A/B testing a mix which sounded bad on tv, and it sounds
fine with a variety of settings in Dynamic Graphics. Do I
have to make it sound horribly bright in order to make it
sound good on tv?

2 - What can you determine sheerly from the visual
representation of the track in Sound Forge? Using heavy
compression / limiting results in an almost uniform heavy
black line -- is that a good thing (looks ugly, so my
visual sense says it must be bad...).

Does a fairly uniform waveform do better in tv station
processing than a more up-and-down waveform?

3 - Is there a specific maximum volume inside Sound Forge
to strive for?

4 - Would using bad speakers help me to anticipate the
impact of the tv station's evil companding?

5 - How do I deal with a spot in which there are some heavy
percussive hits (in this case, tennis balls being hit
HARD). A recent spot with such hits got muddied badly
(although I don't know whether a bassy music track was to
blame or the big hits). I've tried lowering the volume in
the mix just before the hit, so as to make a lower-volume
hit sound somewhat impressive. But is there something else
I can do in SF?

Thanks Very Very much for wading through this. My ears are
worn out trying to figure this on my own.

Bob Cole

Subject:Re: audio for tv -- bulletproofing the mix

Reply by: althoff

Date:6/25/2000 10:17:00 AM

Bob, I recently did music and sfx for our game company ad on local
TV, and it seems to be pretty much what you're doing too. I'll try to
give you my $2 on mixing TV-targetted audio.

First: There is no such thing as dynamics in TV audio. Everything
should be at top, top, top volume. I've found that a limiting
compressor really can't cope with the job, you need something more
ruthless than this. The Steingberg Mastering Edition pack contains a
Loudness Maximizer which works OK, the Waves Ultramaximizer seldom
fails and also Timeworks has one, although I can't seem to remember
the name. I didn't apply much compression until mastering, though.

As for the music and sfx, I tried to keep the target system very lo-
fi, meaning elements with similar frequency charachteristics cannot
share the same timespace. If there's a voiceover and a roaring car on
at the same time, they will garble each other. I tried to keep it in
mind when doing the music. An intense techno pump with some distorted
guitars was the only elements that made it into the final mix, along
with some synthetic pads for the slow moments. The rest was cut out,
not to make the piece sound too garbled.

Mixing this was not easy. The target system has no sub-100hz
capabilities whatsoever, but the mix still had to pump with almost no
bass available. I ended up pitching the song +3 semitones to get the
punch that I needed, as I had by routine mixed the song too bassy. I
normally do floor pieces, meaning there's bass in abundance.

Also, I found that all my hihats and bright frequencies quickly
became distorted, so I had to hi-shelf, EQ and compress them heavily.
In the end, nothing significant to the mix was brighter than 6 000
hz. IMHO it decreased audio quality, but at least I was safe.

Moving on to mastering the piece, I put another 3dB peak slightly
below 100 hz, just in case the viewer had a above-middle range TV
set. Also, some EQ at 3 000 Hz seemed to sharpen the mix.

Finally, I processed the mix with a slight stereo enhancer, some Bass
Maximizer with is in essence a compressor and harmonics generator for
lower frequencies, and finally patched this through the Loudness
Maximizer, to completely wipe out all traces of dynamics in the poor
mix.

Needless to say, the mix sucked, but it did work perfectly for the TV
ad. Also, I noticed that while I think I'd overdone the whole thing,
when listening to the other spots I realized I could have added twice
as much bass and treble and still not be the worst cat in the alley.

So, I guess my tip would be: Don't underdo it. More is better, at
least in the school of commercial spots.

See ya!

Johan Althoff
Sound Designer
O3 Games AB
www.outforce.com





Bob Cole wrote:
>>How do I use SF to prepare an audio track for television?
>>
>>I'm a long-time video editor who has had to do more and
>>more sound work, particularly for low-budget tv spots.
>>They have sometimes gotten mangled by the stations'
>>processing and I'd like to learn how to bulletproof my
>>mixes.
>>
>>New to this forum, but I already found the discussion of
>>normalizing particularly helpful, because it didn't assume
>>ANY knowledge! If you can reply to any of my questions,
>>please don't assume any knowledge on my part.
>>
>>1 - Which Sound Forge tools, in which order, and which
>>settings within them, can help me to bulletproof a mix?
>>Here is the best procedure I've come up with so far.
>>Comments/criticisms are welcome.
>>
>>(a) Paragraphic EQ: shelf off >Also, do you recommend routinely boosting gain at certain
>>frequencies here? If so, which ones?
>>
>>(b) Normalize - Peak, Average or both? Peak: Is Peak
>>somewhat irrelevant, given that I have a good basic level
>>already, and the audio will be mastered onto a master
>>videotape with its own level settings? Average RMS: Is
>>there a good general average setting, or do I set this to
>>compensate for the scan? (In a video, unlike a musical
>>piece, the volume can change significantly. So is the
>>average RMS scan that relevant for a video?)
>>
>>(c) Dynamics - I've been using Graphic -- Is there a
>>specific preset which works best for tv? Ditto with output
>>gain/reduction setting?
>>
>>Brace yourself for a really dumb question -- Why is there
>>an output gain/reduction slider in Graphic Dynamics? I can
>>hear the bigger presence after compression, but aside from
>>the difference in volume... I don't hear the difference.
>>When I implement a hard 6 dB limit in Graphic Dynamics,
>>sometimes it looks as if almost every moment is being
>>limited. I assume this is a good sign that my volume
>>setting in Graphic Dynamics should be set to decrease
>>rather than boost.
>>
>>General guidelines would be much appreciated. I've been
>>A/B testing a mix which sounded bad on tv, and it sounds
>>fine with a variety of settings in Dynamic Graphics. Do I
>>have to make it sound horribly bright in order to make it
>>sound good on tv?
>>
>>2 - What can you determine sheerly from the visual
>>representation of the track in Sound Forge? Using heavy
>>compression / limiting results in an almost uniform heavy
>>black line -- is that a good thing (looks ugly, so my
>>visual sense says it must be bad...).
>>
>>Does a fairly uniform waveform do better in tv station
>>processing than a more up-and-down waveform?
>>
>>3 - Is there a specific maximum volume inside Sound Forge
>>to strive for?
>>
>>4 - Would using bad speakers help me to anticipate the
>>impact of the tv station's evil companding?
>>
>>5 - How do I deal with a spot in which there are some heavy
>>percussive hits (in this case, tennis balls being hit
>>HARD). A recent spot with such hits got muddied badly
>>(although I don't know whether a bassy music track was to
>>blame or the big hits). I've tried lowering the volume in
>>the mix just before the hit, so as to make a lower-volume
>>hit sound somewhat impressive. But is there something else
>>I can do in SF?
>>
>>Thanks Very Very much for wading through this. My ears are
>>worn out trying to figure this on my own.
>>
>>Bob Cole
>>

Subject:Re: audio for tv -- bulletproofing the mix

Reply by: MacMoney

Date:6/25/2000 11:33:00 AM

Hi Bob
I think Johan hit the nail on the head. What I also would try is to
test the mix on a TV with a VCR. When we do TV stuff we always test
the mix by tracking from SF to a VCR and playing the 30 or 60sec spot
on a cheap TV we picked up from a pawn shop,This way you can go back
to make changes in the mix.Good luck

George Ware

Johan Althoff wrote:
>>Bob, I recently did music and sfx for our game company ad on local
>>TV, and it seems to be pretty much what you're doing too. I'll try
to
>>give you my $2 on mixing TV-targetted audio.
>>
>>First: There is no such thing as dynamics in TV audio. Everything
>>should be at top, top, top volume. I've found that a limiting
>>compressor really can't cope with the job, you need something more
>>ruthless than this. The Steingberg Mastering Edition pack contains
a
>>Loudness Maximizer which works OK, the Waves Ultramaximizer seldom
>>fails and also Timeworks has one, although I can't seem to remember
>>the name. I didn't apply much compression until mastering, though.
>>
>>As for the music and sfx, I tried to keep the target system very lo-
>>fi, meaning elements with similar frequency charachteristics cannot
>>share the same timespace. If there's a voiceover and a roaring car
on
>>at the same time, they will garble each other. I tried to keep it
in
>>mind when doing the music. An intense techno pump with some
distorted
>>guitars was the only elements that made it into the final mix,
along
>>with some synthetic pads for the slow moments. The rest was cut
out,
>>not to make the piece sound too garbled.
>>
>>Mixing this was not easy. The target system has no sub-100hz
>>capabilities whatsoever, but the mix still had to pump with almost
no
>>bass available. I ended up pitching the song +3 semitones to get
the
>>punch that I needed, as I had by routine mixed the song too bassy.
I
>>normally do floor pieces, meaning there's bass in abundance.
>>
>>Also, I found that all my hihats and bright frequencies quickly
>>became distorted, so I had to hi-shelf, EQ and compress them
heavily.
>>In the end, nothing significant to the mix was brighter than 6 000
>>hz. IMHO it decreased audio quality, but at least I was safe.
>>
>>Moving on to mastering the piece, I put another 3dB peak slightly
>>below 100 hz, just in case the viewer had a above-middle range TV
>>set. Also, some EQ at 3 000 Hz seemed to sharpen the mix.
>>
>>Finally, I processed the mix with a slight stereo enhancer, some
Bass
>>Maximizer with is in essence a compressor and harmonics generator
for
>>lower frequencies, and finally patched this through the Loudness
>>Maximizer, to completely wipe out all traces of dynamics in the
poor
>>mix.
>>
>>Needless to say, the mix sucked, but it did work perfectly for the
TV
>>ad. Also, I noticed that while I think I'd overdone the whole
thing,
>>when listening to the other spots I realized I could have added
twice
>>as much bass and treble and still not be the worst cat in the alley.
>>
>>So, I guess my tip would be: Don't underdo it. More is better, at
>>least in the school of commercial spots.
>>
>>See ya!
>>
>>Johan Althoff
>>Sound Designer
>>O3 Games AB
>>www.outforce.com
>>
>>
>>
>>
>>
>>Bob Cole wrote:
>>>>How do I use SF to prepare an audio track for television?
>>>>
>>>>I'm a long-time video editor who has had to do more and
>>>>more sound work, particularly for low-budget tv spots.
>>>>They have sometimes gotten mangled by the stations'
>>>>processing and I'd like to learn how to bulletproof my
>>>>mixes.
>>>>
>>>>New to this forum, but I already found the discussion of
>>>>normalizing particularly helpful, because it didn't assume
>>>>ANY knowledge! If you can reply to any of my questions,
>>>>please don't assume any knowledge on my part.
>>>>
>>>>1 - Which Sound Forge tools, in which order, and which
>>>>settings within them, can help me to bulletproof a mix?
>>>>Here is the best procedure I've come up with so far.
>>>>Comments/criticisms are welcome.
>>>>
>>>>(a) Paragraphic EQ: shelf off >>>Also, do you recommend routinely boosting gain at certain
>>>>frequencies here? If so, which ones?
>>>>
>>>>(b) Normalize - Peak, Average or both? Peak: Is Peak
>>>>somewhat irrelevant, given that I have a good basic level
>>>>already, and the audio will be mastered onto a master
>>>>videotape with its own level settings? Average RMS: Is
>>>>there a good general average setting, or do I set this to
>>>>compensate for the scan? (In a video, unlike a musical
>>>>piece, the volume can change significantly. So is the
>>>>average RMS scan that relevant for a video?)
>>>>
>>>>(c) Dynamics - I've been using Graphic -- Is there a
>>>>specific preset which works best for tv? Ditto with output
>>>>gain/reduction setting?
>>>>
>>>>Brace yourself for a really dumb question -- Why is there
>>>>an output gain/reduction slider in Graphic Dynamics? I can
>>>>hear the bigger presence after compression, but aside from
>>>>the difference in volume... I don't hear the difference.
>>>>When I implement a hard 6 dB limit in Graphic Dynamics,
>>>>sometimes it looks as if almost every moment is being
>>>>limited. I assume this is a good sign that my volume
>>>>setting in Graphic Dynamics should be set to decrease
>>>>rather than boost.
>>>>
>>>>General guidelines would be much appreciated. I've been
>>>>A/B testing a mix which sounded bad on tv, and it sounds
>>>>fine with a variety of settings in Dynamic Graphics. Do I
>>>>have to make it sound horribly bright in order to make it
>>>>sound good on tv?
>>>>
>>>>2 - What can you determine sheerly from the visual
>>>>representation of the track in Sound Forge? Using heavy
>>>>compression / limiting results in an almost uniform heavy
>>>>black line -- is that a good thing (looks ugly, so my
>>>>visual sense says it must be bad...).
>>>>
>>>>Does a fairly uniform waveform do better in tv station
>>>>processing than a more up-and-down waveform?
>>>>
>>>>3 - Is there a specific maximum volume inside Sound Forge
>>>>to strive for?
>>>>
>>>>4 - Would using bad speakers help me to anticipate the
>>>>impact of the tv station's evil companding?
>>>>
>>>>5 - How do I deal with a spot in which there are some heavy
>>>>percussive hits (in this case, tennis balls being hit
>>>>HARD). A recent spot with such hits got muddied badly
>>>>(although I don't know whether a bassy music track was to
>>>>blame or the big hits). I've tried lowering the volume in
>>>>the mix just before the hit, so as to make a lower-volume
>>>>hit sound somewhat impressive. But is there something else
>>>>I can do in SF?
>>>>
>>>>Thanks Very Very much for wading through this. My ears are
>>>>worn out trying to figure this on my own.
>>>>
>>>>Bob Cole
>>>>