Really basic advice needed

[JanVigne]

Steve, I would consider the function employed by a lossy program to make your decision regarding format.

MP3 or similar "throws away" data points (read that to mean "signal") for the sake of fitting more data on a smaller storage device. The (questionable) science behind this application revolves around the psycho-acoustics of human perception. Keep in mind perception is considered to be those electro-chemical reactions which exist beyond the ear drum. Our understanding of the perception mechanisms is far from complete. For example, we do not understand how someone can, in an overly crowded and noisy, environment, detect their own name being spoken or the sound of a familiar voice. For that matter, we don't understand how we come to be aware of someone staring at us behind our back from across a crowded room.

We do know human perception is able to fill in certain pieces of information through extrapolation of the incoming signal. With this in mind, the authors of the MP3 format suggested we are unable to clearly perceive quiet sounds when loud sounds are also present. Therefore, MP3 formats will discard signals such as violins when horns are also present. With a solo instrument there are no competing signals to speak of. Therefore, we need to look at bit rate of the format to determine how much data will be discarded by the various MP3 rates. Since bit rates vary from 96 to 320 Kbps (kilo-bits per second) selecting higher recording rates will yield results closer to the quality of WAV or FLAC formats. Of course, the higher the bit rate in MP3 format, the less data you can store on a given media. The thing to remember IMO is any MP3 format will be a "lossy" format which does not provide a full picture of the original event. The "compressed" MP3 file cannot be expanded in the same manner an analog compressor/expander allowed. In the case of lossy formats, once the data has been lost, there is no mechanism which can restore the data points.

There are several compressed or lossless formats available which will allow for higher quality sound than any MP3 format would provide. Some are proprietary to certain storage systems. iTunes is a technically lossless format, as is FLAC. Many listeners have a problem with anything iTunes and avoid the format on principle. FLAC however is considered to be equivalent to a WAV file in sound quality by most users. WAV is the equivalent to CD quality and, other than the higher resolution formats which exist above 16 bit/44.1kHz, will require the greatest amount of storage space. Neither WAV nor FLAC discards data points. What you put in is (technically) what will come out as all data points are represented in the original. FLAC files will take up considerably less storage space than will WAV since FLAC compresses the file in the recording process and expands the file on playback. WAV employs no such compression and the data is stored and read in real time. FLAC compression/expansion occurs "on the fly" or as the process occurs and does present the problem that not all players will handle a FLAC file. WAV on the other hand is a universal format which can be played back on virtually any modern player. IMO this makes WAV the preferred format given the cost of storage space found in today's computer systems. The vast majority of players can handle storage devices up to 16g which would mean you should have a very difficult time exceeding the capacity of such a device. You can pick up SD cards with 16g storage for a few dollars.

WAV files will contain all the information digital systems are capable of capturing and can be upsampled/upconverted to higher resolution formats at any time if so desired. FLAC files can be upsampled/upconverted though the process will be somewhat more complicated than would a WAV file. Upsampling any file, WAV or MP3 cannot replace data lost in the original recording. This means a WAV file can be somewhat upgraded through upsampling/upconversion though for your purposes, this is unlikely to offer any significant improvements in sound quality. Beginning with a MP3 file however, upsampling or upconversion will be a waste of time since there is a significant lack of data contained on the original file.

Once you have captured the recording in WAV, the file can be downsampled to a lower bit rate though some loss of data and sound quality is then to be expected. Therefore, storing the original file in WAV is likely to be the best option since the file can be altered to higher storage capacity or higher sound quality with relative ease. An original MP3 file can be downgraded to provide higher storage but can never be upgraded to higher sound quality.


As to the audible differences between bit rates, I would have to repeat the advice I gave to clients when I was selling audio; I am not responsible for what you cannot hear. Audio has become a commodity used more as sonic wall paper than you and I remember. For many listeners having access to hundreds if not thousands of lower quality songs trumps having fewer songs at a higher sound quality. Certainly, when someone was interested in purchasing a high quality sound system, they were not encouraged to use such a system with their existing MP3 files as the differences in systems would be largely lost with such low quality sources.

[sdelsolray]

Originally Posted by Fran Guidry

Just for jokes I Googled up Harvey Gerst. He's been a busy man.

He owns and operates a recording studio: http://www.itrstudio.com/

He designed speakers for JBL, wrote songs with Roger McGuinn that were on the first Byrds album, designed amps for Acoustic Control, and more: http://tapeop.com/articles/bonus/harvey-and-alex-gerst

And he shares his recordings with us: https://soundcloud.com/harvey-gerst

I think I'll take Harvey's advice.

Fran

One thing Harvey is wonderful at is helping inexperienced folks with recording issues. His famous "great mic threads" over at the Home Recording forum is a good example of the time and care he takes to help.

[dhalbert]

Originally Posted by JanVigne

iTunes is a technically lossless format, as is FLAC.

Do you mean Apple Lossless, aka ALAC? (http://en.wikipedia.org/wiki/Apple_Lossless) I had not heard of it as "iTunes" format specifically.

[JanVigne]

Yes, ALAC.

I stay away from iTunes as much as possible and couldn't off hand think of the name. Steve hasn't expressed any interest in the format so I didn't bother to go further.

[steve s]

Originally Posted by JanVigne

Yes, ALAC.

I stay away from iTunes as much as possible and couldn't off hand think of the name. Steve hasn't expressed any interest in the format so I didn't bother to go further.

My wife has an iphone and ipad. Never crossed my mine that they might be of service here. She keeps them pretty well filled up with photos, so probably not worth my while to go there.

Steve

[Fran Guidry]

Originally Posted by steve s

...
My next big decision point will be when I have to decide whether to stick with the WAV files or go with some level of MP3 in order to fit stuff onto a CD. Do I believe the experts who say that blind A/B tests prove that people can't tell the difference or the experts who say they can?

Steve

If you're talking about a standard music CD that you stick in any player, you have no choice. The audio format for a music CD is 44.1/16 WAV. Even if you feed MP3 or other compressed formats to your CD burner it will convert on the fly to WAV.

This is not the case if you're talking about a data CD, which some players these days can handle. In that case you can use MP3 to get hundreds of songs onto a CD, but an audio CD is good for about 70 minutes of music.

On the subject of audibility of compression, you don't have to believe experts. You can download free software and do the comparisons yourself. Anyone can do this and find out for certain what they can hear and what they cannot. It's sometimes hilarious how resistant some people are to conducting these simple tests.

Fran

[steve s]

Fran,
Thanks for straightening me out on the CD business. I just assumed that what worked for MP3 players would work on CDs. It would appear my streak of being wrong on every single point is unbroken.
Steve

[JanVigne]

"It's sometimes hilarious how resistant some people are to conducting these simple tests."




It's sometime frustrating to see how poorly such "simple tests" are performed.

Maximizing the consistencies between samples and minimizing the multiple variables between something so simple as an "A or B" test is a struggle which persists in every experiment made by every listener who attempts such "simple tests". Well, actually not. Many just plunge in with the idea all of this should be so simple it requires no thought at all. Their results generally reflect a plan with no real thought applied. Most of the tests conducted by the average listeners can easily be kicked over with simple logic. Some are more difficult because we must deal with the frailties of human perception. But a truly successful test is highly unusual, even among the professionals who do this everyday.



First, be aware of the "placebo" and the "no-cebo" effects which dominate every test. Whether you feel this way or that, what you believe, or what you feel you should believe, will cloud your judgement. If you feel "this" is true, then you are 90% of the way to concluding you are already correct. If another person is involved in the testing procedure, quite often they will give off signals - most often these signals are not consciously made - which will tilt the test subject's conclusions in favor of one result.

We could ask the question of auditory memory. But many of us have different answers regarding how long we can retain the sound of a music sample when under test conditions. Are you testing the differences between two samples only? Or, are you testing which sample strikes the listener as more closely resembling "real world music"? There is a distinct difference between the two questions being asked in such tests.

Second, remember that even a slight difference in volume level between two audio samples will lead the listener to predictably choose the louder sample as being better in sound quality. Even if those level differences are momentary and due to dynamic changes in the material. Creating two audio samples which are identical in level is far more difficult than it might seem here since the psycho-acoustics of MP3 compression relies on the louder signal or the lower frequency masking those lower level or upper frequency signals. Doppler effects alter human perception of both low and high frequency signals which makes a simple equal volume comparison unlikely here.

Most importantly IMO, we must accept the fact any playback system can be no better than the source player allows. This is the conflict of high end audio in many cases since a more transparent system has as its purpose the ability to more completely translate the limitations of the source to the perception of the listener. So, where do you set your point where you have a system which is transparent enough? Most certainly, if you are performing tests meant to determine the audibility of the source limitations, you must have a system which is transparent enough to reflect the quality of the source. $39 dollar headphones or computer speakers are unlikely to show any real differences in a low quality source. If your simple test is already limited by the poor quality of the playback system, it's a safe bet your conclusions will be less than we could hope for.

Next, we all listen for our own priorities and, if those priorities are not reflected by easily perceived differences, once again, the conclusion of "no difference" will have been clouded by personal bias and another listener will likely come to another conclusion based on their own values.

Include in all of these variables which lead to conflicted results the idea any test of what you can hear is the least common way in which we listen to music. While there are times when we listen for sound quality, for the most part we listen for enjoyment, not for critical comments. Performing a test in which we are essentially testing the tester leads to conclusions which have little to no bearing on how we will listen when we are after simple enjoyment rather than simple testing.

Then ask yourself which is the best method for testing. Should you perform single or double blind experiments? Is "A/B" enough? Or, should you acquire an "A/B/X" switcher which takes the "either/or" out of the testing procedure? Or, is sighted testing enough? Many listeners feel they are better at distinguishing differences when they understand the changes made in the source and system. The decades long, continuously divisive arguments between the audio objectivists and the audio subjectivists revolve around the simple choice of what music material to use and whether the listener should be aware of when changes have been made.


I can site many other variables which would easily affect the outcome of any comparison between "this" audio sample and "that" audio sample but it should be obvious there is no such thing as a "simple test" when it comes to working with a highly subjective thing such as music.

[kscobie8]

I just deleted a bunch of posts. Most of them with gobs of unnecessary bickering. There might have even been some good information in those posts, but it was lost in the war of egos and endless blathering. It simply wasn't worth the time to edit out the junk just to keep the little bit of decent info.

Stay on topic. Avoid the bickering and cheap shots. Help the OP, please.


Bel isi,
-kyle

[JanVigne]

Never mind that last post.


I do resent the removal of my last few posts though. Why leave Fran's and take away mine?

Just to be clear, IMO Fran's post is snarky and rude and should be removed if you think the thread has deteriorated. The point being Fran is exhibiting the exact behavior suggested by, "If you feel 'this' is true, then you are 90% of the way to concluding you are already correct." To leave his post and remove my reply is, IMO, to show bias toward one form member. Especially in light of the other posts which have been removed.



"Simple tests" - meaning tests without careful thought - can lead to simply incorrect results. OK? Ignore that fact and your results and conclusions are going to suffer.

Hope that passes muster.

[JanVigne]

OK, remove that post too.

[KevWind]

OK Ill amend my post to reflect just your statements and not personalities and adhere to (be nice) , I suggest you do the same.

You state:

Originally Posted by JanVigne: OH! I assume, since you didn't answer my question, you don't have a clue what a "multi frequency, source instrument" is either? Strange you would use it in a sentence then, isn't it

nuff said

Originally Posted by JanVigne:
And we come up with terms such as "multi-frequency/multi-source" which have no definition even by the person who posted the term.

Nope simply dead wrong, yes I used the the terms , however the use and the definition was clearly stated in the SOS article linked and within the context of my posts, which I thought actually self evident "Sorry you don't grasp that concept." (quote is yours BTW)
That you attempted to dismiss the article and my use of those terms, does nothing to invalidate either the article or the terminology.

In my posts I used the phrase "practical" in practical terms, in practical application etc. multiple times. That I was referring to practical "in use in recording" even though you claim otherwise, was in fact evident in almost every post I made. And was actually the subject of the thread in the first place. Acknowledging that a guitar could produce a point source sound, but also stating that if true that has no practical application in recording or listening to a guitar is in no way polar. So in reality there is absolutely no polar aspect to my posts NONE, you are simply wrong.

The fact is although you attempted to dismiss it, the terminology multi- source instrument and multiple frequency, was used by someone other than me (i.e. the author of SOS article and Gerst spoke about the same idea in the analogy about a 3 way speaker system and a guitar). The fact is for all your posting your assertions about point source remain unsubstantiated. There is not one instance, you provided no (link)or a quote where anybody other than you, referred to a guitar as a point source instrument . So we can only assume it is your opinion and terminology and go from there.

The truth is that in "practical recording" terms and mic positions , what the mic will hear is multiple frequencies from multiple locations on the instrument i.e multi sources of guitar sound, strings. sound board , sound hole, some mic positions more so and some less (i.e. 3 ft.out and 2 ft up from the sound hole vs the 6" from neck joint for example) but still "multi-frequency/multi-source"

Experimentation is discouraged. Follow our recipe or fail.

is false. as well as moving again into the "not nice", because
The fact is no one,..... again no one has discouraged experimentation, or claimed "follow or fail" ever...

Can we agree that digital clipping and undesirable distortion occurs at anything above (0 db)

Can we agree that any digital clipping degrades the signal and degrades the sound, unlike analog tape saturation . ?

Can we agree that maintaing peaks below 0 db is best practice ?

If can agree to the above, then why does it matter if different people recommend different (- db levels) isn't it obvious that different people will often recommend different methods? The fact is they will work and not produce clipping. As I have stated several times I would tend to agree that for a solo acoustic guitar recording I would personally have no qualms about even -3 db peaks

[JanVigne]

"-OK Ill amend my post to reflect just your statements and not personalities and adhere to (be nice) , I suggest you do the same."





REALLY?!


C'mon, guy!


What exactly is your idea of being nice?


GEEEEEEEZ!!!

[kscobie8]

Well, I tried. My apologies to the OP. Hope you got some useful info.


Bel isi,
-kyle