Another "pretty" waveform...eh?

[David R. Modny]

...and they wonder why we bitch:

[David R. Modny]

Here's how it used to look before it was "remastered". Mind you, the original was probably taken from a third generation tape at best, and then no-noised, but at least it still *resembles* music. Pick your poison.

[Steve Hoffman]

Well, what is it?

[Larry Naramore]

Originally posted by Steve Hoffman
Well, what is it?

Can't resist, should go to bed and call it a day/night. Anyway it's Track 13. :(

Bad joke? :)

Goodnight!

[David R. Modny]

Originally posted by Steve Hoffman
Well, what is it?


"Awaiting On You All" - All Things Must Pass CD reissue.

Now, I purposely picked this track as it was always, IMO, one of the worst-sounding/recorded tracks to begin with...very Compressed and Spectorized. A whopping 18 db of average dynamic range throughout the song to begin with. We all know that the original CD issue was taken from like a 5th generation tape, and then all the highs were killed with No-Noise. Yet, that CD at LEAST retains the original limited dynamic range.


The "remaster" squashes the overall average dynamic range to approx. 9 db TOTAL! Most of the tops of the waveform (check out the right channel) have also been sheared off via the Ultra Maximixer!

What's the point of having a hi-fi rig and boasting about the 100db+ dynamic range capabilites of a format, when engineers are completely stripping whatever little dynamic range rock music has to begin with, totally away.

[Steve Hoffman]

Geez, scary!

[David R. Modny]

Originally posted by Steve Hoffman
Geez, scary!

The waveform pretty much shows what many people described hearing when the remaster first came out. That is, a LOUD, dynamically deprived sound. Again, I'm not claiming the original CD was any great shakes - in fact it was pretty awful in its own way - it's just that the remaster IMO introduces a whole new bag of worms. Hardly an improvement, unless one's listening on a boombox.

What's the point of using decent, first-generation tapes, and then just mangling the sound because these tools just happen to exist. How could he think that an average dynamic swing of 9db is somehow an improvement???

Loud, Punchy, crap...

[Grant]

Yes, I see it all the time when I rip a track into Cool Edit. I'm used to it.

Just to see how bad they ruin the sound, you can do this too by using the Hardlimiter in Cool Edit Pro or the Wavehammer in Sound Forge. You can try to minimize the effects by adjusting the attack and decay but... A slightly better way is to use the compressor and tailor a smoother compression/limiting scheme, and then boost that.

I sometimes do this by ONLY boosting the volume two or so decibals. But never six or more! Even the default values for the audio editors are 6db. To me, that's just encouraging the less-informed users to ruin their CD-Rs. I can use this tool because I am educated about how dangerous it is, what it's uses are, and how to use it very sparringly!

Some music benifits from this compression/maximization, but I'm guessing most people here don't deal with it.

What is so bad about having to adjust the volume knob?

[Jason Smith]

Another excellent job by our man Jon Astley! I hope he gets to do the Beatles remasters!

He must have severe hearing problems. How could anyone listen to that?:rolleyes:

[Claus]

:mad: :mad: :mad: :confused:

[Dave]

Geez, talk about suckin' the life right out of the music...:rolleyes:

[jroyen]

It is possible to maximize volume, while still maintaining a highly dynamic digital recording. Using this technique there are a great many benefits gained with respect to resolution - mostly in low-level detail.

However, there is a fine-line between maximization and improper usage. It is only when the waveform is clipped, transforming it into a square wave, where this threshold becomes readily apparent. At this point, there are no techniques available to recover the lost bits.

What's most disturbing is that something this apparent and fundamental eludes many seasoned professionals.

Josh

[David R. Modny]

I think it's important to state that Astley hasn't really hard clipped the signal as we know things - it never exceeds O db full scale - but rather, he has just dynamically squashed the signal while maximizing it to full scale in order to achieve maximum loudness. Here is where I think many engineers are letting everything they've heard about resolution and bit depth of 16 bit audio confuse them.

While, yes, it's true that 16 bit digital has its maximum "resolving power" in the upper regions of full scale (the shortest distance between the positive and negative voltage of the signal), I don't believe overall DYNAMICS should ever be compromised in order to achieve this. Many modern engineers seem to think that it's a must to squash the signal so that most of these "upper region bits" gets used. This is INSANE!

Instead, let the music's NATURAL dynamic swing simply PEAK at the highest level possible while recording (without clipping) and LEAVE IT BE. In the case of anything with *real-word* dynamic range...you'll easily fit the original dynamic signal into 16 bit scale and preserve the original dynamic swing of the signal. Unless were talikng about some classical acoustic recording with its softest point 90 db down from maximum, were not going to have to deal with any low-level quantization problems (and even then its questionable as to whether or not we'll actually hear them).

Bottom line - preserve the full dynmics of the original signal in order to achieve maximum fidelity. Leave any compression (preferably analog) as a creative tool on the recording side of things or, if doing digital compression, use lightly and USE ONE'S EARS. As far as maximization/ normalization goes, better to let your input levels peak at the highest possible setting during recording/mastering (without clipping ), and LEAVE IT. Normalization DOESN"T increase resolution. A signal with a maxaimum input of -6b from 0db full scale is a 15 bit resolution signal (6db=1 bit). Normalizing it just raises everything (including the noise floor) lineararly to full scale...it's still 15 bit "resolution".

[jroyen]

I'm certainly not in favor of making sounds louder than they were initially meant to be. Yet, the more amplitude a waveform contains the more information it has available to resolve sound.

Captured too softy and you lose low-level harmonics and depth, too loudly and you lose clarity, dynamics, and focus - on either end it is destructive. Therefore, great care must be taken to ensure optimal conditions in low or very high (>0dB) signal levels.

It is the main reason why early Redbook CDs, in being straight transfers, were described as often having a gritty sound. The solution was as much an eventual compromise of early digital as many engineers had to overcome in mastering for vinyl.

It is a limitation of 16-bit recordings, and lessens to the point of irrelevancy at 24-bits. While, the extra eight bits of 32-bit floating point actually affords the same audio resolution no matter what level it is processed.

Josh

[David R. Modny]

[QUOTE]Originally posted by jroyen
I'm certainly not in favor of making sounds louder than they were initially meant to be. Yet, the more amplitude a waveform contains the more information it has available to resolve sound.

Which I believe letting the peaks of a signal naturally come as close to 0db as possible accomplishes. That way, maximum or near maximum scale has been achieved and no dynamic range compromises to the original signal have to be made. Unless were talking about something with say 70 or above db of dynamic range, a mastering with properly recorded levels (i.e. letting the peaks come as close to 0db as possible) won't even come close to the theoretical quantization noise floor bottom (where all these "nasties" are supposed to take place) of 16 bit, IMO.




It is the main reason why early Redbook CDs, in being straight transfers, were described as often having a gritty sound. The solution was as much an eventual compromise of early digital as many engineers had to overcome in mastering for vinyl.


I find it interesting though, that so-called "straight transferred, Redbook CD's" are essentially what DCC and Steve's work are and have always been (up to this current era of DSD transfers). No workstations used, no post signal processing in the digital domain, etc. Is it the "tube mastering" he uses that simply lessens the inherent "grittiness" of the format? He seems to get pretty decent results. No?

Personally, I'm all for a 24/96, 24/192 world (or the latter number even better as in SACD). But more for the reasons of the latter number. I'm a personal believer that the sampling frequency is the more important factor of the two in determining a more "analog-like" representation of the sound. What I'm saying is that, even with a true 24/96 bit consumer format, or SACD, or whatever,... that a dynamically compressed and maximized signal (as with the original waveform shown) is still gonna sound like ****. :)

[Dave]

David R. Modney,

Isn't it like Steve has always said... "It always boils down to the mastering itself"? Is the bit-rate really all that important?

[David R. Modny]

Originally posted by Dave
David R. Modney,

Isn't it like Steve has always said... "It always boils down to the mastering itself"? Is the bit-rate really all that important?

I think that's my entire point. The mastering is still where the most mucking up can occur (as evidenced by the Astley waveform). In order for the true benefits of DSD to reach the consumer ear, we're going to have to stop having engineers reduce the signal to boombox-like proportions in the first place! What's the point of a "high-resolution" format, when we're buying "remasters" like the original one mentioned in this thread?

[Dave]

Originally posted by David R. Modny
As evidenced by the Jon Astley wave-form.
Agreed and thanks! Sorry about the prev. mis-spelling. :rolleyes:

[Grant]

Originally posted by jroyen
It is possible to maximize volume, while still maintaining a highly dynamic digital recording. Using this technique there are a great many benefits gained with respect to resolution - mostly in low-level detail.

However, there is a fine-line between maximization and improper usage. It is only when the waveform is clipped, transforming it into a square wave, where this threshold becomes readily apparent. At this point, there are no techniques available to recover the lost bits.

What's most disturbing is that something this apparent and fundamental eludes many seasoned professionals.

Josh
Yes. But make no mistake, the dynamics are being compromised to a point. It's all a matter of when it becomes audiable to the listener.

With some types of material, I think you can still get away with a couple of decibels of boost. Then, there are things you don't want to touch.

[Burningfool]

I'm sorry to say that I'm getting a bit lost in the technical details here (even though I love them because I learn things here!) I wonder, however, if anyone here has done a good needle drop of ATMP. After I listened to the remaster I went and dug out an old cassette that I made from a mint orange label Capitol pressing. It sounds great compared to the CD. I'd love to have a clean transfer on CD.

Which leads me to another question: since I know this is not the proper forum to suggest/offer/request such trades, I wonder if some kind soul (or Gort) could direct me to the appropriate place. I'm sure there are lots of good needle drop remasters that would be fun to trade.

Apologies if this is an inappropriate post. Inquiring minds want to know, that's all!

[Grant]

I have done an awlul lot of needle drops over the yaers, too many to count!
Too bad it's mostly stuff i'm sure don't much interest most of the forum members.
I have not done ATMP. I don't own the LP, either.

[Chris Desjardin]

I have done a needle drop of ATMP from a mint, original 1970's Japanese pressing. The sound is amazing! I did a Cool Edit remaster of it, and IMHO, it blows away the sound of both the original and remastered CD's. The sound is light and airy, not loud and compressed like the remastered CD. I am actually going to re-remaster it now that my Cool Edit skills have improved in the last few months. In addition, I just made a needle drop of a mint U.S. promo copy of ATMP (Original 1970's pressing on Apple) and I will compare the sound of the two pressings. Send me a private message if you would like to trade vinyl remasters.

[jkerr]

Chris,

It be very interesting to hear of your comparison of the 2 pressings. Not too long ago I got a mint copy of the Japanese ATMP off of Ebay (for not too much IIRC, and was actually mint for a change!). I agree the sound of this pressing is much better than either cd issue.

[jroyen]

I agree with Grant. Any smart engineer will master within the constraints of the medium, whether they be limitations of tape size and speed, groove size and stylus speed, or bit-size and sampling frequency. Just as that same engineer would master differently for different genres of music.

Therefore, the Astley remaster - based on redbook limitations - cannot be considered a dynamic pop recording - but far outside it. In fact, it is close to being the dictionary definition for everything wrong with modern digital mastering. :)

I would just like to point out that there are still two sides to dynamic range, as being the difference between the loudest and lowest. Moreover, there is a fine-line between everything in a recording falling into place, and everything falling apart - certainly perceived loudness plays an optimal part.

If you look at anything Steve has ever mastered and compared with its catalog counterpart, you would see just how well he understands the importance of perceived loudness. Of course, after knowing what it takes to get tonality correct it may seem like trickery, but probably no more than those used to bypass the limitations in vinyl.

And in some ways, current engineers over-compensate for the differences in resolution of various medium. So in this particular instance, for Astley, bit-rate rate is the more important consideration, as it somewhat helped spawn such extreme maximization. I can only defend its practice, not its abuse. That's not to say that the peak loudness of the latest Ricky Martin or the latest Peter Gabriel remasters are directly correlated either.

Although, normalization is not likely the only thing being employed in those instances. Then normalization does not compress sound either. The average piece of normalization software will only linearly increase gain to automated levels, irrespective of source amplitude - with no clipping, and very little distortion.

That is why waveform pictures, like the above examples, can be extremely deceptive. Because there is a lot for which such a close shot does not account, such as RMS calculations. Yet, some might further argue that RMS Power isn't the most meaningful way to represent perceived loudness. As mathematically, it refers more to current or voltage, and not true power.

Josh

[Sckott]

I did the basic comparison with Joihn Coltrane's "Blue Train". You push the real dynamics right out of the recording when you blow everything to the sky. A simple EQ move would actually improve a lot of recordings, but this kind of butchery is what make low-fi sound "better". It was much better when the LP said "No, I won't do that" and engineers had to abide by the rules, or you had bad product.

With a CD, you can do things like this and get [mostly] away with it.