Quote Originally Posted by RobotCzar
Sorry, they are hard to get even in a library as they are found in the now defunct "Audio" magazine.
It figures. And you complain about someone throwing out a claims with no support. You are guilty as charged my friend.



I mean that they are not accurate to the original performance, they are processed to make them sound "better" (i.e., more sweet, dramatic, etc.). This process would be similar to digital keying in video (where color and contrast are "enhanced" to make the scene prettier or more dramatic).
How do you know what the original performance sounds like, were you there? If you have ever heard a live recording tape before post, you would know that they disperately need some tweaking before mastering.

Isn't post production for tailoring the product to fit the medium. So what do you think we should do? Record a live event, and release it with less the perfect levels, instruments not fully fleshed out, balance not quite right, and dynamics that distort the 16bit medium? ALL recordings have to be tweaked to a degree, but of course as a knowledgeable professional audio engineer you already know these things right?

You have never even heard what a original performance sounds on tape(or hard disk) so by what perspective can you even bring to this topic may I ask?

It makes even less sense to reissue analog recordings done at the equivalent of 12-bits and claim "20-bit" performance.
A recording made with 12bits and no dither would be unlistenable.

How do you know what resolution the analog recording is? Have you every personally measured? This is a inflammitory statement made to support and ignorant perspective. Not helpful to the topic at all. No one, NO ONE has ever measure the resolution of an analog source to make a comparison to digital resolution, so where do you get 12bit from, your bum?

Sampling an analog recording and 20-bits that is performing at significantly less than 16-bits is not going to make them sound like 20 or even 16-bits, despite what the marketing department wants you to believe.
Do you have any evidence to support you claims that analog recordings perform at less than 16bits? Or is this just more of your inflammatory posturing just to make a point(you know you are VERY good at this). You have absolutely no scientific evidence to support your claims that analog recordings(expecially quality ones) do not perform at at least at 16bit level. However there is ample evidence that support the notion that analog recordings carry significantly more bandwidth the 16/44.1khz redbook CD. 1",1 3/4" and 2" magnetic tape with NR can outperform the redbook CD format quite easily.

The noise shaping process (which trades some noise in more audible frequencies for more noise is less audible frequencies) does nothing to reduce noise that is already in the recording. Active processing by a human or a digial algorithm is the only way to lower the noise of such recordings--and that WILL affect audible frequencies (i.e., increase distortion).
If you knew the principles of noise shaping, you would understand the purpose is not to remove the noise in the recording, it is to use the psychoacoustical nature of our ears and push the noise to an area were it is not audible to the human ear. It doesn't trade the noise at all, it simply moves it up in frequency. If you shape the noise so that its prominence is out of our most sensitive area of hearing, you are effectively increasing signal performance by lowering the noise floor. That is why they can claim 20bit performance from 16bit recordings. There are not saying you are getting 20bits of signal quality, they are saying you are getting 20bit performance. BIG difference!

The only thing noise shaping (which is where the quasi "20-bits" come from) can do is to lower noise which can only be heard if the recording is very quiet AND the music is also very quiet (or silent)--hence genre matters. Noise shaping gives the "equivalent"of 18-19 bit noise performance ONLY in the targeted frequency range (noise overall cannot be better than the 16-bits of the playback medium).
If you understood how the ear works, you would also understand that it is not necessary to remove noise in a broadband fashion. Noise in the bass region is masked by signal, and by our ears relative insensitivity in the region. Noise in the upper frequencies are not audible to the ears also based out of relative insensitivity. Obviously you wouldn't have a need for 20bit performance on loud rock and roll where there is little dynamic changes, and the level of noise is well below the recorded level. This is probably the only "genre" of music this applies to. But since the recording world doesn't surround itself around this genre of music, then the genre of music does not play a role in this at all. Trying to make this a technique of genre is a side show to the main point. Since gospel, classic, jazz, world, some types of rap, brass choir, string quartets, small wind ensembles, mass choir, jazz ensemble, big band, organ solos and many other types of "genres" have periods of low level signals, they could benefit from this type of processing. To make your point around the genre of music is majoring in minors.

My on hands experience does not equal your uneducated, non hands on experience. You are just plain wrong about 18-19bit performance from HDCD. According to what I measured from my boards HDCD plug in(and this isn't even from the outboard box which is probably of a better quality than the plug in) between 200-8khz the noise level before processing with HDCD was -96db to the noise floor. That is 16bit performance from a 16bit chain of recording and processing.(which was VERY difficult to do cleanly). After encoding with HDCD, the noise floor in the same area was closer to -119db which is indeed 20bit performance. That is why I purchased the plug in, because it performed as advertised. Now according to my last count, over 120 studios have HDCD processing in the inventory of processors in this country alone. Amoung this 120, a list of the most prestigious and prominent grammy award winning mastering engineers(studios) are amoung them. This includes Bernie Grundman Mastering, Gataway mastering, Skywalker sound, 5.1 entertainment group, Dts entertainment group, and about 20 more that I cannot remember. Do you think one company's marketing department is smart enough to convince these mastering experts that they product can do as claimed, without first proving the point to them?



It is not silly, it is the very definition of "high fidelity". You are correct saying that a live performance cannot be perfectly replicated, but the goal is to get as close as possible. The decision of what row you are in IS dependent on the recording engineer (and producer), but an accuate first row or back row is still the goal.
No recording engineer worth their salt worries about a back row, front row perspective while recording. That is a perspective of a man who sits in his chair in his living room and thinks that he knows all there is about digital recording when he hasn't so much as placed a single mike in his life.

Recording engineers have QUALITY first in mind, and pull together the finest microphones, mixing board, and processing boxes he can get his hands on to acheive that end. Since it is impossible to do a back row recording without either placing the mikes there(which would lead to a reverberant field recording with a poor frequency response and no imaging whatsoever) or they would have to add a ton of reverb to the mix in post processing (which would make the recording sound artificial and unnatural). That is not the "accuracy" that you claim is highly desireable.


A good recording engineer should do what he can to make an accurate recording (given target home audio systems). A recording engineer SHOULD NOT act as an "artist", but in pop music they have to (again, genre matters).
But that would require sweetening, and you are totally against this. Did you know that recording engineers do not have to be an artist at all if that was desireable, even with pop music? Haven't you ever heard of a live studio recording? That can be done with just about any genre of music there is. Anita Baker made two of these recordings, and aside from eq, and some compression, these recording contained no artistic additions at all. The artist decides what artistic approaches they want, and when they finish, then the recording is accurate. What is accurate is based on the artist, producer, or engineers perspective, not yours. The recording itself is just a foundation layed, it is up to the artist to build the building.

What do you mean by a "better artist"? One that makes music that is more pleasing to YOU? Or to me? The point is that "accurate" is potentially objective, "better" is completely subjective.
A artist should make music that is pleasing to them first, then it should be pleasing to you. It is their vision after all that you are sharing, and not the other way around. So the point is, what becomes accurate is based soley on the perspective of the artist, and if the artistic vision includes eq, limiting, compression, fancy panning, dogs barking in the background, sounds panned in a circle frantically and all processed in HDCD, then that is an accurate version the artist wants to portray. And guess what, you have no right to judge it otherwise because it is THEIR vision, not yours.

Showy recordings are like enhanced pictures found in glossy magazines--highly processed and not accurate, but pleasing to many viewers.
If this is what the artist wants to portray, then its accurate to them. Enhanced pictures found in glossy magazines are accurately what an editors wants in his magazine. The judgement of what is accurate belongs to the artist, not you. How would you know anyway?



Old analog recordings often did not achieve 16-bit performance (this is why digital recording is better). New recordings are not quaranteed to acheive 16-bit performance either.
What scientific analysis can you draw from that makes these statements accurate? Most new recording offer much greater performance than 16bit can reproduce. That is why most recordings(especially those headed for SACD and DVD-A release) have to be downsampled to even fit on a 16bit platform. Most new recordings(with wide distribution) are recorded at 24/48khz, 24/96khz, 24/88.2khz, 24/176.4 or DSD format these days. Even small scale studio projects are outputted at 24/96khz or 24/88.2 because the equipment to do this is so cheap now.(computer based DAW come to mind). Almost no one records anything at 16/44.1khz anymore because it leads to too many problems in post processing(no headroom, lack of bits for eq, compression etc). Some of the best sounding digital products come from analog background(early Telarc recordings on Soundstream equipment(which had a frequency response to 50khz, much greater than 16bit 44.1khz can do). Mercury living presence recordings all had information above 20khz in them. Significantly greater bandwidth than 16bit 44.1khz has. Any analog tape above 1" with noise reduction could outperform 16/44.1khz. So you claims come from your head, and not from scientific evidence.

A lot of mixing of tracks recorded at different relative levels will result in a recording with less than the performance of the recording equipment.
This is a convoluted statement if I ever read one. Without a specific sample or bit rate this statement is meaningless. Even if the levels were all different, when summed at the board they would still require more than 16bits to process them, and would required compression and limiting to fit within the 16bit medium.

(Note therefore that multi-track mixdowns are likely to have less than expected performance so 20-bit recording makes some sense--there is more "headroom" for mix downs.) Again, as no live reference exists for multi-tracked recordings and such recordings are very common in pop-rock but not jazz or classical--genre matters.
More convoluted statements, and some plain wrong statements. Your thought process is derived out of the dark ages or not well thought out at all. Classical and jazz are all multitracked these days. What do you think, they only use two microphones for stereo recordings, and five for multichannel?

NO! that is not how it is done. My last recording of big band jazz used 12 microphones. The last large symphony recording I did required 10 microphones(spotlight mikes are necessary in large halls with large orchestras). Now the only way to use minimalist microphone techniques are with small ensembles in small recording spaces. Almost all forms of music require multiple microphones. Now you may only have five that you pull from all of the time, but you definately need more for spotlighting solos or some individual instruments.

Secondly, If I sum all of the outputs of these 10-12 microphones recorded at 24bits, without compression I would be using greater than 24bits of signal. Each microphone input adds DB's to the overall mix. That is why they have invented mixing boards with 64bit floating points.

Nice talking with you, I strive for civility, but it is hard at times.
I am sure it is really difficult, especially when you pretent that you know what you are talking about, and you don't really.

Once again, the HDCD process has a white paper on record with AES. SACD, DVD-A Dolby Digital, and Dts all have white papers on record. The beneifts of SACD and DVD-A have been debated, refuted, and sucessfully rebutted. Dolby digital was completely refuted, and found to perform at 18bits, not the 20bits they claim. Dts and HDCD white papers have not been refuted by the peers of its inventors. There is a reason for this, and that reason is they perform, or have performance characteristics as stated in the white papers submitted. Now you who complains that engineers do not follow scientific methods, are guilty of the same. You have thrown off a bunch of statements that are neither scientific or particularly accurate. What they show is that you have cursory knowledge, but trying to present defined statements.

Have you ever watch an engineer mix audio? The how do you know they don't follow scientific methonds? Or what methods the use in general?
Have you ever recorded or measured any audio? Then how do you know what the resolution of any recording(analog or digtal) is.
Have you ever had the job of placing microphones for a recording? Then how do you know what a correct technique is, or in what instances more than 5 are needed?
Have you ever done a multitrack recording in your life and measured the resolution of your results? Then how do you know the resolution of varying levels of multitracked channels are.

If the answer to these question are what I think they are, then you have just wasted a alot of bandwidth with a bunch of self thought up foolishness.

Lastly, your effort to demonize audio engineers is gross and insulting. Remember, they are sitting behind the mixing boards moving the controls. You are sitting in your rocking chair listening. Why in the heck should I listening to a rocking chair quarterback(who normally doesn't know mixing from a rat turd), rather than listening to the quarterback who really runs the game.