Audio Test Feedback

[trueindie]

Hey guys,

Can you take a listen to these Edit:*Three* files in this link and let me know which one sounds better, or if they all sound horrible? I'm doing this as a test to show someone to get some audio/video podcast work, so any advice and help will be truly appreciated.

http://torontopicturecompany.com/temp/
Audiotest1 doesn't start until like 10 seconds.

Also APE and Alcove, if you guys have some advice on how to best treat sound for the web, that would be greatly appreciated.

Best,
Aveek

 

[AudioPostExpert]

Can you take a listen to these Edit:*Three* files in this link and let me know which one sounds better, or if they all sound horrible?

They all sound moderately horrible! Probably #4 sounds the least bad/most natural but to be honest I would clean it up a little and apply a fair amount of EQ. It sounds a bit like it's been recorded in a padded coffin!

if you guys have some advice on how to best treat sound for the web, that would be greatly appreciated.

With the web, you've often got to contend with devices which have extremely poor audio reproduction, smartphones, etc. For this reason, it's wise to keep the volume up by restricting the dynamic range and also bare in mind that anything below about 160Hz is simply not going to be reproduced by some/many playback devices. You shouldn't remove everything below 160Hz (some people may be listening with earphones or other playback systems) but just make sure there's nothing too important in those lower freqs. A reasonable target for something like Youtube would be about -16LKFS, which is about 8dB higher than broadcast specs and no peaks above -2dBTP (same as broadcast specs).

G

 

[trueindie]

They all sound moderately horrible! Probably #4 sounds the least bad/most natural but to be honest I would clean it up a little and apply a fair amount of EQ. It sounds a bit like it's been recorded in a padded coffin!

Padded Coffin sounds like a compliment . I got accused of recording this in an "empty concrete cube." So, I'll take padded coffin.

When you say apply EQ, what does that mean? I already applied EQ. I took out the low frequencies. I raised up the high frequencies. If I may ask, would you make this drier, and lower the mid-frequencies, as that's what I suspect you would do? I'm not sure that I'm explaining this right, as I'm unfamiliar with any technical lingo, but I hope you're able to understand it. Maybe I'll post a picture of the eq applied later today.

With the web, you've often got to contend with devices which have extremely poor audio reproduction, smartphones, etc. For this reason, it's wise to keep the volume up by restricting the dynamic range and also bare in mind that anything below about 160Hz is simply not going to be reproduced by some/many playback devices. You shouldn't remove everything below 160Hz (some people may be listening with earphones or other playback systems) but just make sure there's nothing too important in those lower freqs. A reasonable target for something like Youtube would be about -16LKFS, which is about 8dB higher than broadcast specs and no peaks above -2dBTP (same as broadcast specs).
G

Thank you

From the feedback I've gotten, it seems to me that certain aspects of sound are a matter of taste and not about correct or incorrect. Some people like reverb, some people hate it. Some people like the dry sounds they remember from the 70s radio stations, and prefer that, and some prefer a little more wetness. For myself, I seem to prefer the sound a bit wet, not too much, just a touch.

I'll post a picture of the EQ later today. Thanks Ape!

BTW: #s 2,3 & 4 are basically, #1 procesed. & #4 is #3 de-gated.
Any chance you'd like to take a stab at it APE? I'd love to hear what you'd do to it. Just a thought .

 

[trueindie]

Here's what I did for #4, apart from some further de-humming and spectral work

EQ

Dynamics (what does that even mean? why are dynamics called dynamics???)

De-esser

Limiter

My guess is that veteran audio guys prefer this EQ. Makes it sound more like radio

Cheers.
Aveek

 

[AudioPostExpert]

Any chance you'd like to take a stab at it APE? I'd love to hear what you'd do to it. Just a thought .

I would take a look at it for you but unfortunately at the moment I can't. A couple of weeks or so is probably the earliest I could get round to it.

EQ

Mmmm. The HPF is OK but I'd have done the rest of it quite differently. I can't say exactly what I'd do without actually playing with it myself but there's certainly a lot of resonance in the lower freqs I'd want to substantially reduce, probably around 250Hz-400Hz. I also would not have boosted the HF as you have done and would probably have added a little more presence somewhere around 700Hz-1.2kHz.

Dynamics (what does that even mean? why are dynamics called dynamics???)

They're called "dynamics" processors because they directly and primarily process/affect the dynamic range.

The compressor is a little harsh, too high a ratio and too high a threshold and the gate is also very harsh, particularly on #2. Although again I'd need to play with it myself to know exactly what settings to use. The limiter is probably OK although as it doesn't appear to be a True Peak limiter I'd maybe be a little more conservative on the peak level, maybe -2.5dB.

My guess is that veteran audio guys prefer this EQ. Makes it sound more like radio

Nope, if anything that's pretty much the opposite of what a veteran audio guy would likely prefer for radio! As a general rule, there is no general rule about what EQ to apply! Usually I will start with a HPF set at 80Hz (say 12dB per octave) and a LPF set at 12kHz but that's just a starting point. The filter settings I end up with and the specific EQ boosts and cuts are a function of the resonances of the particular voice, along with the type and positioning of the mic. In other words, it's down to what I'm hearing and experience. In general though I'm aiming to cut freqs with EQ rather than boost.

Remember also that the order in which you apply your processing makes a difference. For example, you are greatly boosting the high freqs (above 2kHz), which is where the essing lives and then applying a lot of de-essing, in effect fighting the processing you've already applied!

If I may ask, would you make this drier, and lower the mid-frequencies, as that's what I suspect you would do?

I'm not sure I'd try and make it any drier than it already is and if anything I'd be looking to add back in some mid freqs but as I mentioned above, I'd also want to reduce some of the lower resonances. Remember that adding reverb is in effect smearing the dialogue, adding reflections which effectively muddy the dialogue and cause a loss of clarity. Something I'd be wary of doing if the audience is likely to be listening on low quality playback devices like smartphones which already suffer from a lack of clarity. On the other hand, completely dry dialogue tends to sound unnatural, so it's a judgement call but if you are going to add any reverb I would err on the conservative, less is more, side.

G

 

[trueindie]

I would take a look at it for you but unfortunately at the moment I can't. A couple of weeks or so is probably the earliest I could get round to it.

Heh, heh. I was just kidding. Not necessary. I appreciate the time you've already spent listening to it.

Mmmm. The HPF is OK but I'd have done the rest of it quite differently. I can't say exactly what I'd do without actually playing with it myself but there's certainly a lot of resonance in the lower freqs I'd want to substantially reduce, probably around 250Hz-400Hz. I also would not have boosted the HF as you have done and would probably have added a little more presence somewhere around 700Hz-1.2kHz.

Interesting. I'll try that out. Personally, I seem to like the HF. Makes it sound a little "tinny?" My ears seem to enjoy that sound. But I suppose it's not about what I enjoy? that really sucks.

They're called "dynamics" processors because they directly and primarily process/affect the dynamic range.

Oh, that makes so much sense. Thanks.

The compressor is a little harsh, too high a ratio and too high a threshold and the gate is also very harsh, particularly on #2. Although again I'd need to play with it myself to know exactly what settings to use. The limiter is probably OK although as it doesn't appear to be a True Peak limiter I'd maybe be a little more conservative on the peak level, maybe -2.5dB.

Got it. I guess I'll take the gate off completely. I've come to the conclusion that it makes dialogue sound worse. I thought I had reduced it to the point where it was only slightly affecting the sound. I thought I had let in enough of the noise floor. But I think I'll just take it off completely.

Nope, if anything that's pretty much the opposite of what a veteran audio guy would likely prefer for radio! As a general rule, there is no general rule about what EQ to apply! Usually I will start with a HPF set at 80Hz (say 12dB per octave) and a LPF set at 12kHz but that's just a starting point. The filter settings I end up with and the specific EQ boosts and cuts are a function of the resonances of the particular voice, along with the type and positioning of the mic. In other words, it's down to what I'm hearing and experience. In general though I'm aiming to cut freqs with EQ rather than boost.

Remember also that the order in which you apply your processing makes a difference. For example, you are greatly boosting the high freqs (above 2kHz), which is where the essing lives and then applying a lot of de-essing, in effect fighting the processing you've already applied!

I'm not sure I'd try and make it any drier than it already is and if anything I'd be looking to add back in some mid freqs but as I mentioned above, I'd also want to reduce some of the lower resonances. Remember that adding reverb is in effect smearing the dialogue, adding reflections which effectively muddy the dialogue and cause a loss of clarity. Something I'd be wary of doing if the audience is likely to be listening on low quality playback devices like smartphones which already suffer from a lack of clarity. On the other hand, completely dry dialogue tends to sound unnatural, so it's a judgement call but if you are going to add any reverb I would err on the conservative, less is more, side.
G

Extremely insightful and helpful. Thank you sir!
Have a great evening.
Aveek

 

[trueindie]

The limiter is probably OK although as it doesn't appear to be a True Peak limiter I'd maybe be a little more conservative on the peak level, maybe -2.5dB.
G

Now exactly what is the implication there? How can you tell if something is a true peak limter, and are they better? I've seen people on youtube use Izotope Ozone for mastering their songs, but instead of using the limiter that comes with Ozone, they used the L2 Ultramaximizer. Why do they do that? Is the Ultramaximizer a better limiter?

I purchased the entire Izotope Advanced Audio correction Suite, and I thought that I had all the plugins that I could possibly need. But after watching that video where the person used L2 with Ozone, and after reading a bunch of comments on the L2, I started looking at Waves.

They have a great bundle deal until the end of February. Would you recommend this one,
http://www.waves.com/bundles/sound-design-suite
or would you say that I have more than I need.

Thanks again,
Aveek

 

[AudioPostExpert]

HI seem to like the HF. Makes it sound a little "tinny?" My ears seem to enjoy that sound. But I suppose it's not about what I enjoy?

Right, it's not about what you personally enjoy but regardless, what you are hearing is a function of your listening equipment/environment. How do you know that you're not adding a lot of HF to compensate for a lack of HF in your system? Pretty much all audio post eventually comes back to this question, how much do you trust your listening environment? It's a tough problem that many commercial audio post facilities spend hundreds of thousands or even millions answering!

Regarding dialogue specifically, there is very little of any worth in DX above 12kHz and probably little above 8kHz or so. Boosting the HF as you have, is mainly boosting just noise/interference. A bit of a boost around 2-3kHz can often help DX but rarely is it beneficial to boost freqs higher than this.

I guess I'll take the gate off completely. I've come to the conclusion that it makes dialogue sound worse.

Gates are a bit of a blunt instrument. Very useful under certain circumstances but extremely gentle use or none at all is usually recommended for very exposed DX. A gate with a "knee" control is usually better in this regard.

Now exactly what is the implication there? How can you tell if something is a true peak limter, and are they better?

It's not so much about a TP limiter being intrinsically better, it's just a slightly different tool. A TP limiter is required for broadcast applications as specs stipulate peak limits in True Peak rather than sample peak, which is what the music industry generally uses. True Peak (TP) indicates peak values once the digital audio data is converted back to an analogue audio signal, sample peak indicates the maximum value of the digital data itself. There can be a significant difference between sample peak and TP (TP is always higher than sample peak) but on DX I wouldn't expect it to be more than half a dB.

If a limiter is a TP limiter it will specifically state that it is, there is an additional processing overhead for TP limiting. BTW, the Waves L2 is not a TP limiter. When format converting for the web, this TP vs sample peak can come into play, causing clipping distortion where none is indicated by the sample values. In your circumstances I'm not sure it would be worth specifically buying a TP limiter or using anything better than your current limiter. If you want to see the TP value, I believe RX3/4 will provide that info (along with the integrated loudness of an audio file). Keep below -2dBTP and you'll be fine, so for DX that means a sample peak of -2.5dBFS should see you consistently below -2dBTP.

G

 

[trueindie]

Thanks