Do you like your music audio-processed?

[Tom Wells]

One step of encoding, straight to an analog transmitter causes no problems and can be generally smooth and transparent at 192k mp3 encoding. When multiple encoding happens anywhere in the chain, it is correct that bad things happen, and happen fast.
Same as "half-tone" reproduction of photos in printing. One time is OK, but a half-tone of a half tone almost always creates a distracting Moire pattern. But most stations' audio chains now involve multiple digital steps, so I will agree that professional broadcast should be using some kind of lossless strorage.

For tape hiss, nothing was better than National Semicondictor's DNR ( Dynamic Noise Reduction ).
It required no pre encoding, and could be set to a level that killed hiss while not impacting the program material audibly.

 

[elchupacabras]

For tape hiss, nothing was better than National Semicondictor's DNR ( Dynamic Noise Reduction ).
It required no pre encoding, and could be set to a level that killed hiss while not impacting the program material audibly.

Ditto. Best NR ever invented that for the most part didn't change the phase of the source material like Dolby did.

 

[The F Mister]

Let me be clear about one thing. There is no comparison between coded audio and linear audio. Always go for linear in professional environment. I was comparing just those drawbacks of coded audio between the drawbacks of the former standards (tape in this case). Let's take those tube amplifiers, yes they may be warm sounding and better to some. But they are not able to bring the music over 1-1. They color the audio (just what makes them appealing). This I find a drawback, that others may like okay I can understand that even I like a certain analog sound from time to time. But you could never use them as reference, hack they sound different every minute as they warm up.

 

[jesseg]

lossy or lossless, regardless... after 3 pages about processing for taste often just so things are "the same level"...

I'm surprised none of you mentioned ReplayGain. Shame, shame, shame. ;D

http://en.wikipedia.org/wiki/Replay_Gain
http://replaygain.hydrogenaudio.org/

"But Jesse, my playout system doesn't support ReplayGain."

Well, nobody has ever accused me of not having an answer for everything

WaveGain
http://www.rarewares.org/files/others/wavegain-1.2.8.zip
http://www.rarewares.org/files/others/WaveGain 1.2.6.dmg (leopard UB)
http://www.rarewares.org/files/others/wavegain-1.2.8srcs.zip (source)
http://members.home.nl/w.speek/wavegain.htm (Win32 front-end)

MP3Gain
http://mp3gain.sourceforge.net/

AACGain
http://www.rarewares.org/files/aac/aacgain_1_8.zip
http://www.rarewares.org/files/aac/iGainSetup.msi (front-end + AACGain v1.8)

^^ all of the above will work with ANY playout system.

 

[BofH]

And even better there's now SeeDeclip Pro which will literally restore hypercompressed/clipped source WAV files and at the same time apply Replay Gain RMS levelling to give you clean, consistent level cuts ready for broadcast!

http://www.cutestudio.net/data/products/audio/seedeclip/index.php

Download and have a play with the free version! It's an amazing product and should be a must for all radio stations looking to make the most of bad source material from modern CD releases. The Pro version also includes a command line executable that can be used for batch processing. If anyone wants a copy of my batch file, send me a PM!

 

[rrsounds]

lossy or lossless, regardless... after 3 pages about processing for taste often just so things are "the same level"...

I'm surprised none of you mentioned ReplayGain. Shame, shame, shame. ;D

All these systems have one thing in common: They pick a single gain adjustment for the entire piece.

Which is good in that it doesn't mess with levels during a song.

Which is bad if the song has wide dynamic range. But considering how little dynamic range anything has these days, no one will probably notice. ;D

But if a song has a low open, do not expect these software solutions to perform mixing miracles.


Kind Regards,
David

 

[konbaasiang]

All these systems have one thing in common: They pick a single gain adjustment for the entire piece.

True. This can be a great thing, because it gives you a lot of freedom in how to set up the audio processor at the end of the chain. If the audio processor doesn't have to worry about gain riding, or surprises, setting it up becomes less of a compromise -- you can set it for just the right amount of compression, for incredible consistency without a hint of pumping.

///Leif

 

[jesseg]

And even better there's now SeeDeclip Pro which will literally restore hypercompressed/clipped source WAV files and at the same time apply Replay Gain RMS levelling to give you clean, consistent level cuts ready for broadcast!

I'm going to buy this. As a mastering engineer, this is really invaluable. BECAUSE IT ACTUALLY WORKS!!!

I sent comparison demo to Leif a few hours ago actually, and he was also impressed. It's actually not only unclipping pre-clipped (before final brickwall) audio.... AND it's somewhat restoring the other audio that got IMD because of the pre-clipped material.

I would say without a doubt this is the FIRST time I have EVER heard this done. In fact, just 2 days ago I remember commenting on something where I agreed that it was not yet possible. Now I have to remember where that was, and go correct myself

Awesome. Totally awesome.


Oh btw, here's those samples
http://ictybtihky.com/forumfiles/radioinfo/Say It Wrong - Original.flac
http://ictybtihky.com/forumfiles/radioinfo/Say It Wrong - SeeDeClipped.flac
(29 second clips, for non-commercial educational purposes, covered under Copyright Act of 1976, 17 U.S.C. § 107)

 

[rrsounds]

All these systems have one thing in common: They pick a single gain adjustment for the entire piece.

True. This can be a great thing, because it gives you a lot of freedom in how to set up the audio processor at the end of the chain. If the audio processor doesn't have to worry about gain riding, or surprises, setting it up becomes less of a compromise -- you can set it for just the right amount of compression, for incredible consistency without a hint of pumping.

///Leif

In my opinion, the best use for single-gain-value file processing is when the audio is already dense but perhaps at uneven levels, source to source.

It will NOT make a song with a low level opening and/or ending mix any better, nor will it 'fix' the musical pieces with wide variations.

The best one can say is that such software probably doesn't hurt, as long as it only increases gain. But keep in mind that if it ever REDUCES the level of any song, it is throwing away bits of information at the low end of the dynamic range, and if a processor later attempts to bring up those low levels the result could conceivably be ugly.

On the other hand, an AGC can do something that the non-dynamic single-value software can't: it can fix levels at any point WITHIN a certain piece.

And it goes almost without saying, that if an AGC is designed well, it will have "incredible consistency without a hint of pumping." ;D

Kind Regards,
David

 

[Tom Wells]

When I'm dubbing from vinyl, shellac, or tape, knowing that I'll be listening on AM, even with Breakaway and the "cheap tube 2-pass wet box", I still find myself de-fading intros and outros when I'm editing the dub. Some songs I'll hit 5 times on the end with each chunk being another 3 db up until
the last bit is only -6db from the body of the song. Sounds great, I hear the whole song, and sounds tight on air like WLS did.
I'm careful to not go too far as to lose impact on a song that WAS low-level at the beginning, and I never save the result until I've listened to it on air. Sometimes I "fatten" the whole song using Nero's "dynamic processor", where you draw the delinearity curve, then set attack and release times. Naturally, I do what sounds best, which to me means it sounds like WLS circa 1966.
My question is, I wonder how much artistic integrity do I compromise by such fiddling?
Classical music listeners prefer far less processing. Antique music archivists record every scratch, daring not modify what they got out of the groove in any way. I've decided to make the cleaned, optimized, leveled, processed and AM'd versions of these songs my standard
reference. Based on what they'd sound like on WLS, not whether they'd be "better" straight off the phono tube pre-amp listening on the headphone socket with only triode and one op-amp between the vinyl and my ears. (which is breathtaking)

Does anyone else not just process for listening, but actually do that listening on an RF modulated/radio path?

 

[konbaasiang]

David, of course the AGC still has to be well designed.

All I'm saying is that replaygain can help the AGC. No matter how well designed the AGC is (assuming the AGC doesn't use a couple of minutes of look-ahead), things can still get better by pre-levelling the audio by a single gain value, which does not in itself add to the pumping. This pre-levelling essentially constitutes full-song look-ahead, except it's done during ingest instead of during airing, so it adds no latency or gain.

This is assuming that you want to be able to keep a sense of dynamics (for example running the agc at 2:1 or 3:1). If you want the AGC to eliminate any trace of long term dynamics, then there's no point of pre-levelling -- any ol' AGC can do that just fine.


Tom, I do process just for listening (can't live without levelling), *and* I do some of my listening on an FM path! BW 1W exciter in my office, tuners at the front porch, living room, kitchen and bathroom .

I generally use BBP with straight reference settings for the RF-path, and Live with straight reference settings for the monitors in my office.


Jesse, I bought SeeDeeClip! I am thoroughly impressed. Listening to "American Idiot" album at the moment, and for the first time it's NOT grating at my ears. Truly astounding considering what the original CD looks like.

Best,
///Leif

 

[rrsounds]

David, of course the AGC still has to be well designed.

All I'm saying is that replaygain can help the AGC. No matter how well designed the AGC is (assuming the AGC doesn't use a couple of minutes of look-ahead), things can still get better by pre-levelling the audio by a single gain value, which does not in itself add to the pumping. This pre-levelling essentially constitutes full-song look-ahead, except it's done during ingest instead of during airing, so it adds no latency or gain.

This is assuming that you want to be able to keep a sense of dynamics (for example running the agc at 2:1 or 3:1). If you want the AGC to eliminate any trace of long term dynamics, then there's no point of pre-levelling -- any ol' AGC can do that just fine.
<snip>

I think one of the reason the "one size fits all" idea sticks in my craw is that this is the way Dolby does audio for TV: sample an entire show/movie, etc. and create a single number that describes the average loudness.

It TOTALLY ignores the problems people have with loudness variations within the program itself.

Now, if we were to take YOUR concept a step further, and incorporate into this type of software an honest-to-God AGC (looking ahead over the entire file before making dynamic changes) then I can see all kinds of possibilities. ;D

But as long as people believe an AGC has to be emasculated to sound good, OR that it has to be audible to work at all, I will have a job.

Kind Regards
David Reaves

 

[konbaasiang]

Dolby does sample the entire program (with a voice-detector, so that it supposedly only measures voice), and distill it down to one loudness value -- the DIALNORM value.

However, that's not an AGC. Dolby *also* has an AGC (yes, it's awful, it's wideband, and makes nobody happy), and use dialnorm only to get the audio into the sweetspot of this AGC. When it gets louder, the AGC ducks. When it gets quieter, the AGC boosts.

Again, I am NOT saying file-by-file normalizing is an AGC! Heck no. It's simply not true, and besides, I'd be out of a job too!

All I'm saying is that by normalizing the files first, and THEN using a good AGC (setup more conservatively than usual), you can have a better result than you would with that same AGC alone.

Why do I claim this so stubbornly?

Well, I'm glad I asked.

Imagine a well-mastered, quiet, dynamic track from the 80s.. And then imagine a balls-to-the-wall cut from today, with no dynamics whatsoever except for the fadeout. The modern cut also happens to have a much higher noise floor than the 80s cut -- let's say it was mastered through a reel-to-reel deck for extra flavour, before hitting the digital buzzsaw.

Let's say we want the AGC to do all the work, and we want it to have enough power to bring up the quiet parts of the 80s track.. That means a LOT of possible gain. Assuming the song has an average level of -15, and we want to gain 20dBs beyond that. That's 35dB! It will sound fine on the nice, clean 80s track though.

Then, the modern cut comes along. Average level -3. AGC does its job, ducks way down and stays there -- until the song starts fading out. Where the AGC had 20dB of available gain boost in the case of the 80s song, it now has 35dB of gain above the average level of the song! That's a tremendous amount, and can make noise clearly audible even in good recordings.

What could possibly be done to improve the situation?

Pre-normalizing. ReplayGain would push that modern track down 12dB, so that the average is -15 just like the 80s cut. Thus, the AGC will have an equal amount of dynamic gain available for both tracks: 20dB.


What else can I say? It works, and it's not a hard concept to grasp. We STILL need that AGC at the end though, for sure! Without that, we'd still have intros down in the mud.

Best,
///Leif

 

[DJboutit3]

I run a online radio station I try to have the best sound possable while the station is streaming at 56kbps.
Commerical FM stations use a little to much processing they need to turn the bass down about 10% to 15% & the treble down 40% to 50% with to much treble it sounds like there is some wind or a snake in the background. I do not care for that much treble to begin with I really think turning it up a little is ok but most radios should not be able to have more treble than that I love bass.

 

[rrsounds]

Dolby does sample the entire program (with a voice-detector, so that it supposedly only measures voice), and distill it down to one loudness value -- the DIALNORM value.

However, that's not an AGC. Dolby *also* has an AGC (yes, it's awful, it's wideband, and makes nobody happy), and use dialnorm only to get the audio into the sweetspot of this AGC. When it gets louder, the AGC ducks. When it gets quieter, the AGC boosts.

Again, I am NOT saying file-by-file normalizing is an AGC! Heck no. It's simply not true, and besides, I'd be out of a job too!

All I'm saying is that by normalizing the files first, and THEN using a good AGC (setup more conservatively than usual), you can have a better result than you would with that same AGC alone.

Why do I claim this so stubbornly?

Well, I'm glad I asked.

Imagine a well-mastered, quiet, dynamic track from the 80s.. And then imagine a balls-to-the-wall cut from today, with no dynamics whatsoever except for the fadeout. The modern cut also happens to have a much higher noise floor than the 80s cut -- let's say it was mastered through a reel-to-reel deck for extra flavour, before hitting the digital buzzsaw.

Let's say we want the AGC to do all the work, and we want it to have enough power to bring up the quiet parts of the 80s track.. That means a LOT of possible gain. Assuming the song has an average level of -15, and we want to gain 20dBs beyond that. That's 35dB! It will sound fine on the nice, clean 80s track though.

Then, the modern cut comes along. Average level -3. AGC does its job, ducks way down and stays there -- until the song starts fading out. Where the AGC had 20dB of available gain boost in the case of the 80s song, it now has 35dB of gain above the average level of the song! That's a tremendous amount, and can make noise clearly audible even in good recordings.

What could possibly be done to improve the situation?

Pre-normalizing. ReplayGain would push that modern track down 12dB, so that the average is -15 just like the 80s cut. Thus, the AGC will have an equal amount of dynamic gain available for both tracks: 20dB.


What else can I say? It works, and it's not a hard concept to grasp. We STILL need that AGC at the end though, for sure! Without that, we'd still have intros down in the mud.

Best,
///Leif

Leif, you're pretty glib about all this. We'll have to have a discussion this weekend at the Dutch Processing Day gathering, over a couple of Heinekens, if you wish. ;D I'll even buy.

First, are you talking about a dolby transmission processor? What's model number? Last discussion I had with a dolby exec, they claimed there was no need for transmission AGC, that with Dialnorm metadata, any sort of dynamic control violated their concept of preserving "artistic intention." My suggestion of incorporating an intelligent multiband AGC was met wth serious disdain (which made me feel good, actually, because I sensed an opportunity!). I know the "dolby Volume" algorithm has some kind of loudness-related automatic level control, but that is consumer-based.

Next matter...

If you are suggesting that all hot-mastered songs can be lowered by 12 dB to make them closer in perceived loudness to not-so-hot mastered songs, this is basically advocating the truncation of already nasty 16-bit recordings into 14-bit recordings.

Which might be fine as long as there are no low sections in those songs, or any reverb tails. Perhaps 90% (or even more) of today's songs might qualify ... but what about the remainder?

Even then, it is probably still fine if that is indeed THAT, and the file is only meant to be listened to with little or no processing. OR if the file can be saved, and maintained, in a greater than 16-bit format, in which case it won't need to be truncated anyway (and as long as the software is smart enough to 'know' this).

Maybe I'm missing something, but while I really do agree with the need for file level consistency, I simply disagree that reducing a song's bit depth can ever be a good thing when there's even a slight chance that someday it will be followed by an amount of processing that would lay bare this weakness.

In an AGC, we can have an extreme interim bit depth, where there is no risk of truncation of low level material while being processed. What comes out can be tailored to fit into the available bit-depth. And while most of the equipment I know of can handle 24-bit I/O, the "CD" storage standard (i.e., 16-bit, 44.1kHz) which most stations adhere to, is and will most likely remain 16-bit for the foreseeable future. Which means these newly truncated "14-bit" songs could be around for a LONG time.

One last thing:
I might remind you (assuming you've discovered this already), that ALL processor manufacturers have only the best intentions, but there will ALWAYS be more users who push the equipment as far is it can go to the processing extreme, than will even CONSIDER using the prettier presets we might wish them to use. :

IOW, while Replay Gain and their ilk may tend to allow a reduction in processing, it's very optimistic to expect that the bulk of users will follow your suggestion. :'(

Kind Regards,
David

 

[jesseg]

If you are suggesting that all hot-mastered songs can be lowered by 12 dB to make them closer in perceived loudness to not-so-hot mastered songs, this is basically advocating the truncation of already nasty 16-bit recordings into 14-bit recordings.

First off, ReplayGain done properly doesn't truncate the bits. It re-quantizes the new LSB, using dithering and usually noise-shaping.

Here's something you can try out... try blind ABXing some WAV/FLAC originals run through lossyWAV.

http://wiki.hydrogenaudio.org/index.php?title=Lossywav
http://en.wikipedia.org/wiki/LossyWAV

What it's purpose is for is inaudibly removing unused bits in certain frame sizes matching lossless codecs so that they perform MUCH more efficiently. It's not lossless after lossyWAV (duh) but the whole point is that you can't hear it until it's used at extremes.

But even at very very safe settings, you're going to find that MOST songs are able to be reduced to 14bits (and even less) without ANY audible change at all. Don't believe me, look at the stats, TONS of research, and most important of all... blind ABX test some samples for yourself. See if your ears can beat the dozens of experienced ABX listeners that helped make lossyWAV what it is over the last few years.

I've actually seen quite a number of songs that average to around 12bits, without any audible difference. And ya know... it's most often the new produced songs that are slammed to all hell and back. The exact same tracks you mentioned would cause a problem for being bit-reduced. Again, don't believe me, or the years of research & stats... you really must see and hear this for yourself.

Even then, it is probably still fine if that is indeed THAT, and the file is only meant to be listened to with little or no processing. OR if the file can be saved, and maintained, in a greater than 16-bit format, in which case it won't need to be truncated anyway (and as long as the software is smart enough to 'know' this).

Yeah, but even at 24bit depth there's still no reason to truncate. Even a simple shaped TPDF can help, even at 24bits. Just because you personally can't hear something on your own gear, doesn't mean it never will be heard, so do it anyways.

IOW, while Replay Gain and their ilk may tend to allow a reduction in processing, it's very optimistic to expect that the bulk of users will follow your suggestion. :'(

Well... if they decide not to without taking the time to educate themselves, then it's just a decision made out of fear. If they decide not to after educating themselves about it, then they have found a better way around this issue, or just LIKE a very strong AGC riding everything. Either way, the education is what empowers them to make a good decision.

 

[konbaasiang]

Hmm.. I looked up a few definitions of "glib".. well.. Thanks, I guess? .

I do have a bit of experience with TV audio and metadata, from designing the software for the OctiMax and Aeromax line of products for Linear Acoustic, from the ground up.

The AGC in AC3 (Dolby Digital) I mentioned is part of the metadata, and is not actually applied to the audio but sent along as a control value (1 value every 256 samples, applied through windowing in the decoder), but that doesn't make it any less of an AGC. There are actually three different gain values: DIALNORM, which is constant for each program, DYNRNG which is meant to provide reasonable dynamic compression for home viewing, and finally COMPR which is heavy compression/limiting, meant to prevent overloading of RF modulators downstream.

Here's how the audio chain is supposed to work according to dolby:

- Program comes in from network, has metadata with proper DIALNORM value set.
- Program AND metadata gets sent to encoder (let's use the old Dolby 562 as an example)
- Encoder analyzes audio with respect to DIALNORM and creates DYNRNG and COMPR metadata (i.e. runs a compressor/agc)
- Encoder outputs an AC3 bitstream
(snip)
- AC3 bitstream gets decoded inside consumer set-top box, DYNRNG or COMPR applied, and audio is output.

However, metadata usually gets lots at one or several steps, and even if it didn't, Dolby's compressor doesn't work all that well. Thus, I have a job in TV audio processing .

They don't like to call it an AGC, because hollywood wouldn't like that, and it can be defeated (by digging through menus in the receiver).. But, a system that analyzes the audio (in realtime as it's encoded!), creates gain values to reduce the dynamic range, and finally applies the gain value further downstream.. Well, if it looks like an AGC, and smells like an AGC, what could it be?

Dolby Volume is another story altogether. That IS a wideband compressor, and a marketing department. Imagine how good that wideband compressor would sound if they had Vorsis' marketing department :-D.



Reducing the audio to 14 bits?

Sure, why not!

It should have been 14 bits to begin with. I listen to plenty of classic rock and other well mastered recordings that average -12dB or less, and even extremely dynamic classical music, all in 16-bit format. Never once (in otherwise well recorded cuts) have I heard a bit of quantization noise!

Worrying about the bits at the lower end of the scale, in a smashed clipped 16-bit recording, is worrying about the WRONG bits!

Let's play with the numbers. Assume we're clipping just 1% off the top of the waveform.. That's essentially an error of 1%, affecting the MSB.

How many LSBs would we have to discard to come up with the same voltage error?
I believe the answer is 9. Starting with 16 bits, removing 9 bits, we have 7-bit audio -- 128 different voltage levels. Roughly as accurate as 1% of clipping.

Yes, this is an apples to oranges comparison, but it paints a colourful picture, doesn't it?


I would *much* rather listen to a well mastered and properly dithered 14-bit track (or even 12-bit!) than a smashed 16-bit track.

I've found the true noise floor of most recordings (even 24-bit DVD-A's) to be about -70 to -80dB. 16-bit gives you 96dB SNR, whereas 12-bit gives you 72dB. 12-bit is cutting it awfully close (especially if you want to AGC afterwards), but 14 is fine. Older digital recordings, which were not pushing for maximum recordings, essentially had this resolution anyway, and sounded great -- much better than modern smashed recordings.

Looking forward to seeing you this weekend! This will be a lot of fun. I accept, but make it Belgian beer . Man, I haven't had Chimay for so long.. Not available here.

I'm looking forward to hear / play with the Ariane too, I actually haven't had a chance to play with one yet.

///Leif

 

[jesseg]

The CD Audio standard itself was actually designed for 83dB reference levels as reflected in K-System K-20, and exemplified by ReplayGain. The fact that we're still using CDs at all shows the wisdom Sony & Philips had when they created the standard... and as Leif said, lots of those early CDs (from 1982 to 1985 especially) still sound better than MOST CDs that are released today. And it's no secret that today's AD/DA converters (and dithering) are of MUCH better quality than they had back then. It's the top end of the scale as Leif so eloquently pointed out.

Btw, I bought SeeDeClip Duo Pro the other day as well, and have been continually loving it for a few days now. I've already used it in one master... to de-slamify 1 track before starting on it last night. I'm so loving it.

 

[rrsounds]

Ya know, I REALLY like that part from the medical doctors' Hippocratic oath that says "First, do no harm." I've even considered using it in advertisements. ;D

If you are suggesting that all hot-mastered songs can be lowered by 12 dB to make them closer in perceived loudness to not-so-hot mastered songs, this is basically advocating the truncation of already nasty 16-bit recordings into 14-bit recordings.

First off, ReplayGain done properly doesn't truncate the bits. It re-quantizes the new LSB, using dithering and usually noise-shaping.

You're actually reinforcing some of my points.
To begin...
My antenna ALWAYS goes up when someone uses the qualifier: "done properly."

Next, one cannot assume ANYthing about activity in the lowest number of bits by only examining the overall loudness level. The automatic assumption that, just because the top bits are trashed, it's therefore OK to trash the bottom bits, is specious. You may get lucky, heck, you may always be lucky! But, personally, I try to keep luck and chance at arm's length.

Further, I'm sure you respect the idea that dithering is not something to be applied casually or at numerous stages. Let's say the CD mastering engineer struggled to dither material down to 20 dB below the CD noise floor. Maybe he spent weeks on it, getting it just right...Is ReplayGain going to respect that when it blindly reduces gain a further 12 dB? Always?

Here's something you can try out... try blind ABXing some WAV/FLAC originals run through lossyWAV.

http://wiki.hydrogenaudio.org/index.php?title=Lossywav
http://en.wikipedia.org/wiki/LossyWAV

What it's purpose is for is inaudibly removing unused bits in certain frame sizes matching lossless codecs so that they perform MUCH more efficiently. It's not lossless after lossyWAV (duh) but the whole point is that you can't hear it until it's used at extremes.

But even at very very safe settings, you're going to find that MOST songs are able to be reduced to 14bits (and even less) without ANY audible change at all. Don't believe me, look at the stats, TONS of research, and most important of all... blind ABX test some samples for yourself. See if your ears can beat the dozens of experienced ABX listeners that helped make lossyWAV what it is over the last few years.

I've actually seen quite a number of songs that average to around 12bits, without any audible difference. And ya know... it's most often the new produced songs that are slammed to all hell and back. The exact same tracks you mentioned would cause a problem for being bit-reduced. Again, don't believe me, or the years of research & stats... you really must see and hear this for yourself.

I'm confident you can find thousands of examples of songs that have limited bit-depth, or that if bit-reduced, don't present obvious side effects. It the risk of sounding sarcastic I would respectfully answer "Duh!"
That's the world we live in! But that description does not fit every song, which is the crux of my argument.

IOW, just because a destructive technique works transparently with certain files, maybe even a statistically large number of them, doesn't make its use categorically applicable to ALL files. Particularly since we have zero knowledge of how the future will treat those files. Future processing, future personnel, etc.

Over and over, in this post and others, I read: "listen for yourself!" But almost without fail, these examples are files that have not been post-processed in any way, and post processing is a HUGE unknown! My gut feeling: to assume that in our industry there will be little or no processing after ReplayGain (or lossyWAV, or whatever) is, at best, standing upon shakey ground.
Cuz, hey, even though some 128kbs MP3s sound acceptable until they're processed. ;D ... still they DO get processed.

IOW, while Replay Gain and their ilk may tend to allow a reduction in processing, it's very optimistic to expect that the bulk of users will follow your suggestion. :'(

Well... if they decide not to without taking the time to educate themselves, then it's just a decision made out of fear. If they decide not to after educating themselves about it, then they have found a better way around this issue, or just LIKE a very strong AGC riding everything. Either way, the education is what empowers them to make a good decision.

Or they just don't know.
Education only comes when someone has the desire (and the time!) to gain knowledge. The 'tweekers' amongst us are in that category. But once again, my experience is that tweekers are a very small minority of the universe of people who are called upon to use all these tools we are giving them. We HOPE they use them with intelligence, but we must never ASSUME that will be the case! Have you listened to the radio lately?

As a designer, I have what I call my "Guardrail at the Grand Canyon" rule. There's a guardrail at the Grand Canyon, so tourists won't fall and kill themselves. Now, rock climbers don't need a guardrail, they need a rope! But we don't hand out ropes to tourists. They get the guardrail.

You may consider yourself to be a "rock climber," rather than a "tourist," and, God bless ya, YOU may have a great skill set. Grab yerself a rope!
But I try not to confuse the two categories, because MOST people we will meet belong in the 'tourist' category, for good reason.

I have to very carefully consider the outcome before I start handing out ropes to them.


I believe you misunderstood what I said here:

Even then, it is probably still fine if that is indeed THAT, and the file is only meant to be listened to with little or no processing. OR if the file can be saved, and maintained, in a greater than 16-bit format, in which case it won't need to be truncated anyway (and as long as the software is smart enough to 'know' this).

When you responded:

Yeah, but even at 24bit depth there's still no reason to truncate. Even a simple shaped TPDF can help, even at 24bits. Just because you personally can't hear something on your own gear, doesn't mean it never will be heard, so do it anyways.

Because my point was that a reduced level 16-bit signal will fit 24 bits just fine; of course you won't need to truncate it. Or do anything else. Human hearing is an amazing thing, but it doesn't have 144 db dynamic range! But, unfortunately OTOH, neither do most radio stations have 24-bit storage systems.


And finally, that last line of yours sums up my whole reason for doubting the wholesale incorporation of ReplayGain etc:

"Just because you personally can't hear something on your own gear, doesn't mean it never will be heard"

BINGO! Words to live by!

Kind Regards,
David

 

[littlejohn]

Gentlemen -

The last several posts remind me of the country church during services (I grew up in the very rural Southeastern US) when Brother Eddie was testifying. He testified about his fight with temptation when he and Sister Jane were climbing in the hills. And testified about the second fight with temptation as they continued their climb. As he began his testimony of his third fight with temptation, an LOLwBH (Little Ol Lady with Blue Hair). stood up and said "Reverend, make Brother Eddie be quiet. He ain't testifyin' any more, he's braggin'!".

Gentlemen, more testimony and less brag, whattaya say?