ampdb

[COCHEDELAFERTE]

Hi all,

Learning the attached file "DeepNote.csd" i found in Dr tutos, i try to enhance the output volume via the ampdb opcode, which is set in the example file at 68 (db). If i make that (to 72 or more), i get in the console a lot of "out of limits samples". In the fact, this message had no real bad effect, the file is played as well, with the output sound enhanced as attempted. But in all cases i'm interested to understand why this message, and how to modify the tinstrument to have a perfect console output with any ampdb parameter, not only 68 db or less. If someone can take a quick look and explain me that...

DeepNote.csd


Cheers,

Sergio


 

[Adam Puckett-2]

Out-of-range samples means you have too high a volume when rendering your piece.

On 1/31/12, COCHEDELAFERTE <[hidden email]> wrote:

> Hi all,
>
> Learning the attached file "DeepNote.csd" i found in Dr tutos, i try to
> enhance the output volume via the ampdb opcode, which is set in the example
> file at 68 (db). If i make that (to 72 or more), i get in the console a lot
> of "out of limits samples". In the fact, this message had no real bad
> effect, the file is played as well, with the output sound enhanced as
> attempted. But in all cases i'm interested to understand why this message,
> and how to modify the tinstrument to have a perfect console output with any
> ampdb parameter, not only 68 db or less. If someone can take a quick look
> and explain me that...
>
> http://csound.1045644.n5.nabble.com/file/n5445967/DeepNote.csd DeepNote.csd
>
>
> Cheers,
>
> Sergio
>
>
>
>
> --
> View this message in context:
> http://csound.1045644.n5.nabble.com/ampdb-tp5445967p5445967.html
> Sent from the Csound - General mailing list archive at Nabble.com.
>
>
> Send bugs reports to the Sourceforge bug tracker
>             https://sourceforge.net/tracker/?group_id=81968&atid=564599
> Discussions of bugs and features can be posted here
> To unsubscribe, send email [hidden email] with body "unsubscribe
> csound"
>
>

Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email [hidden email] with body "unsubscribe csound"

[Richard Dobson]

On 31/01/2012 22:57, COCHEDELAFERTE wrote:

> Hi all,
>
> Learning the attached file "DeepNote.csd" i found in Dr tutos, i try to
> enhance the output volume via the ampdb opcode, which is set in the example
> file at 68 (db). If i make that (to 72 or more), i get in the console a lot
> of "out of limits samples". In the fact, this message had no real bad
> effect, the file is played as well, with the output sound enhanced as
> attempted. But in all cases i'm interested to understand why this message,
> and how to modify the tinstrument to have a perfect console output with any
> ampdb parameter, not only 68 db or less. If someone can take a quick look
> and explain me that...
>
> http://csound.1045644.n5.nabble.com/file/n5445967/DeepNote.csd DeepNote.csd
>
>

It is adding 30 tones, which mathematically (not allowing for possible
cancellations between components) gives a level of approx +30dBFS, or
"way over" digital peak. So, it has to be scaled downwards accordingly
to bring it within range, which the code duly does, in a particular way.
It is living a bit dangerously, indeed; for  the usual 16bit range
maximum is in ampdb terms 96dB, so the target should really be 66 (or
even 60) not 68. The output is in practice reported as  some -1.7dB
below digital peak, which is really borderline but just about OK. Thus,
rasising the value given to ampdb even just a little will take you into
overload. The output is really already as loud as it can be. If you want
more "impact" etc, you have to resort to the usual artifice of
compression, further synthetic enhancement, et al.

These days the recommended approach is to use ampdbfs with a value such
as -30, as dBs only really make sense relative to a maximum, which by
convention is 0dB. You can also to set 0dbfs = 1, to obviate any wrong
impression that the raw amplitude numbers relate in any literal way to
the output sample format.

Richard Dobson


Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email [hidden email] with body "unsubscribe csound"

[Mark Van Peteghem-2]

I suspect that it is out of limit for only some small periods. In that
case it may still sound ok, but most people prefer not to do that. The
listener can always turn up the volume of his computer or hifi
installation :-)

Mark

Op 31/01/2012 23:57, COCHEDELAFERTE schreef:

> Hi all,
>
> Learning the attached file "DeepNote.csd" i found in Dr tutos, i try to
> enhance the output volume via the ampdb opcode, which is set in the example
> file at 68 (db). If i make that (to 72 or more), i get in the console a lot
> of "out of limits samples". In the fact, this message had no real bad
> effect, the file is played as well, with the output sound enhanced as
> attempted. But in all cases i'm interested to understand why this message,
> and how to modify the tinstrument to have a perfect console output with any
> ampdb parameter, not only 68 db or less. If someone can take a quick look
> and explain me that...
>
> http://csound.1045644.n5.nabble.com/file/n5445967/DeepNote.csd DeepNote.csd
>
>
> Cheers,
>
> Sergio
>
>
>

Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email [hidden email] with body "unsubscribe csound"

[COCHEDELAFERTE]

Hi all tree,


Thank's greatly for your replies. Now i'll try several things as possible with Odbfs (last part of Richard's explanations) or also, just to see what happens, completely without amdb, i mean with usual amplitude notation instead. In the fact i found that the sound isn't particularly strong in this piece, even with 30 tons, comparatively anothers usual examples or things i yet composed.


Cheers,

Sergio

[peiman]

It is adding 30 tones, which mathematically (not allowing for possible cancellations between components) gives a level of approx +30dBFS, or "way over" digital peak. So, it has to be scaled downwards accordingly to bring it within range, which the code duly does, in a particular way. It is living a bit dangerously, indeed; for  the usual 16bit range maximum is in ampdb terms 96dB, so the target should really be 66 (or even 60) not 68. The output is in practice reported as  some -1.7dB below digital peak, which is really borderline but just about OK. Thus, rasising the value given to ampdb even just a little will take you into overload. The output is really already as loud as it can be. If you want more "impact" etc, you have to resort to the usual artifice of compression, further synthetic enhancement, et al.

If I may also add that the perception of loudness depends on frequency and spectral make-up as well as amplitude. If your mix covers a wide range of spectral frequencies then it will sound more full in terms of volume (think in terms of orchestration). This is because different frequencies stimulate different regions on the basilar membrane in the inner ear, while once a region on the membrane is excited it becomes temporarily desensitised. So mixing two tones of the same pitch will not sound as loud as mixing two different pitches (assuming that the tones have the same amplitude). Also we are a lot more sensitive to higher frequencies than we are to lower frequencies.

Best,

Peiman         

[COCHEDELAFERTE]

Hi again,


1) Thank's, Peiman, for the acoustical explanations you added this day.

2) I tried to use the common amplitude declaration (just a p-field) instead the ampdb opcode : same problem if i enhance too, i'm becoming again out of limits. The question is really in the 30 tones*.

3) Richard, if you read me again, can you please give a bit more details (or just a link) about what you mean with "the usual artifice of compression, further synthetic enhancement", maybe that can help me other hand.



* Why the 3à tones ? It's related here : http://csound.1045644.n5.nabble.com/Good-vibration-td5108275.html


Cheers,

Sergio