IPB

Welcome Guest ( Log In | Register )

4 Pages V  < 1 2 3 4 >  
Reply to this topicStart new topic
Normalize Vs Volume increase
AndyH-ha
post Jun 30 2011, 22:25
Post #26





Group: Members
Posts: 2207
Joined: 31-August 05
Member No.: 24222



It may be true that most editors work in floating point, regardless of what your source is (this is not true of the several I use). but if you don't do everything at one go, without saving anything until everything is finished, there will still be conversions back to integer every time you write to disk, with the resulting increased quantization error. That may be ok if you have no choice but, as I wrote earlier, unless your computer resources are puny, it is easier and more reasonable to just record in floating point and never leave it until you are done - with everything.



Sound Forge is really so primitive that it cannot record in floating point format? Can it convert to floating point after recording? The only downside of this (done one time, as the first step after recording) is the time it takes.

There are people who will point out that some LP processing software only works in 16 bit integer. This is true of a few. WaveRepair is one that is (otherwise) very good at certain things but if you are not using such a program, there is no reason to restrict yourself to its limitations.

This post has been edited by db1989: Jul 1 2011, 13:23
Reason for edit: As on previous page
Go to the top of the page
+Quote Post
mixminus1
post Jun 30 2011, 22:36
Post #27





Group: Members
Posts: 688
Joined: 23-February 05
Member No.: 20097



QUOTE (tinpanalley @ Jun 30 2011, 14:16) *
Ok, I'm trying Sound Forge now. Aiming for -6db as a peak. The preamp's little light is telling me it's clipping, but Sound Forge says -7.0 has been the max.

By "preamp", do you mean the USB phono pre? If so, back off! smile.gif

You definitely don't want to clip the preamp - if you can capture at 24-bit (which it sounds like you can in Sound Forge), the levels you've been using (-12 dB peaks) are just fine.

I would be rather surprised if Sound Forge didn't work at floating point internally, even if it can't save it to disk (although that is rather strange).

When you say it will only let you do 24-bit and not float, do you mean when you open up a new project, or when you try to save/export something to WAV?


--------------------
"Not sure what the question is, but the answer is probably no."
Go to the top of the page
+Quote Post
tinpanalley
post Jun 30 2011, 22:39
Post #28





Group: Members
Posts: 124
Joined: 7-May 07
Member No.: 43262



No, just that when I choose 32-bit (IEEE float) it tells me
"An error occurred while opening an audio device. An unsupported media type was requested.
USB Preamp (USB Audio Codec) does not support 32-bit floating point input."
Go to the top of the page
+Quote Post
tropicalfish
post Jun 30 2011, 22:51
Post #29





Group: Members
Posts: 112
Joined: 10-April 10
Member No.: 79758



QUOTE (db1989 @ Jun 30 2011, 13:43) *
QUOTE (tropicalfish @ Jun 30 2011, 19:25) *
Then after that, Amplify. It'll automatically find the right amount to amplify to get your highest peak to 0dB.

If you use normalize, then all the tracks (if you record each track separately) will lose the intentional volume differences and they will appear/sound as if they're the same loudness.
As will also occur as a result of your suggestion to amplify to 0 dB, the only difference being that this option is free of the other’s idiosyncrasy whereby it treats both channels separately (as I noted in the past and as is still the case according to DVDDoug).

??

If you have multiple tracks (and you select all of them) and amplify them all at the same time, it will raise all of the tracks by the same amount until the highest peak in any one of the tracks meets 0dB.

Not sure what you mean. Normalize raises all of the tracks by different amounts so that the highest peak in each individual track meets 0dB.

I'm using Audacity 1.3.13




This post has been edited by tropicalfish: Jun 30 2011, 23:01
Go to the top of the page
+Quote Post
tinpanalley
post Jun 30 2011, 23:05
Post #30





Group: Members
Posts: 124
Joined: 7-May 07
Member No.: 43262



You see, that's what I thought. huh.gif
Go to the top of the page
+Quote Post
tropicalfish
post Jun 30 2011, 23:22
Post #31





Group: Members
Posts: 112
Joined: 10-April 10
Member No.: 79758



There are no 32-bit devices, so why use it? You might as well stick with 16 bit.
Go to the top of the page
+Quote Post
tinpanalley
post Jun 30 2011, 23:52
Post #32





Group: Members
Posts: 124
Joined: 7-May 07
Member No.: 43262



QUOTE (tropicalfish @ Jun 30 2011, 18:22) *
There are no 32-bit devices, so why use it? You might as well stick with 16 bit.


I don't know, I didn't suggest that, it's being suggested to me.
Go to the top of the page
+Quote Post
tropicalfish
post Jul 1 2011, 01:15
Post #33





Group: Members
Posts: 112
Joined: 10-April 10
Member No.: 79758



So you want to record in floating point, and then convert to 16-bit at the end?

I can do that in Audacity and Reaper (at 32-bit FP).
Go to the top of the page
+Quote Post
tinpanalley
post Jul 1 2011, 01:28
Post #34





Group: Members
Posts: 124
Joined: 7-May 07
Member No.: 43262



I don't really care how I record. I just wanna get a sense of what I asked before about getting the best sound possible.

1. What do you recommend I do with my settings when I capture vinyl, for someone like me who is all about the purity of the sound as it is reproduced by the stylus and cartridge?
Go to the top of the page
+Quote Post
DVDdoug
post Jul 1 2011, 01:53
Post #35





Group: Members
Posts: 2565
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE
1. What do you recommend I do with my settings when I capture vinyl, for someone like me who is all about the purity of the sound as it is reproduced by the stylus and cartridge?
I think that's been answered, and I think you are doing fine.

Don't your digital recordings sound identical to the vinyl when played back on the same equipment at the same volume?

This post has been edited by DVDdoug: Jul 1 2011, 01:55
Go to the top of the page
+Quote Post
tinpanalley
post Jul 1 2011, 04:08
Post #36





Group: Members
Posts: 124
Joined: 7-May 07
Member No.: 43262



All the 24 and 32 bit stuff still has me really confused. It's fine, I just need to do the research and studying myself. It's frustrating for me because as with video I have this thing where I NEED to know what is happening technically throughout the process but I don't have enough background or experience with audio to get my head around it.

There's an analog signal being turned into something digital via the preamp which is also boosting the signal. That newly amplified signal is then being sent digitally (as 1s and 0s) via a USB cable into a computer which takes the input and copies those 1s and 0s onto a harddrive. The software represents that digitized analog signal as a waveform which I can save in various containers (.wav, .flac, etc) and with various filters, settings, and processes applied. I get all of that (IF that was all correct). It's the bits I don't get.

I don't get:
- if the signal looks like it's clipping on the light of the preamp, why is it only hitting -7 on the meter in Sound Forge? Is there another variable somewhere? The Windows input level or something? Why aren't they equal?
- How does the bit amount affect the recording?
- Is the bit depth something the RECORDING HAS or something the SIGNAL COMES WITH from the turntable?

Look, I've got this great turntable and the guys at needledoctor.com helped me pick the best stylus and cartridge for my setup. I don't wanna get so-so sound. I'm the guy that will only use FLAC on his media player (a Cowon). I use gapless. I like my classical having the appropriately low levels. I own Sinatra's Capitol albums on mono because I don't like the duophonic crap that was produced in the 50s when the labels wanted everyone to buy a stereo hifi. I can't be ok with "yeah, that sounds pretty good". I wanna have the best I can and understand why it's the best that I can do with my vinyl captures. Does that make sense? Am I being too picky? sad.gif ...geez, whiney dude, huh?
Go to the top of the page
+Quote Post
AndyH-ha
post Jul 1 2011, 06:06
Post #37





Group: Members
Posts: 2207
Joined: 31-August 05
Member No.: 24222



There exist a continuously varying analogue signal from out of the analogue part of the recording chain. It is sampled by the ADC (in your case, inside the USB device). The bit depth is the precision to which it is sampled.

2 to the power of the number of bits is the number of different voltage values that can be output from the DAC (i.e. 2^16=65536 steps between minimum negative voltage and maximum positive voltage -- half of these for greater that 0, half for less than zero, minus one or two for the sign).

The ADC must “choose,” from these 65536 possible values, the one that is closest to the “actual” value of the signal voltage. The more bits it can use (i.e. 24 bits), the smaller each step, so the closer the choice will be to the real analogue voltage (as measured with infinite precision, or at least more precision than the ADC can manage).

24 bits is the limit of the hardware. However, no soundcard can actually do 24 bits. Thermal noise fills the lowest order bits due to the nature of not living in a extra cold part of the universe, so voltages very (very) near zero are lost in noise.

16,777,216 steps for 24 bit covers exactly the same range as 65,536 steps for 16 bit, so the quantization error (difference between “actual” voltage and what the ADC can record) is that much smaller (2^24/2^16). If fact it isn’t possible to hear the difference but when you do post recording processing, the same kind of errors occur in each calculation, so they add up.

There are a variety of 32 bit floating point formats but they are basically a 24 bit mantissa and a 8 bit exponent. This lets the data scale as necessary. Recording is not different between 24 bit and floating point but calculations errors are smaller and it is almost impossible to clip (unless you forget about >0dB values when you convert to integer).

Any format, 16 bit integer, 24 bit integer, floating point (32 and 64 bit varieties) works fine for LP recordings, which have so much intrinsic noise that they obscure the differences. Still, why not do as well as you can?
Go to the top of the page
+Quote Post
db1989
post Jul 1 2011, 13:25
Post #38





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



QUOTE (tropicalfish @ Jun 30 2011, 22:51) *
QUOTE (db1989 @ Jun 30 2011, 13:43) *
QUOTE (tropicalfish @ Jun 30 2011, 19:25) *
Then after that, Amplify. It'll automatically find the right amount to amplify to get your highest peak to 0dB.

If you use normalize, then all the tracks (if you record each track separately) will lose the intentional volume differences and they will appear/sound as if they're the same loudness.
As will also occur as a result of your suggestion to amplify to 0 dB, the only difference being that this option is free of the other’s idiosyncrasy whereby it treats both channels separately (as I noted in the past and as is still the case according to DVDDoug).

??

If you have multiple tracks (and you select all of them) and amplify them all at the same time, it will raise all of the tracks by the same amount until the highest peak in any one of the tracks meets 0dB.

Not sure what you mean. Normalize raises all of the tracks by different amounts so that the highest peak in each individual track meets 0dB.

[pic]
Sorry; I hadn’t considered you were referring to tracks that were vertically stacked (i.e. not horizontally).
Go to the top of the page
+Quote Post
2Bdecided
post Jul 1 2011, 18:19
Post #39


ReplayGain developer


Group: Developer
Posts: 5089
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



QUOTE (AndyH-ha @ Jun 30 2011, 20:51) *
I have seen a DC offset in a few LP recordings that was on the LP itself, part of what I got off the disk
Come on Andy - think that one through properly and have a good laugh at yourself! wink.gif Hint: how would a DC offset be stored in a record groove? How would a cartridge detect it?

Cheers,
David.
Go to the top of the page
+Quote Post
2Bdecided
post Jul 1 2011, 18:25
Post #40


ReplayGain developer


Group: Developer
Posts: 5089
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



QUOTE (tinpanalley @ Jul 1 2011, 04:08) *
I own Sinatra's Capitol albums on mono because I don't like the duophonic crap that was produced in the 50s when the labels wanted everyone to buy a stereo hifi.
Aren't the official Capitol CD releases already the properly mono?

QUOTE
I can't be ok with "yeah, that sounds pretty good". I wanna have the best I can and understand why it's the best that I can do with my vinyl captures. Does that make sense? Am I being too picky? sad.gif ...geez, whiney dude, huh?
What matters is whether you can hear a difference. You already have everything you need in order to find out. Go on, do some listening wink.gif


Peaking 12dB down while recording at 16-bit is just leaving the territory where people here will tell you "there isn't a cat in hell's chance of hearing a difference", and heading into the territory where "you probably can't hear a difference but just maybe sometimes theoretically you might".

Recording at 24-bit won't do any harm, and will almost inevitably cause a placebo induced sensation that the sound quality has improved, if you believe that it should.


Two different cartridges are going to sound 1000x more different than the difference caused by any of this stuff.

Cheers,
David.
Go to the top of the page
+Quote Post
tinpanalley
post Jul 1 2011, 19:51
Post #41





Group: Members
Posts: 124
Joined: 7-May 07
Member No.: 43262



Thanks for the wonderful feedback. Just wanted to get to this first...

QUOTE (2Bdecided @ Jul 1 2011, 13:25) *
Aren't the official Capitol CD releases already the properly mono?


No CD has ever been produced of the Capitol Mono masters... sadly. Only the odd track here and there. The only thing out is the horrible remaster from the late 80s which attempted to "fix" the Duophonic masters and ended up adding bass and excessive reverb to his voice and killing some of the background instruments in the process. A new remaster cleaned those up but they're still just the duophonic versions cleaned up. The UK Capitol CDs are better than any of the Us ones though which is sad. A bit like Vera Lynn's US CDs sounding better than the UK ones. That effort should be made here. Anyway... different thread, I suppose.
Go to the top of the page
+Quote Post
AndyH-ha
post Jul 1 2011, 20:36
Post #42





Group: Members
Posts: 2207
Joined: 31-August 05
Member No.: 24222



QUOTE
Come on Andy - think that one through properly and have a good laugh at yourself! wink.gif Hint: how would a DC offset be stored in a record groove? How would a cartridge detect it?

The answer is in: how is a DC offset stored in the recording on one’s hard drive?
Go to the top of the page
+Quote Post
DVDdoug
post Jul 1 2011, 21:46
Post #43





Group: Members
Posts: 2565
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE
It's the bits I don't get.
OK, I'll try to simplify it a 'bit'... wink.gif

Digital audio is a series of numbers (or "values".). Each number represent the amplitude (height) of the wave at any a single point in time. You essentially "connect the dots" to create the wave-shape. If you zoom-in with Audacity, you will see the wave represented as a series of stair steps*. But don't worry.... When you play the audio file, the steps are smoothed-out by a low-pass filter following the digital-to-analog converter. Some audio editors will show a smooth-filtered waveform... I'm not sure about Sound Forge.)


A "bit" is a binary "digit" (base-2).

With 8 bits, the highest number we can represent (or can count to) is 1111 1111, which converts to 255 decimal.

With 16 bits, we can count to 1111 1111 1111 1111 (65,535 in decimal).
(We usually write the bits in groups of 4 because it's easier to read and its easier for programmers to convert to hexadecimal = base 16.) ohmy.gif

The actual format and conversion to decimal varies because one bit can be used for a the +/- sign. The values in a 16-bit WAV file range from -32,768 to +32,767. And with floating-point, some bits are used for the mantissa and some for the exponent, as Andy explained.

A 16 bit file does not go "louder" than an 8-bit file, because all of the formats are scaled.** The 16-bit file, has more, smaller, steps = More resolution. Floating-point is scaled differently and you can have a "taller" wave. But when you play-back a floating-point file, it still has to go through an integer digital-to-analog converter, so it won't go "louder" in the real world either.


* Of course, digital audio is quantized (digitized) in both amplitude (vertical) and time (horizontal). The bit depth represents the amplitude resolution, and the sample rate kHz, represents the time resolution.

** In mathematics, this "scaling" to adjust of the formats to have the same volume/amplitude would be called "normalizing", but in audio processing normalize has a slightly different meaning.
Go to the top of the page
+Quote Post
DVDdoug
post Jul 1 2011, 22:12
Post #44





Group: Members
Posts: 2565
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE (AndyH-ha @ Jul 1 2011, 12:36) *
QUOTE
Come on Andy - think that one through properly and have a good laugh at yourself! wink.gif Hint: how would a DC offset be stored in a record groove? How would a cartridge detect it?

The answer is in: how is a DC offset stored in the recording on one's hard drive?
We're getting a bit off topic here, but DC = zero Hz which cannot be stored on a record, or reproduced by a phono cartridge which requires movement to generate a voltage. However, you can easily store "zero Hz", or a constant offset value in a WAV file.

But, I believe you can have an average-offset, or asymmetry on a record where the positive half-cycles are larger than the negative half-cycles, or vice-versa.


BTW - GoldWave's offset correction algorithm can actually introduce a true DC offset component if you feed-in an asymmetric waveform. I assume it works by adding/subtracting a constant value to/from each sample, so that the new-average of all samples is zero. This algorithm can create a "click" if there s silence at the beginning/end of a file. Since I discovered this, I fix offset with a regular-old high-pass filter. (I believe the fix-offset algorithm has been corrected in the new GoldWave beta release.)

This post has been edited by DVDdoug: Jul 1 2011, 22:27
Go to the top of the page
+Quote Post
2Bdecided
post Jul 4 2011, 12:34
Post #45


ReplayGain developer


Group: Developer
Posts: 5089
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



QUOTE (DVDdoug @ Jul 1 2011, 22:12) *
We're getting a bit off topic here, but DC = zero Hz which cannot be stored on a record, or reproduced by a phono cartridge which requires movement to generate a voltage.
And even if it didn't, what would the DC offset be relative to, given that the record groove itself is moving inwards the centre of the record, and the arm+cartridge have got to follow it!

Cheers,
David.
Go to the top of the page
+Quote Post
AndyH-ha
post Jul 4 2011, 20:32
Post #46





Group: Members
Posts: 2207
Joined: 31-August 05
Member No.: 24222



My experience is based on CoolEdit Statistics reporting a few percent DC offset on a few LPs over the years. These were isolated cases. I normally record three or four albums at a time. There was no DC offset on most of 700 to 800 disks. It is unlikely that something changed electrically in my system for one of those three or four LP recorded on some odd day, then immediately went right again when playing the next album, preventing further DC offset in the next recording.

I know it is not the electrical condition of the LP. What you seem to be saying is that if I used recording equipment with an offset, so the recording I end up with on hard drive measures with a DC offset because of that recording chain problem, just cutting the recording to LP would eliminate that offset. If this is true, I don't understand the mechanism.
Go to the top of the page
+Quote Post
2Bdecided
post Jul 4 2011, 23:34
Post #47


ReplayGain developer


Group: Developer
Posts: 5089
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



I think you may think I might be being clever than I am (?!) ...

QUOTE (AndyH-ha @ Jul 4 2011, 20:32) *
If this is true, I don't understand the mechanism.
If you recorded something with a 40000% DC offset, it would move the entire record groove about 1 inch to the left.

How would you know this when you played it back?

Cheers,
David.

This post has been edited by 2Bdecided: Jul 4 2011, 23:35
Go to the top of the page
+Quote Post
AndyH-ha
post Jul 5 2011, 08:38
Post #48





Group: Members
Posts: 2207
Joined: 31-August 05
Member No.: 24222



Perhaps I am just too ignorant of what is involved. The waveform is displaced by some amount from zero. This is normally caused by a DC voltage on the analogue signal line. It exist in the analogue signal. When converting to digital it is retained as a numerical component of each sample value. The file can be copied elsewhere, within the computer or to another, and the offset does not go away.

Since there is a way to detect and remove it in an editor, there has to be a way to add it where it did not originally exist. In a digital recording it is just another part of what makes up the sample, in a sense like 60Hz hum. It doesn't matter whether it was actually caused by DC where there should have been none, or by something else.

Also, I don't seen any reason there could not be a DC voltage on the signal line going into the cutting amplifier, adding an offset to the recording being fed to it. However, I take it by your extreme example that cutting the signal to disk would effectively remove the offset. I can see there is no reference available on the LP but if this means the signal becomes what it would have been had there been no offset to begin with, then either
What I 've found on a few LP recordings is actually something else that just happens to look the same or
my system somehow, very infrequently, glitches just for the duration of one LP.

I never gave it any thought before. The few times it was there I just removed it. Since most of my recordings show no such problem, I did not consider the odd, very infrequent case a sign that my equipment was acting up, any more than I thought the clicks and pops were the fault of my system rather than the condition of the LP.
Go to the top of the page
+Quote Post
2Bdecided
post Jul 5 2011, 14:08
Post #49


ReplayGain developer


Group: Developer
Posts: 5089
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



Vinyl is an intrinsically AC-coupled format. It has no concept of DC.

Any DC selectively reported by your software is because your software's idea of 0 DC did not match your equipment's - maybe due to different timescales / time constants for removing DC (e.g. most software just sums all values in the whole file), maybe because there was some assymetry in response after DC was removed (e.g. distortion) which means different signals will appear to have different amounts of DC.

Cheers,
David.
Go to the top of the page
+Quote Post
pdq
post Jul 5 2011, 16:14
Post #50





Group: Members
Posts: 3391
Joined: 1-September 05
From: SE Pennsylvania
Member No.: 24233



Think of it this way. If there is a DC offset in the signal when the master for the LP is created, all of the grooves will be displaced a few microns in the same direction. When you drop the needle into the first groove, you have just applied a DC correction by placing it in the center of the offset groove.
Go to the top of the page
+Quote Post

4 Pages V  < 1 2 3 4 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 22nd August 2014 - 03:05