VO distort on iPhone speaker

[mowmow]

Hi,

I'm having a problem with iPhone 3G speaker when playing back my podcast file (just a plain voice track without any music).
It sounds good and never distort on both A5Xs and PCs little speakers but it distort easily on iPhone speaker.

Is there any tricks in general to make it not to distort and make it sounds good on any small speakers even though when you turn it loud?

I use plugins such as oxford EQ, oxford supresser, oxford inflator, oxford limitter etc.
I'm looking at VU meter and adjust the level so that it stays close to 0VU.
I noticed when I put lots of lows (like radio DJ) it distort on small speakers easily but sometimes it distort even when sound is thin. Maybe I have to dip more low-mids? I'm confused.

If there are mastering engineers, I'd appreciate if you can give me a hint.

[playitloudpro]

Try rolling of at 35 htz about 3 db

Also cut at 200 htz by 1 db with narrow q

I do a lot of voice over work and those are some. Of the suggestions I have received here on gs

[Llitsor]

I'd just roll off ALL the low end, as the poxy iPhone speaker is never going to be able to reproduce it anyway. It could well be causing the distortion.
And don't ponce about removing 1 or 2db's.. You'll want to wipe off loads more than that.

[mowmow]

Originally Posted by playitloudpro

Try rolling of at 35 htz about 3 db

Also cut at 200 htz by 1 db with narrow q

I do a lot of voice over work and those are some. Of the suggestions I have received here on gs

Thankx playitloudpro.
I will definately try that.
So I guess some of the low frequencies are problem to small speakers.
I wonder if most mastering engineers think about playing back on such a small speaker like iPhone speaker.
It is almost like a miracle to be able to play back on any speaker and sounds good.
I wish I knew how much low end is enough or too much.

[ReSeRe]

Originally Posted by mowmow

So I guess some of the low frequencies are problem to small speakers.

A-L-W-A-Y-S

[mowmow]

Originally Posted by Llitsor

I'd just roll off ALL the low end, as the poxy iPhone speaker is never going to be able to reproduce it anyway. It could well be causing the distortion.
And don't ponce about removing 1 or 2db's.. You'll want to wipe off loads more than that.

Thanks Llitsor.
I guess I have to seriously clean up the low end.
What if I want lots of low end on the voice?
It is kind of difficult to have lots of low end and not to distort on small speakers.

I remember back in 90s, young engineer mixed rap record with tons of low end. And the mastering guy at Stering shaved off almost all of them. He went to the mastering house and he said "where is my low end men?" LOL
But I though the mastered CD on the radio sounded great. It was a hit record.

So how do you know if it is too much low end? This only comes from experience? I guess there is some trick to shave off the low end and still make it sounds like there is a lots of low end.

[mowmow]

Originally Posted by ReSeRe

A-L-W-A-Y-S

Is there any books that I can read that you recommend?
I really want to know about how to treat low end professionally.

[playitloudpro]

everything i have read about ring tones seems to make me believe anything below 35 htz is of no benefit. VOX wise if you are trying to keep it warm most of what i read says boost gently at 100-110 for more fullness but still notch cut at 200.

The other trick i learned from this site is to do a EQ sweep. It allows you to find the exact frequency that is honky or distorted so you can make cut there.

Also probably no reason to go over -6 db in regards to general level assuming you have compressed it mildly.

Like i said i am no engineer but people on here have been great to me and i read a lot hope it helps even if a little.

one last suggestion is to test the audio in a cheap crap pair of headphones like someone might have for working out. if it manages to sound clean there it may have better chance on phone small speakers

[Chris Wilson]

In my experience, voiceovers will 'clip' crappy computer speakers if the levels peak above about -4dBFS.

For some reason, if the sounds fill the entire audible frequency spectrum, like a Lady Gaga track or something, it does ok up there towards -1 or 0 dB, but for voice only spots, I've found it generally works better to back the levels off a bit.

Unless your voiceover artist is Barry White, there will not likely be much going on at 35Hz.


Very common in voiceover jobs:

80 Hz high pass, 250Hz dip, 2k shelf boost (maybe a little peak at 11k-13k) and maybe a little L1 to get that annoying commercial sound (if it's that kind of job).

Notch around 8K for mouth noises (restoring air above)

Maybe notch 200-350Hz with a wide Q

[mowmow]

Originally Posted by Chris Wilson

In my experience, voiceovers will 'clip' crappy computer speakers if the levels peak above about -4dBFS.

For some reason, if the sounds fill the entire audible frequency spectrum, like a Lady Gaga track or something, it does ok up there towards -1 or 0 dB, but for voice only spots, I've found it generally works better to back the levels off a bit.

It's strange isn't it?
Sometimes thin voice distort all over the place more than big fat voice with more over all levels.
It is difficult to know which type of sound will distort on small speaker without the real test.
In any case, I guess sudden peaks have to be compressed with short attack setting and make it even.

Originally Posted by Chris Wilson

Very common in voiceover jobs:

80 Hz cut, 250Hz dip, 2k shelf boost (maybe a little peak at 11k-13k) and sometimes touch it up with an L1 to get it up in your face (if it's that kind of job).

Notch around 8K for mouth noises (restoring air above)

Maybe notch 200-350Hz with a wide Q

Wow, it's a good tips.
I'll try this. thank you!

I set limiter(on master section) output level to -3 and I still heard distortion on sudden peaks.
I wonder which compressor or limiter plugin is the best for plain VO work.

[Adam Dempsey]

You've already isolated the distortion issue to the built in speaker, so don't use the speaker. Play via an iPhone/iPod dock and take a line out. Optimizing playback of the source is primary. Processing, secondary - always.

[mowmow]

Originally Posted by Adam Dempsey

You've already isolated the distortion issue to the built in speaker, so don't use the speaker. Play via an iPhone/iPod dock and take a line out. Optimizing playback of the source is primary. Processing, secondary - always.

I didn't know iPhone dock had a line out since I don't have it.
What do you mean by optimizing it?
Do you mean mixing using iPhone speaker?

[Jean Doe]

Try mixing through your mac's internal speakers. At times this helps isolating small speakers' problems (until you get experienced enough to predict problems immediately).
The idea that low end is ALWAYS a problem is flawed ime. It just depends at what level the low end has peaks in comparison to the mids and highs and how clean it is i.e. how much distortion and how much dissonance there is.
So in practice, here's a starting point. Hipass filter out everything that sounds wooly. Add a bit more compression than you normally would. Add a bit more deessing than you normally would. And a add a bit of limiting leaving some headroom. Now try playing back on small speakers (like your built in computer speakers) and listen to exactly where it starts distorting. Is it on a vowel that has a bit too much 400hz, then cut there. No sweeping necessary if you have a good ear.

[mowmow]

Originally Posted by Jean Doe

Try mixing through your mac's internal speakers. At times this helps isolating small speakers' problems (until you get experienced enough to predict problems immediately).
The idea that low end is ALWAYS a problem is flawed ime. It just depends at what level the low end has peaks in comparison to the mids and highs and how clean it is i.e. how much distortion and how much dissonance there is.
So in practice, here's a starting point. Hipass filter out everything that sounds wooly. Add a bit more compression than you normally would. Add a bit more deessing than you normally would. And a add a bit of limiting leaving some headroom. Now try playing back on small speakers (like your built in computer speakers) and listen to exactly where it starts distorting. Is it on a vowel that has a bit too much 400hz, then cut there. No sweeping necessary if you have a good ear.

Wow! What a advice! Realy logical down to the point.
Thank you sir.

I read this over and over and practice until I get it in my soul.

[playitloudpro]

The people on gs are top notch I learned from your post too. Thanks

[ReSeRe]

Originally Posted by playitloudpro

everything i have read about ring tones seems to make me believe anything below 35 htz is of no benefit. VOX wise if you are trying to keep it warm most of what i read says boost gently at 100-110 for more fullness but still notch cut at 200.

The other trick i learned from this site is to do a EQ sweep. It allows you to find the exact frequency that is honky or distorted so you can make cut there.

Also probably no reason to go over -6 db in regards to general level assuming you have compressed it mildly.

Like i said i am no engineer but people on here have been great to me and i read a lot hope it helps even if a little.

one last suggestion is to test the audio in a cheap crap pair of headphones like someone might have for working out. if it manages to sound clean there it may have better chance on phone small speakers

nice learns.

@Jean Doe: maybe this looks newbish to you. no problemo.
there is ALWAYS a problem with very low end on the small speakers simply because those speakers CANT play it, so you don't know wtf's going down there. ASAT.
i agree it's better than nothing your advice: try to compress it/level down/low cut. in blind i mean. and God help us all.
and sweep it's doing by a lot of experienced ppl: for search, for study, for compare, for finding the problem. in an analog, clean way.

OP: i'm too lazy to find threads about books there but they're definetly are. pls search.
one piece of advice in low end kingdom: get yourself a GOOD can. it will help a lot. till u'll have a sub or a full range midfield monitor.

[Jean Doe]

Originally Posted by ReSeRe

@Jean Doe: maybe this looks newbish to you. no problemo.
there is ALWAYS a problem with very low end on the small speakers simply because those speakers CANT play it, so you don't know wtf's going down there. ASAT.
i agree it's better than nothing your advice: try to compress it/level down/low cut. in blind i mean. and God help us all.
and sweep it's doing by a lot of experienced ppl: for search, for study, for compare, for finding the problem. in an analog, clean way.

I respectfully disagree. Try using a test tone oscillator playing a sine wave over e.g. your laptop's speakers start the tone at say 700Hz. Turn up your speakers just loud enough to hear that annoying buzzy distortion. Now sweep the frequency down to 80Hz. Notice the buzz disappear.
Now I'm not trying to prove that there are less problems with low frequencies than with mids and highs. What I'm saying is that these problems are related to low end dissonance and untamed peaks. Once those are eliminated (see sinewave test) the problems disappear.

Not really sure what you mean with that sweep comment. "Study"? "Analog clean way"?
All I was trying to say with my previous post (and I wasn't trying to offend you in any way) is that sweeping the spectrum is bad practice (at least in my book) unless it is done to narrow in on a part of the spectrum that one already had identified as being the target before the sweep.

[mowmow]

I've been experimenting with guide lines you gave me in this thread, thank you all.

I noticed that these below can cause distortion on iPhone speaker.
1) sudden peaks
2) average RMS level
3) sustained low end level compare to the other band
4) frequencies of sustained low end.

It seems OK if I set the limiter output level to -2 and average RMS level set to below 0VU or below as long as the sustained low end level and frequency is gentle.

It is difficult to solve #3 and #4 if I want big fat bigger than live voice.
It seems like iPhone don't like sustained lows below 300Hz at all.
iPhone seems to like boosting around 1.4K~2KHz area.
Correct me if I'm wrong.

[Th3_uN1Qu3]

The best thing you can do is buy TrueRTA (you'll need it anyway) and borrow a calibrated mic, then measure the iphone's speaker response. You'll likely find a few nasty peaks that need to be eq'd out, or emphasized (depending on how annoying you want to be).

Most phone loudspeakers will cut below 300Hz. It doesn't really matter if your podcast has a 300Hz cut or not since it won't get played anyway. If it gets distorted when your audio has stuff under 300Hz then you're clipping somewhere. I would strongly recommend against doing a sharp low cut at 300Hz since people with regular computer speakers will find your podcast "nasal".

[mowmow]

Originally Posted by Th3_uN1Qu3

The best thing you can do is buy TrueRTA (you'll need it anyway) and borrow a calibrated mic, then measure the iphone's speaker response. You'll likely find a few nasty peaks that need to be eq'd out, or emphasized (depending on how annoying you want to be).

Most phone loudspeakers will cut below 300Hz. It doesn't really matter if your podcast has a 300Hz cut or not since it won't get played anyway. If it gets distorted when your audio has stuff under 300Hz then you're clipping somewhere. I would strongly recommend against doing a sharp low cut at 300Hz since people with regular computer speakers will find your podcast "nasal".

Thanks Th3_uN1Qu3.
I tried dipping at 300Hz about 3~5db with Q width 2.0.
If I dip too much, it start to sound too thin.
After getting used to this dipping, other sound files which I didn't dip that much start to sound funny.
There are something around this frequency that I don't like about. It's a strange sound area when I boost it.

The trueRTA seems interesting.
I will look into that.

[ReSeRe]

Originally Posted by Jean Doe

I respectfully disagree. Try using a test tone oscillator playing a sine wave over e.g. your laptop's speakers start the tone at say 700Hz. Turn up your speakers just loud enough to hear that annoying buzzy distortion. Now sweep the frequency down to 80Hz. Notice the buzz disappear.
Now I'm not trying to prove that there are less problems with low frequencies than with mids and highs. What I'm saying is that these problems are related to low end dissonance and untamed peaks. Once those are eliminated (see sinewave test) the problems disappear.

Not really sure what you mean with that sweep comment. "Study"? "Analog clean way"?
All I was trying to say with my previous post (and I wasn't trying to offend you in any way) is that sweeping the spectrum is bad practice (at least in my book) unless it is done to narrow in on a part of the spectrum that one already had identified as being the target before the sweep.

no, u didn't offend me in any way.

Clean, analog: trying to say thru a console. the feeling and the result, "in my book" is way better and faster than in software mode. (step 1&2) the diagnosis and the way to go.
yes, if u're sure the rest is good, identify the problem and go only there with a tiny and tinyier Q. better solving is digital due that. (step3)

About the 700HZ/80Hz test.. uhm, well, i got your POV somehow, but disagree.

Have a nice day.

[mowmow]

I just bought iPhone4.
It sounds so different from iPhone3G.
iPhone4 has less distortion but lot more compression sound, it almost sounds annoying.

I guess Apple try to solve distiortion problem with compressor.

[Th3_uN1Qu3]

Not surprising, they've been doing it for a while in their Macbooks. Instead of doing the right thing aka using a proper speaker (just look at Nokia and SE), they try to solve it by throwing more DSP at it. Bass-ackwards if you'd ask me.