[kphooligann]

I'm somewhat of a relic of an audio enthusiast or audiophile.  I got my start with a turntable based system I bought with money earned as a paperboy about 35 years ago.  But I think I have some questions about MP3 and wav playback that others may have as well. Here they are:
 
1) Will a good DAC or Soundcard do anything worthwhile, $200 worth, for MP3 playback?
2) If one "plays" an 128k MP3 file, I assume the computers motherboard or soundcard will "see" a 128k signal.
    Also, that those devices will convert the signal to an analog signal put out through the analog out jacks. But, what signal should I                           
    expect as output from the spdif jack?
3) I am under the impression that "red book" CD means a 16 bit sample taken 44,100 times per second of an analog signal, is this 
    correct?
4) MP3 files are composed of what size samples taken how many times a second?
5) Oversampling, in theory, can not improve sound.  Am I correct?
6) Radio Paradise and Cinemix streams are so called lossless streams, correct?
7) In conversion of an analog signal to an MP3 file, not all sound frequencies are treated equally, correct?
8) Playback of an MP3 file requires a codec, right?  Is this software that converts the file into an analog signal? Or, software that 
    converts the file into another digital signal?  Or, non of the above.
9) Given that outside of Radio Paradise and Cinemix, all the music I listen to is MP3 quality, would my next audio dollar be better 
    spent on a headphone amp or headphones rather than a DAC?  BTW, I currently use DR 150 headphones 
    and have a set of T50RP 'phones on the way.
10) Can the difference between wav files and MP3 files be stated in terms of sample size and sampling rate?
       And, if so, what is it?
 
Thanks ahead of time for your feedback!

 

[Chodi]

Quote:

I'm somewhat of a relic of an audio enthusiast or audiophile.  I got my start with a turntable based system I bought with money earned as a paperboy about 35 years ago.  But I think I have some questions about MP3 and wav playback that others may have as well. Here they are:
 
1) Will a good DAC or Soundcard do anything worthwhile, $200 worth, for MP3 playback?
Probably not much if any improvement in 128kbs mp3
2) If one "plays" an 128k MP3 file, I assume the computers motherboard or soundcard will "see" a 128k signal.
Yes
    Also, that those devices will convert the signal to an analog signal put out through the analog out jacks. But, what signal should I                           
    expect as output from the spdif jack?
The spdif jack is just a digital stream meant to be connected to an  external dac
3) I am under the impression that "red book" CD means a 16 bit sample taken 44,100 times per second of an analog signal, is this 
    correct?
I can't even imagine why you are worried about this
4) MP3 files are composed of what size samples taken how many times a second?
See the answer above
5) Oversampling, in theory, can not improve sound.  Am I correct?
There is a debate about this but in general at this level the answer would be no it cannot
6) Radio Paradise and Cinemix streams are so called lossless streams, correct?
I have no idea
7) In conversion of an analog signal to an MP3 file, not all sound frequencies are treated equally, correct?
You are correct that is one reason it is a lossy format
8) Playback of an MP3 file requires a codec, right?  Is this software that converts the file into an analog signal? Or, software that 
    converts the file into another digital signal?  Or, non of the above.
A codec is a set of instructions which in this case allows the conversion to analog or vice versa
9) Given that outside of Radio Paradise and Cinemix, all the music I listen to is MP3 quality, would my next audio dollar be better 
    spent on a headphone amp or headphones rather than a DAC?  BTW, I currently use DR 150 headphones 
    and have a set of T50RP 'phones on the way.
Your best choice would be to get as far away from mp3 as possible. At 128kbps it is pretty bad quality. If you must use it then there is little point in spending for high end equipment.
10) Can the difference between wav files and MP3 files be stated in terms of sample size and sampling rate?
       And, if so, what is it?
WAV is lossless but takes a lot of file space. That is why flac is so popular as it can greatly reduce the size of the file and produce the same lossless results.
 
Thanks ahead of time for your feedback!

If you must use mp3 for some reason then you must go up the 320kbps to approach lossless reproduction. Anything less than 192k is so lossy that the playback equipment is not of much importance. Buying quality phones to listen to 128k mp3 is a real waste of good headphones.

 

[kphooligann]

Thank You Chodi,
 
I'm not complaining, just defending question 3.  Well, let's say the upper limit of human hearing is 15 kHz. Let's then round 44.1 kHz to 45 kHz. So now, we sample a sound signal, which can be modeled as the superposition of many pure tone sine waves upon one another at 3 times the highest audible component frequency and that's accurate?  I'd like to see 160 kHz sampling rates, representing eight samples of a 20 kHz signal per one full cycle. How many bits should that sample be I can't guess.
 
 I imagine that the best examples of recorded music in the last fifty years were the Reference Recordings direct to disc LPs.  These recordings were uncompressed and analog, the master was directly cut from a microphone feed.  Red book CD can't possibly match that!
 
I'm just trying to nail down the diferences between MP3 and wav files and how MP3 files are ultimately converted to analog when played on a computer or MP3 player and passed through an outboard DAC.
 
 I gather most outboard DACs do not decode MP3 files but are fed bit streams of 16 bit samples at 44.1 kHz rate or maybe 24 bit samples at a 96 kHz rate.  I don't know, but, those DACs that do have indicators, don't indicate for any MP3 bit stream but indicate for wav file frequencies.  This is why I wonder what's coming out of an spdif jack.
 
BTW, there is an interesting web site, http://www.noiseaddicts.com , which will let one hear back to back MP3 and wav files.  I found the differences very subtle with the DR 150 'phones plugged directly into a computer.

 

[cel4145]

I'm just trying to nail down the diferences between MP3 and wav files and how MP3 files are ultimately converted to analog when played on a computer or MP3 player and passed through an outboard DAC.

See this explanation: http://www.soundonsound.com/sos/may00/articles/mp3.htm

 

[ClieOS]

You seem to misunderstanding what 44.1kHz means - Nyquist–Shannon sampling theorem states that you need double the sampling rate to get the full range of what you are trying to sample. So sampling rate of 44.1kHz will actually give you the frequency response up to 22.05kHz, which is only a bit more than what human hearing range is usually defined (20kHz).
 
Also, following any DAC stage is a LPF (low-pass filter) stage to filter out what is above 20kHz. The reason of doing so is because the sampling process will create noise and the sampling process will push the noise over 20kHz (by design). So whatever is beyond 20kHz (can also be up to 22kHz depends on the filter design) is filtered out and not audible. This is true to even 96kHz or 192kHz sampling - anything beyond 20kHz should not make it to your headphone. Then why do we oversample? Because the higher you sample, the higher the noise will be push upward, making filtering an easier job and result in more linear phase in the FR that you want. But you are still getting under 20kHz of sound. Here is a good white paper from Lavry Engineering explaining oversampling in detail.
 

 

 

[kphooligann]

Thank You ClieOS,
 
I ran into a reference to Nyquist Shannon during my web searches to answer my own questions.  I guess I'm skeptical of it because I doubt I could understand the proof.
After pre-Calculus my math grades dropped like a rock.  But, I'm willing to believe it rather than continuing to be silly about it. I'll check out the Lavry Paper.

 

[kphooligann]

Thank You cel4145,
 
I'm sure the Sound on Sound article will go well with the YouTube video by Longbottomrx titled Mp3 Encoding and Decoding and a diagram of "MP3 Audio" I found at analog.com
in their A Beginner's Guide to DSP .  I was for a short while under the impression that MP3 files were created directly from analog but now I realize they are mostly made from red book cd wav files or bit streams.  I imagine I can think of decoded MP3 as a recreation or facsimile of an original wav file or bit stream.  Let me read on.

 

[Currawong]

Quote:

I'm somewhat of a relic of an audio enthusiast or audiophile.  I got my start with a turntable based system I bought with money earned as a paperboy about 35 years ago.  But I think I have some questions about MP3 and wav playback that others may have as well. Here they are:
 
1) Will a good DAC or Soundcard do anything worthwhile, $200 worth, for MP3 playback?
2) If one "plays" an 128k MP3 file, I assume the computers motherboard or soundcard will "see" a 128k signal.
    Also, that those devices will convert the signal to an analog signal put out through the analog out jacks. But, what signal should I                           
    expect as output from the spdif jack?
3) I am under the impression that "red book" CD means a 16 bit sample taken 44,100 times per second of an analog signal, is this 
    correct?
4) MP3 files are composed of what size samples taken how many times a second?
5) Oversampling, in theory, can not improve sound.  Am I correct?
6) Radio Paradise and Cinemix streams are so called lossless streams, correct?
7) In conversion of an analog signal to an MP3 file, not all sound frequencies are treated equally, correct?
8) Playback of an MP3 file requires a codec, right?  Is this software that converts the file into an analog signal? Or, software that 
    converts the file into another digital signal?  Or, non of the above.
9) Given that outside of Radio Paradise and Cinemix, all the music I listen to is MP3 quality, would my next audio dollar be better 
    spent on a headphone amp or headphones rather than a DAC?  BTW, I currently use DR 150 headphones 
    and have a set of T50RP 'phones on the way.
10) Can the difference between wav files and MP3 files be stated in terms of sample size and sampling rate?
       And, if so, what is it?
 
Thanks ahead of time for your feedback!

 
Lets see how I go at answering these:
 
1. A better DAC probably wont do anything for old MP3s which are very compressed, as they will show up the audible artefacts created by the compressed. Later variable compression algorithms are pretty good to the point that some members of the forum had trouble telling apart 128k and 320k MP3 files.
 
2. The "k" in MP3 bit rates refers to kilo-bits-per-second, which is the rate of data transmission required for playback. It is different to the 16 bit, 44.1 kHz (kilo Hertz) being referred to as the most common sampling rate used. What goes out an S/PDIF jack, without getting into the technicalities, is an uncompressed PCM signal. The computer decompresses the MP3 (or AAC) file and sends the uncompressed PCM data to the DAC. Unfortunately as MP3 compression is "lossy", that doesn't make the data lossless. If you have, say, a FLAC or ALAC file, the compression used by those is "lossless" and retains all the information about the music.
 
3. Red Book refers to the format for an audio CD. That data is 16/44.1 (to use the shorthand for it), but once you copy a CD to your computer each track becomes an audio file of one type or another.
 
4. See 2. 
 
5. It's too early to worry about oversampling and upsampling. They refer to things that occur inside DACs which are complex and contentious. 
 
6. Radio Paradise (one of my favourite streaming stations) sends a lossy signal, equivalent to a 64, 128 or 192 kbps MP3. If such a thing as lossless streaming existed, it would take up a huge amount of bandwidth and cost a huge amount of money, as 10x the data rate would be required per stream.
 
7. Yes. Quieter sounds or sounds about 16 kHz are removed first (depending on settings). Newer VBR compression methods use heavier compression during quieter parts of the music and lighter compression during more complex parts.
 
8. A codec allows the computer to convert a type of file to a format necessary for playback. So an MP3 codec would allow it to be converted to digital PCM data. Not analog. Analog conversion is done inside a DA chip.
 
9. Headphones should come first, but what comes next depends on what you are using for playback. I'd probably say a DAC and amp should come together, as one without the other probably wont produce a lot of noticeable benefit, beyond an amp increasing the amount of bass somewhat.
 
10. WAV and AIFF files are uncompressed audio data. MP3 files are lossy compressed audio data. The numerical difference between them is the bit rate. A 16/44.1 WAV or AIFF file has a bit rate of 1411 kbps. Depending on the compression, an MP3 will have a considerably lower bit rate.

 

[kphooligann]

Thank You Currawong,
 
Thank you for your answers. I have one more question.  Is the following statement true, false, or neither:
"In MP3 audio,  a red book CD pcm signal can be encoded into one of several MP3 bitstreams, such as 32, 64, 96, or 128 kbs.  Upon decoding, a pcm signal having 16 bit bit depth and 44.1 kHz frequency is produced."

 

[Currawong]

Quote:

Thank You Currawong,
 
Thank you for your answers. I have one more question.  Is the following statement true, false, or neither:
"In MP3 audio,  a red book CD pcm signal can be encoded into one of several MP3 bitstreams, such as 32, 64, 96, or 128 kbs.  Upon decoding, a pcm signal having 16 bit bit depth and 44.1 kHz frequency is produced."

 
You could say:
 
"In MP3 audio, PCM audio data from a Red Book CD can be encoded into one of several MP3 bitstreams, such as 32, 64, 96, or 128 kbs.  Upon decoding, PCM audio data having 16 bit bit depth and 44.1 kHz frequency is produced."
 
But this is is perfectly possible independent of CDs. Just remember that "Red Book" refers to the way data is organised on a CD for audio play back. You could as easily make a data CD with AIFF or WAV files on it which has the same data (but it wouldn't play back in a CD player).

 

[LordOctron]

Okay, I gave my best to sum-up the en/de-coding process in laymen terms, perhaps it's of use...
 
With a CD you start with a digital representation (PCM/WAV) of the original analog signal, bit-depth and sample-rate are the "limits of precision" for the digital representation, this limits are specified by the CD-A Red Book standard. MP3 compression (as most other lossy audio compressions) basically does two things, first it tries to find the "least audible information" in the signal and removes em (that's the lossy part), then it makes use of traditional compression methods to reduce the size even further (that's the non lossy part). The size of the resulting MP3 depends on the bit-rate that was specified when compressing the file, the bit-rate again influences how much "least audible information" has to be removed (which influences quality). Now, when playing the MP3 you have to convert the compressed data back to a PCM/WAV again, that's what the MP3 decoder does. Then the PCM signal gets passed to your output device (sound-card or standalone DAC) which converts the digital PCM to a analog signal which you can feed to your amp etc.

 

[kphooligann]

Thanks M'Lord every little bit helps (no pun intended).

 

[Hutnicks]

Quote:

 
Lets see how I go at answering these:
 
1. A better DAC probably wont do anything for old MP3s which are very compressed, as they will show up the audible artefacts created by the compressed. Later variable compression algorithms are pretty good to the point that some members of the forum had trouble telling apart 128k and 320k MP3 files.
 
2. The "k" in MP3 bit rates refers to kilo-bits-per-second, which is the rate of data transmission required for playback. It is different to the 16 bit, 44.1 kHz (kilo Hertz) being referred to as the most common sampling rate used. What goes out an S/PDIF jack, without getting into the technicalities, is an uncompressed PCM signal. The computer decompresses the MP3 (or AAC) file and sends the uncompressed PCM data to the DAC. Unfortunately as MP3 compression is "lossy", that doesn't make the data lossless. If you have, say, a FLAC or ALAC file, the compression used by those is "lossless" and retains all the information about the music.
 
3. Red Book refers to the format for an audio CD. That data is 16/44.1 (to use the shorthand for it), but once you copy a CD to your computer each track becomes an audio file of one type or another.
 
4. See 2. 
 
5. It's too early to worry about oversampling and upsampling. They refer to things that occur inside DACs which are complex and contentious. 
 
6. Radio Paradise (one of my favourite streaming stations) sends a lossy signal, equivalent to a 64, 128 or 192 kbps MP3. If such a thing as lossless streaming existed, it would take up a huge amount of bandwidth and cost a huge amount of money, as 10x the data rate would be required per stream.
 
7. Yes. Quieter sounds or sounds about 16 kHz are removed first (depending on settings). Newer VBR compression methods use heavier compression during quieter parts of the music and lighter compression during more complex parts.
 
8. A codec allows the computer to convert a type of file to a format necessary for playback. So an MP3 codec would allow it to be converted to digital PCM data. Not analog. Analog conversion is done inside a DA chip.
 
9. Headphones should come first, but what comes next depends on what you are using for playback. I'd probably say a DAC and amp should come together, as one without the other probably wont produce a lot of noticeable benefit, beyond an amp increasing the amount of bass somewhat.
 
10. WAV and AIFF files are uncompressed audio data. MP3 files are lossy compressed audio data. The numerical difference between them is the bit rate. A 16/44.1 WAV or AIFF file has a bit rate of 1411 kbps. Depending on the compression, an MP3 will have a considerably lower bit rate.

To your point 7 here. The other element that has not been touched upon is the encoder itself. There is a tremendous difference in MP3 encoders and most of the time you never know what your file has been encoded with unless you have done it yourself. This results in a wide variety of playback quality for the same encoded bitrate. In short 2 320Kbit files of the same song can sound vastly different.
 
As mentioned previously running as far and fast from MP3 as you can  is your first, best upgrade path.