Routing arrangement for A/B comparisons?

Counterparts · Post by **Counterparts** » Thu Apr 15, 2004 7:18 am

I was wondering how people here route things so that one can easily compare the effect of (for example) the Optimaster plugin compared to the unprocessed signal, such that the relative volume levels are the same?

I have been thinking along the lines of taking both the input and output of the Optimaster (or whatever) into a little mixer, a stereo channel for each, tweaking the volume levels in that mixer, then simply muting / unmuting the relevant channel to make the comparison.

What about if you have a chain of devices, and wanted to compare the processed to unprocessed signals at any point in the chain (again, with equalised volume levels)?

TIA,

Royston

Immanuel · Post by **Immanuel** » Thu Apr 15, 2004 8:50 am

Controll room (in your mixer folder)
j9k's VU meter (not the peak meter - the VU meter).

Counterparts · Post by **Counterparts** » Thu Apr 15, 2004 9:18 am

Hi Immanuel

Immanuel wrote:
Controll room (in your mixer folder)

Ah - gotcha! That looks like what I was after (I had been considering making a modular equivalent with j9k's switch devices). There's still a pile of devices in SFP that I haven't got round to checking out properly yet...

j9k's VU meter (not the peak meter - the VU meter).

The 'VU100' device, right?

I should have been more specific; I am talking about making a comparison with one's ears, not with a meter.

Thanks,

Royston

Immanuel · Post by **Immanuel** » Thu Apr 15, 2004 9:22 am

I am talking about this one
http://www.planetz.com/forums/viewtopic ... orum=16&10
I think it does the job well enough. You can make small tweaks afterwards, but I brings you close.

I find it hard to make it equal by ear. As soon as you compress, and compare it to non-compressed stuff, something will be louder and other things will be softer.

I recomend you give it a go (it is not too much of a hassle) and write back your comments here afterwards.

Counterparts · Post by **Counterparts** » Thu Apr 15, 2004 9:40 am

I don't quite agree...OK, a compressor raises the effective listening volume, but if you bring the volume of the dry signal up to match the compressed signal, then you can more accurately compare how the compressor has changed the signal dynamics & the
tonal quality / frequency content of the signal - that's the reason I want to make A/B comparisons.

I can't see how you can make that sort of comparison/judgement with a meter

Hmmm...on the subject of metering there is something I'm a bit confused about:

In Vinco, you can switch between a peak-style 'LED' meter and a VU meter.

What confuses me is that if I set the compressor up with the peak meter (nothing distorting, input, compression & output levels how I want them) and then switch to using the VU meter, then (input or ouput) the needle's just permanently wedged up against the right hand side of the dial. What's the point of that?

Royston

Immanuel · Post by **Immanuel** » Thu Apr 15, 2004 5:48 pm

What I mean is, that I also do level compensation, when comparing compressed vs. noncompressed signal. I just find it hard to actually make a perfect levelcompensation - because it is 2 different signals with different sonic qualities. That is why I like to have a VU meter to help me - NOT a peak meter.

Post by **hubird** » Thu Apr 15, 2004 7:24 pm

On 2004-04-15 10:40, Counterparts wrote:
What confuses me is that if I set the compressor up with the peak meter (nothing distorting, input, compression & output levels how I want them) and then switch to using the VU meter, then (input or ouput) the needle's just permanently wedged up against the right hand side of the dial. What's the point of that?

Royston

I was asking myself exact the same question this evening, Cntrprts!
And not alone this evening!
Just not using the Vinco but the PSP Vintage Warmer.
After compressing I think the resulting sound is good, but the Vu meter says it's war?!
Anyone knows more about this?

BTW, me too I make the top volume level equal for both dry and compressed processing, if comparing.
It's the only way to make hearable what the compression does to the processed music.

You need an exact peak level meter, in decimals.
I therefor use the output peak meter of the program, like Waveburner, Cubase or Peak offers (on mac).

I accept that the compressed part sounds louder than the dry part, after all that's part of the game, that's what compressing IS all about (in the first place).

It seems useless to me to equal the subjective volumes ('ear pressure').
At best if you wanne check just the 'colouring' of the compressing process, but it's getting academic a bit then.
cheerz

_________________
Let There Be Music!

[ This Message was edited by: hubird on 2004-04-15 20:27 ]

Immanuel · Post by **Immanuel** » Thu Apr 15, 2004 8:04 pm

Not realy.
Because as a listener, I will have a prefered listening volume. If you have compressed your music more, I will just turn it down. I don't dial my amp in for numbers and just keep it there. I adjust to what I feel is comfortable. Therefor a/b-ing at the same percieved volume would be way above just achademic - if you target group is like me in that aspect ...

Counterparts · Post by **Counterparts** » Fri Apr 16, 2004 3:36 am

Immanuel wrote:

What I mean is, that I also do level compensation, when comparing compressed vs. noncompressed signal. I just find it hard to actually make a perfect levelcompensation - because it is 2 different signals with different sonic qualities. That is why I like to have a VU meter to help me - NOT a peak meter.

Ah, gotcha now - thanks for the clarification.

Royston

Counterparts · Post by **Counterparts** » Fri Apr 16, 2004 3:42 am

hubird wrote:
...but the Vu meter says it's war?!

Scarey isn't it!

I keep thinking somethings going to melt...

Royston

wolf · Post by **wolf** » Fri Apr 16, 2004 4:06 am

On 2004-04-15 20:24, hubird wrote:
After compressing I think the resulting sound is good, but the Vu meter says it's war?!
Anyone knows more about this?

a peak meter reacts faster, that's why you see peaks only occasionally, if at all. A VU-Meter (like the one from BBC) has a far longer integration time, which leads in return to a slower reacting needle.

I accept that the compressed part sounds louder than the dry part, after all that's part of the game, that's what compressing IS all about (in the first place).

In mastering stage yes. I assume you know, why everybody wants to make his recording lowder ?
In the mixing or even creating stage a compressor can do much more

a compressor makes the signal quiter in the first place due to limiting the signal dynamic.

It seems useless to me to equal the subjective volumes ('ear pressure').

better called loudness

At best if you wanne check just the 'colouring' of the compressing process, but it's getting academic a bit then.

yep, too, but in no way academic, in fact it is very important to know, how a compressor influences the sound (not only coloring, it also changes the way e.g. a kick kicks and therefor the groove) in mastering stage, as you don't wanna change the overall perception, just make it louder. .. because louder tracks are perceived as better quality, even if it's not the case.
That's beside others, why you shouldn't mix at too loud volumes, because you always think it sounds good

best
wolfgang

Post by **hubird** » Fri Apr 16, 2004 7:14 am

thanks Wolf for your comments.
I know there is a -14dB 'gap' between Peak and VU.
It's scary to see the peak meter staying totally at the right max, while the VU meter stays below zero.
I guess I need more studying, the PSP manual is good but not basic enough for a DIY type a guy like me

Post by **garyb** » Fri Apr 16, 2004 3:16 pm

no VU probs here..

war?
you really like it when there's no more dynamics,huh?

compressors can't make the maximum volume any louder,but they do make the PERCEIVED overall volume louder by making the louder sounds quieter and the quiet sounds louder(less dynamics).if gain is added to match the original peak level obviously the sound at that peak level will be more intense,louder,more dense.

put that less dymanic track among others with normal dynamics and it will seem quite loud in comparison even at lower peaks.

i don't want to come off like the great know it all,much respect to the pros here!....whatever works to make your stuff sound right is correct. i'm just saying something for the sake saying it. for those who don't know.....

i don't know but i might just use the bypass switch.if you are adjusting gain to match perceived loundness, the bypass works great. you could also use a couple of mixer channels and mute/unmute....

wolf · Post by **wolf** » Fri Apr 16, 2004 3:32 pm

@hubird:
you're welcome

btw, the vu meter is average reading, while the peak meter .. well, reads peaks.
showing the average of a signal is more like the human ear is receiving sound, while we aren't able to hear peaks like the peak meter is showing us.
In return you should use a peak meter to check clipping and a vu meter (better is a rms meter, which is really close to what loudness the human ear receives) to check, what the loudness dependant on released recordings is.

puh, nuff said .. now I've to prepare me for the live gig this evening

have fun!

wolfgang

_________________
worldless productions
<a href="http://www.worldless.com">worldless.com</a>

[ This Message was edited by: wolf on 2004-04-16 16:35 ]

kensuguro · Post by **kensuguro** » Fri Apr 16, 2004 3:34 pm

I also found out (probably posted this somewhere before) that compressing decreases the distance the speaker driver needs to move (less dynamics) and there fore allows cheap ass speakers and hi-fi speakers to achieve similar results. So, it also helps in making your stuff sound similar accross different playback units. I think that was the conclusion I reached. Dunno any books or numbers that back this theory up tho.

By the way, I route things into the dyna mixer and mute/unmute to make a/b comparison. I don't turn up the dry signal tho, cuz most of the time, when I do masters, my aim is to change the sound characteristic. If I have a multi stage processing chain, I connect all the outputs to different channels on dynamixer, and do the mute/unmute or solo to make comparisons. I've been doing this for quite a while now.

[ This Message was edited by: kensuguro on 2004-04-16 16:37 ]

Immanuel · Post by **Immanuel** » Sat Apr 17, 2004 2:38 pm

On 2004-04-16 08:14, hubird wrote:
I know there is a -14dB 'gap' between Peak and VU.

Only in some cases is this true. The VU meter reads the average level - as wolf said it. If your average level is 14dB below the the peak level, then it is true. But you may very well have an average level, wich is either more or less than 14dB from peak level. The diffence between the average level and the peak level is called the crest factor.

I do not know, where that -14dB truth comes from. I know, that Bob Katz has made a pdf about mastering, wich has been read be a LOT of people all over the world. He wrote, that he usually trims his vu-meters to -14 (as far as I remember it - it was last year, that I read it). This may on the internet has changed to vu-level being 14dB below peak level.

I think your music would probably mostly have a crest factor of less than 14dB (it has been some weeks

).

Immanuel · Post by **Immanuel** » Sat Apr 17, 2004 2:53 pm

On 2004-04-16 16:34, kensuguro wrote:
I also found out (probably posted this somewhere before) that compressing decreases the distance the speaker driver needs to move (less dynamics) and there fore allows cheap ass speakers and hi-fi speakers to achieve similar results. So, it also helps in making your stuff sound similar accross different playback units. I think that was the conclusion I reached. Dunno any books or numbers that back this theory up tho.

It is very simple - therefor it depends on the scenario. A speaker needs to move more to produce higher output. If you adjust the volume on the amplifier to the same percieved loudness, then the peaks will be lower volume, if the track has been compressed. This will be easyer to reproduce for the speaker + amplifer.

side note: if you make digital clipping, you can get a special kind of distortion, becourse you cut the peaks. This has the side effect, that the waveform will move quicker from minimum to maximum amplitude. This will be mroe demaning on your amp + speakers. The weaker the power amplifier (compared to the playback volume) + the longer and thinner cables to the bass driver (to the entire speaker, if you don't bi-wire), the harder will it be for the amplifier to comtroll the bass driver, and therefor the bassdriver may try and move longer than asked (because the amplifier can not controll it to stay in the heavily DC-offset position wich in effect is created by digital hard clipping). This puts the speaker at risk of being damaged. Tweaters are, however (as far, as I know) the most often damaged speakers due to digital clipping. Ofcoarse speaker characteristics play in too.

Post by **hubird** » Sat Apr 17, 2004 5:34 pm

@ immanuel:
thanks for your comments

Concerning the -14 dB issue, here some words about the PSP Vintage Meters' meters, grabbed from the manual:

PSP VintageMeters’ meters are designed to work similar to real VU and PPM meters and as well as over indicators.
To ensure that they can be used as a useful tool in every situation we supplied them with a set of parameters that are aimed to adjust the meters' behavior to every practical situation.
....
Standard VU meters are specified to work with 300ms integration time - every other integration time setting gives a response that is not compatible with standard VU time response but allows you to adjust it to your particular needs.
As the VU meter show a kind of average level, it has to be calibrated to be useful for real applications.

In normal situation the VU meter shows -14dB value relative to peak value.
That is why mastering and post production engineers decided to use it as reference level for music.

Nowadays hot level practice are the cause for average levels to be much closer to peak value - that is why we decided to give a user such a wide reference level adjustment range.

PPM meters are Pseudo Peak Meters. They show the level value very close to digital peak values.
Typically they have 10ms attack and about 1000-2000ms release times. In many cases they are more practical then Digital peak meters.
You can achieve perfect digital peak metering by setting attack to 0ms.

---------------------

In practice I know how to handle with compression, and I benefit from a discussion on planetz about 'multi-stage' compression, which in fact is playing around with different 'knie' settings, if I may say it this way

I was just amazed by the huge difference between the Vu and Peak meter's behaviour behaviour, being the one 100% normal, the other totally mad

Anyway, compressing still is the most complicated tech/musical process I know, it keeps me going on thinking and retrying.

Now I must say sorry to Counterpart for carrying away the subject!

_________________
Let There Be Music!

[ This Message was edited by: hubird on 2004-04-17 18:34 ]

Immanuel · Post by **Immanuel** » Sun Apr 18, 2004 1:28 pm

On 2004-04-17 18:34, hubird wrote:
Standard VU meters are specified to work with 300ms integration time ...

In normal situation the VU meter shows -14dB value relative to peak value.

Nowadays hot level practice are the cause for average levels to be much closer to peak value.

Now here are 3 statements from the manual. Statement 1 and 3 are ok, but statement 2 is not compaible with statement 1 and 3. Lets say you have a snare drum track at beat 120. The snare plays quarternotes with -2dB peak level. In another track you have a very think pad playing -2dB peak. You will just not get -16dB on both tracks with a VU-meter. If it was so normal to have VU to be -14 compared to peak, then you wouldn't realy need VU anymore for working in digital. The -14dB is no rule of thumb. Try running some songs from different decades and different genres and play them thru a peak meter and a VU-meter. You will just not get that -14dB all that often. Run a sine thru the VU-meter, and you will get around -3 (I might remember wrong here). Take a song and normalise it to 0dB. Play it thru the VU-meter. Now put it thru a brick wall limmiter set to -2dB and normalise it to 0dB again. You VU-meter will now measure around 2dB more.

So to repeat myself the -14dB rule does not realy exist. Writen in a manual or not. I do not believe it.

Post by **hubird** » Sun Apr 18, 2004 1:53 pm

ok, thanks