I have read all the tutuorials on both Amplify and Normalize. I have also searched the Forum on this topic and read everything that I could find. I do not have any DC Offset in my recordings and my system seems to be reasonably balanced. I have therefore decided that I should use Amplify instead of Normalize. But I notice that the default settings in Audacity are different for the two methods. Normalize defaults to -1.0 dB. Amplify defaults to 0.0. Is there a reason for this? When I use Amplify which setting do you recommend that I use?
Charlie Nash
I usually use Normalize as it retains my preferred setting (Amplify always defaults to 0dB in Audacity 1.3.12 but I usually prefer to keep a bit of headroom before exporting, so generally Normalize to around -0.1 dB). The case where Amplify should be used rather than Normalize is if the peak levels in left and right channels are significantly (and deliberately) different.
Normalize will always amplify the left and right channels independently so that both channels have the same peak level. Often this is fine, but occasionally one channel will have higher peaks than the other channel that you want to retain (keep the original left/right balance) in such cases you need to use the Amplify effect as this amplifies Left and Right by the same amount and so retains the original Left/Right balance. You can set the target level of the amplify effect to anything you like - as with Normalize I usually adjust it to around -0.1 dB before exporting the final audio file, but this needs to be set each time the effect is used as it reverts back to 0dB in Audacity 1.3.12 (some lower level in Audacity 1.2.6)
one pushes both channels to max so stereo may become unbalanced
one pushes both until the highest one hits its max but both by the same amount so if one is lower to start it stays lower
its just numbers - defaults the coder liked i suspect
i always leave at least 1dB under the max with either
some folks push to 1/2 or closer - maybe even touching 0.0
you really cant hear the difference
so why risk problems that could occur
with some playback devices
Snip from my reply to your workflow/filter posting
<<<I use the Amplify effect and not the Normalize effect to bring the level up to -1.0dB (Amplify works equally on both channels, Normalize works on each channel independently and thus the stereo balance can be impaired)>>>
WC
or corrected.
Most modern commercial recordings have been mastered with the maximum peak level close to or at 0dB. If the transfer of a tape or record into digital has come through with one channel peaking higher than the other, it may be due to an unintentional left/right imbalance in the playback/Audacity recording. For example, if recording through a pre-amp with independent channel controls, they may not have been set exactly the same, or for a cassette there may be a small alignment error of the tracks on tape with the play heads (if this latter issue occurs regularly the tape head alignment should be corrected). In such cases, Normalize can be useful to bring left and right back into balance.
It would be really nice to have an option to select “Link stereo channels” so that the user can choose which they want.
It would be especially useful for using in batch mode (chains).
There is already a feature request for this. If you agree, please vote for it.
An option to calculating the track gains together to avoid changing the stereo balance: When multiple tracks are selected, only the highest peak level of ALL tracks should be considered to adjust each track by the same amount so that the volume of each track relative to each other will stay the same. (1 votes)
(if you do not have an account for the wiki, you can voice your vote here and one of the forum moderators will transfer it to the wiki).
WC,
I’m a little confused. While your recommendation sounds correct, is this not what normalize already does?
Also, I find it interesting that you choose to use “Amplify” whereas Steve chooses to use “Normalize”. I gather, however, from reading your past Postings that you do a lot of Vinyl transfer, whereas Steve states that he does very little Vinyl transfer. Would this be the primary reason that the two of you opt for different methods?
Charlie
Since WC is not on-line at the moment but we are…
Having done hundreds of vinyl transfers, waxcylinder has his set-up precisely tuned for producing high quality recordings from vinyl. He knows that the recording that he gets will be very close to the original sound of the record and that the left/right balance is correct.
If say the left channel of his recording has a higher peak than anything on the right channel, then that is because that is the way that it is on his original vinyl.
In such a case, if he used the “Normalize” effect, the left/right balance would be changed because both the left and right channels would be adjusted independently to create a maximum peak at the set level in both channels. That is not what he wants - he wants to amplify both channels equally so that the left/right balance is unchanged. To achieve equal left and right channel amplification, the Amplify effect must be used.
I will generally use Normalise for the reasons described earlier, though I’ll use Amplify if I specifically need to amplify both channels by the same amount.
Both approaches are perfectly valid which is why I would urge people to vote for having the option to choose “Link stereo Channels”.
WC is on the line - but he was in the middle of typing a long reply when the forum crashed - and the msg got lost
In the meantime Steve has said most of what I was going to say.
I am pretty confident that my equipment is well balanced. As Steve says, think about a recording where the left channel say has a very loud noise that is not present on the right channel. Normalize will look at each channel independently and thus will probably over-amplify the right channel - whereas Amplify will amplify them both the same, retaining the original balance.
Unless you are unlucky, I would expect the kit that you have that you listed for us to be pretty well balanced provided you set it up right. You may however find the odd original recording which is not properly balanced - and you may want to adjust that, consider Normalize in that case.
WC
Steve,
I went out to the Wiki a few minutes ago and read the recommendation regarding this. It sounds like the whole situation might be clarified somewhat if Normalize and Amplify were transformed into only one instead of two choices, but with the option that you and WC are recommending. Seems like everyone would then have what they needed (channels either tied together or calculated separately) without the confusion of Amplify vs Normalize.
As for my situation, ie, transferring Vinyl to CD, it seems like WC’s vast experience with this dictates that I should follow his methodology of using Amplify. Do you think it makes any difference?
Charlie
Charlie,
Yup that is what we would indeed like …
===========================================
one piece of advice - I would suggest that you start out working with some of the LPs that you care about least as your technique and skills will undoubtedly improve with time. I mistakenly started out with some of my favourite LPs and then had to go back and do them all over again when I realized that my capture and processing had improved. Even now I’ll get the odd track on my iPod shuffle and think I could have done that better
WC
WC
I understand what you are advising on starting out with least favorite LPs. Already in just a couple of weeks my techniques have changed and minor tweaking has already yielded big improvements in quality (largely thanks to the ideas and advice that you and Steve have shared with me). That being agreed upon, however, the temptation is absolutely overwhelming to work on some long term favorites that are plagued with Clicks, the background roar of Vinyl, etc. Being weak in character, I must pull out some oldfavorites, even at the expense of redoing them when necessary. It has been incredibly rewarding to hear Dvorak’s “Aus der Neuen Welt” without numerous distracting clicks and that horrible background roar that I hate so much in the ultra-quite passages. Also, I have a “Music Box” on my new car. What a wonderful invention. I simply must get my favorite old LPs into my car’s Music Box.
Also, listening to music in an auto is why I am interesting in Dynamic Compression. Dvorak is probably a perfect example of the problem of Classical music in a noisy environment. I agree, when possible, I like to hear quiet when the Composer intends for it to be, and loud likewise. However, when I move out of the quietness of my Den or away from my earphones, I have another situation entirely. Hence, I am interesting in Compressed CDs for my auto only. I have downloaded and played around with Chris’s Dynamic Compressor. Have not made any adjustments in the defaults, other than the Ratio (per advise in other postings). I certainly do not want to overdo it. Even the .50 Ration seems to take too much of the finesse out of the presentation. When you use Compression for your auto, do you use Audacity Compress or Chris’ Dynamic Compressor? What parameters are you generally using? Any other advice on Compression?
Charlie
I never use compression not evn for in-car use (where I rely on the volume knob - and in my little 2-seater with the lid off you can’t hear the audio at all
Steve should be able to offer some advice on this, but there are a couple of other users’ of Chris’s Compressor on the forum (notably Koz) perhaps thay might comment.
You are lucky in starting out with ClickRepair in your toolbag. I used to do all my early repairs by hand - using the Draw tool in 1.2 originally and later when 1.3 became stable enough I was using the Repair command. But then Koz gave me a steer towards CR and that speeded up my process no end - and gave ne better results!
WC
Chris’s dynamic compressor was written because Chris had exactly the same problem listening to classical music in his car, so he created “Chris’s dynamic compressor” and did an excellent job. In my opinion it is by far the best compressor available for this particular purpose. I find some other compressors easier to work with for other purposes, but for evening out the volume of classical music for car listening, this is the one. I find the default settings to be just about right. Koz usually recommends increasing the compression amount a little (the first control at the top).
For noise/pop/click/crackle removal I use “Gnome Wave Cleaner”, which is free and excellent, but only available on Linux.
Any kit with a single knob for R/L without a center indent may well be off between channels. It may well be off anyway do to component tolerances.
And original recordings may well be off too.
I prefer to just push both channels up to their max (-1.0 in my case).
Should I ever hear the difference I might go back and redo it but more likely to turn the knob on the amp.
For most of us average folks it makes no difference. We can’t tell a 1dB delta like someone who is trained andor works with this stuff all the time.
What is the other compressor you would suggest?
I note that Chris’s is good for home use too!
I currently have an outboard hardware compressor [cheapstuff (<$30) thing to test the concept on the stereo] that makes the music much better (compresses to about 28dB range) except for a click every time it goes from very soft to very loud which is annoying [and older models of this compressor did not have].
I hope to redo all the main classical cds and compress them to under 30dB range so they work at home. No cd in the car yet
Classic Compressor v.117 is a good free VST compressor from the now departed “Kjaerhus Audio”. Classic Compressor by Kjaerhus Audio - Dynamics (Compressor / Limiter) Plugin VST
(the other VST effects in the “Classic” series by “Kjaerhus Audio” are also good).
“ReaComp” is another good free VST compressor, though some of the more advanced features such as side-chain and midi control are not available in Audacity.
There is also a multi-band compressor in the same plug-in pack.
(other “ReaPlug” effects are also good REAPER | ReaPlugs )
Steve Harris’s SC4 is a good (free and open source) LADSPA compressor (cross-platform support), though it lacks “lookahead”. It is available in the LADSPA plug-in pack http://audacityteam.org/download/beta_windows
The new (Audacity 1.3.12) Audacity compressor is not bad, though is a bit quirky in how it works. ( http://manual.audacityteam.org/index.php?title=Compressor ). It now uses lookahead and can compress based on peaks or RMS values.
For Linux there is “Jamin” which can be used with Audacity in real time, routing the audio with Jack for real time processing (requires a fairly fast computer). [JAMin - About]
I enjoyed reading this thread and learned a few tricks.
However, I have a possibly unusual situation: Perhaps 70% of my source material was originally recorded in the monaural era–pre 1960. It is typically 2 channel, but has the same signal in both channels.
How is the choice between amplify and normalize affected when dealing with such material, where the levels in each channel are identical, second by second. Offhand, I’m wondering if the choice of amplify versus normalize then becomes much less relevant?
On a related point:
Much of my source material is 2 channel mono, pre 1960, and mp3 format. By default, I run all mp3s through mp3Gain before committing to my hard drive. I do a lot of playback directly from my PC (not from CDs) and find that mp3Gain helps to even out perceived volume levels between multiple tracks in shuffle play, so I am not constantly fiddling with the volume.
Perhaps only 15% of my mp3s are ever processed by Audacity at all. I only break out Audacity if I hear something I don’t like----a noisy fade, a too long tail, a particular click, etc.
For this 15%, should I undo the mp3Gain changes BEFORE altering in Audacity–particularly if I intend to use normalize or amplify? Historically, I have NOT undone mp3Gain changes before fiddling in Audacity and notice no issues.
FYI, my limited understanding of mp3Gain is that it does NOT alter the file per se—but rather only adjusts the instructions given to the playback unit regarding desired volume. If my understanding is correct, perhaps the mp3Gain changes needn’t be undone, but maybe should be re-done after Audacity normalizing or amplifying?
The choice of Amplify vs. Normalize may still be relevant.
If, for example, these are recordings from a mono record player, then the left/right channels should be identical. However, depending on the equipment and how it was set up, one channel may be louder or quieter than the other channel. In this case Normalizing will be beneficial in that it will correct the imbalance.
On the other hand, if these are stereo recordings from a stereo record player playing a mono record, then the reverse could be true. If the recording had been set up so that the left/right balance was carefully set so that the music was equal loudness in both left and right channels, then it is possible that there could be a crackle on the record that is louder on one side than the other. If this crackle is the loudest thing on the recording, then it will knock the left/right balance to one side if Normalize is used. This is an unlikely but possible scenario.
The third case is that the left/right channels are identical - for example on a ripped high quality commercial CD of an old mono recording. In this case the two effects will be identical.
That is also my understanding, though I don’t know a lot about mp3Gain and don’t use it myself.
Perhaps you could do some experiments, for example, take a very quiet recording, apply mp3Gain, then Normalize it, then apply MP3Gain again. What is the effect at each stage? This could be interesting/useful information to other forum users, so if you have the time and inclination perhaps you could experiment, then write up your findings in a new topic. I’d certainly be interested
Steve:
Thanks for the comments. I think I will conduct some tests with mp3Gain out of curiosity. I’ll have to find some specific source material that might be particularly revealing on this point.
Regarding mono/stereo and normalize/amplify, your comments bring up another question I have been pondering.
Imagine the typical mp3 of a typical studio session conducted in 1955. You open it up in Audacity and see 2 channels. Superficially, the channels look identical, but because of the possibilities you noted in your response, there may be some differences. Or they may in fact be identical.
What is the downside of doing a “stereo track to mono” in Audacity for such material?
Possible advantages:
You end up with a smaller file, which is always good, all other things being equal.
Any left/right imbalances that you noted above in a couple of your scenarios would be eliminated.
The only disadvantage that comes to mind immediately is that some playback configurations MIGHT result in sound then being heard through only 1 speaker rather than both left and right. That would be a show-stopper of course, but I’ve not come across that on my equipment.
Can “stereo track to mono” introduce any unwanted effects?
At what point in the workflow should one make that conversion? I’d guess it should be done before any amplification or fading, but I’m not clear regarding de-clicking or noise reduction.
Historically, I think I have converted to mono at various points in the workflow and haven’t yet noted any issues.
Any comments?
It can, but it is so rare as to be a “freak” event. I’ve seen a grand total of “1” such occurrence here on the forum. The “stereo” track was essentially mono, but the left and right channels were 180 degrees out of phase. (one channel was inverted - possibly due to bad wiring).
In normal circumstances I would convert to mono as a first step as it makes file sizes smaller and processing quicker, but it is always worth just checking after doing anything that the result has not done something weird. Audio is very capable of doing all sorts of surprising and weird things, so it’s best to always have at least a quick listen to the results of every operation before proceeding further.