I have listened carefully to .mp3 vs vinly converted version. I have noticed the difference before. I have compared tracks with their waveform in Audacity, and most of the time what I hear matches. I have stopped listening to my entire mp3 collection.
Much of this is very preliminary, as 'there is a problem with my system' I am not using any sort of Hi- Fi amplifiers at the moment.
I am sorry I do not have the confidence to buy any sort of amplifier on the market, except that what I have heard sounds decent to me, I know I am out of my league here. I had a bad experience with the Lepai digital amp that one expert on You Tube recommended, PC speaker boards and the PAM 8403 are much better in my opinion.
I realize that to 'return to vinyl' needs a very good system to make full use of the potential that is on vinyl.
I looked at the tracks by zooming out in Audacity - they peaks stay where they are. I dont know how the FLAC format reconstructs lost information.
Hi,
Mp3 are not like linear format which reproduce at output what is at input: they use destructive algorythm to 'butcher' your source linear file ( in order to have less data in the final output file): it is based on frequency and amplitude present in the recording.
It's based on 'masking effect': when 2 signal are present, let's say a lower frequency one and an high frequency one, depending on their relative level if the lower frequency one is louder it will 'mask' ( in part) the higher one.
So the algorythm makes a constant analysis of the input signal and decimate the 'masked' frequency in your initial signal to lower the volume of data to be encoded.
Then a data compression is applied to save on the data volume to be saved ( but it's not audio dynamic compression).
It is clever approach as it's based on psychoacoustic but it have drawback: it introduce distortion of the initial signal and the harmonic content of treated source is modifyed.
That's the typical 'TR-808 syndrome'* (as we called it) with cymbal encoded in low rates.
If encoded at higher rates it is less and less audible but still present and some can still identify the effect in good condition ( low distortion loudspeaker, low background noise environment, direct instant comparison, etc,etc,).
FLAC are entirely different: they only perform data compression. If only focused on audio they can be considered as linear, but the resulting files are then treated in a way their final data 'weight' is reduced. It's a format compression type of file, audio remains untouched.
Ok,
I believe you when you say you heard difference between different formats. But when you say what you see match what you heard i'm sceptical**.
Let's take your previous example of screenshots of peaks: you compare the same file on different level of zoom, so on different timescale. Of course the end redults will be different despite what you observe is still the same thing!
Let's take an analogy: you take a walk into a forest. At ground level there is a large amount of variation of tree size: there is low height and higher ones. There is even variation in between the higher ones, some are bigger than others and it seems there is a large amount of variability even looking at the same species.
Now let's take a step back at the same forest but on your way home, on a road on a hill some kilometers away from it: now the canopy seems very consitant in term of height, no large variation on it, despite what you saw earlier when walking in it. Some trees ( the smaller ones) even seems to have disapeared...
This is what happen when you 'zoom' in the waveform in audacity: you compare a large time scale ( your landscape view) to a small timescale ( your walk in forest): there is no change in the audio, only the scale you look at it.
I'm pretty sure you did not took notice of this kind of things and made a preset view setting to compare your files at same timescale when comparing your visual representation of your different files. So you probably compared apple to orange.
Other thing, i told you about rms level of your files: have you performed an analysis of each files and then compensated the louder one to meet the lower one?
If not then here again it's apple to orange as the intrinsic level of each file is different from track to track...
I don't know Audacity as i use different tools but you'll have to learn all this very boring things to perform any kind of visual analysis/comparison before performing any kind of meaningful comparison and have conclusion you can rely on.
Gear, gear,gear... well i agree there is a treshold you should not cross about quality but i would not obsess over it either.
About amplifier: a good old chipamp used in it's SOA ( safe operating area/range) was/is more than enough for some very serious about quality brands you'll find in almost every recording/mixing facility around the world. To assess the quality of what you as end listener will evaluate in the end.
And it is the same with classD or whatever.
It's mainly in the application/use and coherency in the chain which is important in my view.
Try to make a coherent/homogenous reproducing chain ( no need for a fancy 'whatever link' in your chain if the other links are not on par) and you'll be happy.
*: TR-808 is a famous Roland 80's beatbox where cymbals (ride) sounds are highly unatural ( despite being musical!). You can find example easily over the web to have an idea of why we used that name xith mp3 comparison
**: sceptical but not condescending. I've taught all this professionaly to vutur audio engineer at a previous time of my life and observed the same kind of things over and over again. So i anticipate it is the same with you. I might be wrong but... 😉
Mp3 are not like linear format which reproduce at output what is at input: they use destructive algorythm to 'butcher' your source linear file ( in order to have less data in the final output file): it is based on frequency and amplitude present in the recording.
It's based on 'masking effect': when 2 signal are present, let's say a lower frequency one and an high frequency one, depending on their relative level if the lower frequency one is louder it will 'mask' ( in part) the higher one.
So the algorythm makes a constant analysis of the input signal and decimate the 'masked' frequency in your initial signal to lower the volume of data to be encoded.
Then a data compression is applied to save on the data volume to be saved ( but it's not audio dynamic compression).
It is clever approach as it's based on psychoacoustic but it have drawback: it introduce distortion of the initial signal and the harmonic content of treated source is modifyed.
That's the typical 'TR-808 syndrome'* (as we called it) with cymbal encoded in low rates.
If encoded at higher rates it is less and less audible but still present and some can still identify the effect in good condition ( low distortion loudspeaker, low background noise environment, direct instant comparison, etc,etc,).
FLAC are entirely different: they only perform data compression. If only focused on audio they can be considered as linear, but the resulting files are then treated in a way their final data 'weight' is reduced. It's a format compression type of file, audio remains untouched.
Ok,
I believe you when you say you heard difference between different formats. But when you say what you see match what you heard i'm sceptical**.
Let's take your previous example of screenshots of peaks: you compare the same file on different level of zoom, so on different timescale. Of course the end redults will be different despite what you observe is still the same thing!
Let's take an analogy: you take a walk into a forest. At ground level there is a large amount of variation of tree size: there is low height and higher ones. There is even variation in between the higher ones, some are bigger than others and it seems there is a large amount of variability even looking at the same species.
Now let's take a step back at the same forest but on your way home, on a road on a hill some kilometers away from it: now the canopy seems very consitant in term of height, no large variation on it, despite what you saw earlier when walking in it. Some trees ( the smaller ones) even seems to have disapeared...
This is what happen when you 'zoom' in the waveform in audacity: you compare a large time scale ( your landscape view) to a small timescale ( your walk in forest): there is no change in the audio, only the scale you look at it.
I'm pretty sure you did not took notice of this kind of things and made a preset view setting to compare your files at same timescale when comparing your visual representation of your different files. So you probably compared apple to orange.
Other thing, i told you about rms level of your files: have you performed an analysis of each files and then compensated the louder one to meet the lower one?
If not then here again it's apple to orange as the intrinsic level of each file is different from track to track...
I don't know Audacity as i use different tools but you'll have to learn all this very boring things to perform any kind of visual analysis/comparison before performing any kind of meaningful comparison and have conclusion you can rely on.
Gear, gear,gear... well i agree there is a treshold you should not cross about quality but i would not obsess over it either.
About amplifier: a good old chipamp used in it's SOA ( safe operating area/range) was/is more than enough for some very serious about quality brands you'll find in almost every recording/mixing facility around the world. To assess the quality of what you as end listener will evaluate in the end.
And it is the same with classD or whatever.
It's mainly in the application/use and coherency in the chain which is important in my view.
Try to make a coherent/homogenous reproducing chain ( no need for a fancy 'whatever link' in your chain if the other links are not on par) and you'll be happy.
*: TR-808 is a famous Roland 80's beatbox where cymbals (ride) sounds are highly unatural ( despite being musical!). You can find example easily over the web to have an idea of why we used that name xith mp3 comparison
**: sceptical but not condescending. I've taught all this professionaly to vutur audio engineer at a previous time of my life and observed the same kind of things over and over again. So i anticipate it is the same with you. I might be wrong but... 😉
I have done correlations between MP3s and the original WAVs.
A 128 kbps MP3 encoding typically introduces 10-15% distortion.
A 256 kbps MP3 encoding introduces 1-3% distortion.
I am counting as distortion all errors other than a constant time delay. These errors are usually bucketed into frequency response, harmonics, noise, and phase shift.
Ed
A 128 kbps MP3 encoding typically introduces 10-15% distortion.
A 256 kbps MP3 encoding introduces 1-3% distortion.
I am counting as distortion all errors other than a constant time delay. These errors are usually bucketed into frequency response, harmonics, noise, and phase shift.
Ed
Member
Joined 2009
Paid Member
I have to agree.
I already have the cheapest turntable in the world, possibly in the universe.
Maybe a next-level turntable is in order: get something with a cartridge that requires 3-4g tracking. Or wait until the used ones are repaired or
I am simply not confident experimenting with vintage quality turntables, maybe that is Phase 3 of my project.
How do 'good' turntables perform under the strobe disk test? There is a reviewer here who does such a test and I have included a screenshot of the relevant attempted 'clip' from about 9:00 onwards. The strobe markings seem to 'pulse' forwards: is this detectable audibly?
Quartz- locked turntables do not seem to have this problem, in my recollection.
I know there is a thread covering this, quoted above, if there is a section in the thread covering this aspect of turntable speed analysis, I would appreciate it if a link could be provided.
Quartz- locked turntables do not seem to have this problem, in my recollection.
I know there is a thread covering this, quoted above, if there is a section in the thread covering this aspect of turntable speed analysis, I would appreciate it if a link could be provided.
I used Audacity to find peaks in the audio file. The file, by the way, is hosted on the Internet Archive.
FLAC file normalized, no clipping found under "Analyze > Find Clipping"
The file without any normalization, and again, "Find Clipping". I can hear the difference, with adjusted volume. The track is still a little distorted, though, but much better to listen to. This was not a blind test, though that could be arranged or I could splice the track. Output levels were matched manually, I admit, as far as I could manage, listening was on headphones.
Attachments
As i already told: if there isn't clipping in the initial file, the difference you heard are from:
_ the difference in level which aren't matched ( very probably as you did it by ear and manually),
_ something in your reproducing chain which behave differently as you are closer to 0dbfs ( DAC very probably).
For comparison purpose you must rely on RMS level as this is how our brain integrate/interpret level: Audacity should have a analysis tool to know this.
Probably labeled RMS or AVERAGE level.
First perform a test on the initial file it will give a value which will be your target (for your second normalized file).
Normalize your file, redo an average analysis: it'll tell you by how much db the file had been boosted to reach 0dbfs on peaks.
The difference between the two file is equal to the attenuation you'll need to compare both file at same level.
If Audacity is transparent ( which i have no doubt about) then you'll just have prooven it can modify level without impact on the audio.
To verify this you can load both file in a DAW ( multitrack) and reverse polarity on one file, adjust the level on the normalised track ( eg: -1,42db ( minus 1,42db) on the normalized file track's fader), push playback and... have zero level output.
It is known as a null test and one of the best way to see if digital treatment is blameless: if you have sound, it'll tell there was difference induced in the digital treatment.
If you still ear difference when replaying the file then it's either one of the two reason given before, either expectation bias.
And don't dismiss expectation bias it's a very powerful 'feature' of human being! 😉
_ the difference in level which aren't matched ( very probably as you did it by ear and manually),
_ something in your reproducing chain which behave differently as you are closer to 0dbfs ( DAC very probably).
For comparison purpose you must rely on RMS level as this is how our brain integrate/interpret level: Audacity should have a analysis tool to know this.
Probably labeled RMS or AVERAGE level.
First perform a test on the initial file it will give a value which will be your target (for your second normalized file).
Normalize your file, redo an average analysis: it'll tell you by how much db the file had been boosted to reach 0dbfs on peaks.
The difference between the two file is equal to the attenuation you'll need to compare both file at same level.
If Audacity is transparent ( which i have no doubt about) then you'll just have prooven it can modify level without impact on the audio.
To verify this you can load both file in a DAW ( multitrack) and reverse polarity on one file, adjust the level on the normalised track ( eg: -1,42db ( minus 1,42db) on the normalized file track's fader), push playback and... have zero level output.
It is known as a null test and one of the best way to see if digital treatment is blameless: if you have sound, it'll tell there was difference induced in the digital treatment.
If you still ear difference when replaying the file then it's either one of the two reason given before, either expectation bias.
And don't dismiss expectation bias it's a very powerful 'feature' of human being! 😉
Exactly. MP3 is best described as "butchering" as you say. Because that's what it does -- it elliminates parts of the frequency spectrum at every point in time.
No analog vs digital comparison should be done using MP3 files...
I understand you're looking for a good turntable. There is much more to a good turntable than just speed smoothness.
Any really good belt drive, direct drive or idler drive turntable will not have problems with speed smoothness. And no, you don't need quartz lock. Quartz Lock has nothing to do with speed smoothness; QL only helps to achieve the exact same absolute speed value over long hours -- even if the turntable has not so good wow&flutter performance.
And, flutter performance also depends a lot on the tonearm and cartridge used... the big elephant in the room.
If what you're looking is getting the best quality out of vinyl then look for a turntable with a high quality bearing, high quality tonearm, and well designed plinth. Either modern or vintage.
You can get a truly good system with little money if you know how to restore/refurbish vintage turntables. This is DIYaudio after all. For example you can get a broken Thorens TD125 for reasonable money and fix it (the main board is essentially a wien bridge quadrature oscillator). And that one -to put an example- was Thorens' top of the line turntable of the late 60s and one of the best turntables ever made. Just to put an example, there will be hundreds of other desirable models.
I hope, that doesn´t come as a surprise for you😉
This goes for all aspects of life.
It has taken me close to 50 years to get, where I am now, and what I have forund out during these years: There are no shortcut´s.
I thought it was about getting a more than satisfactory sound from a limited budget, which is absolutely possible.
A 2. hand Rega-, Project-, Argon-deck (or one of the millions of japanese DD´s from the eighties) and a AT-VM95/SH cartridge
will get you very close to that 😉
Don´t use FleeBay for turntable. Find one in one of your local charity stores and give it some love, care and adjustment.
I have yet to find figures on how much drag the stylus causes - and how much the platter has to weigh to make those effects very small, unnoticeable, actually.
There are turntables too heavy to be lifted by one person.
I will be trying my hand at restoring a vintage turntable, but apart from that I think I will have to settle for an 'intermediate' model until I have enough knowledge to mess with a real one, not to mention some space to place the thing.
I do not think I have that much time, really which is why the sudden interest in exploring for a lifetime music collection and looking for the best sound I can afford, more in terms of space and time than money.
Good to know. There are no charity stores here.
No, i'm not speaking about stylus drag. At all.
The tonearm mass and cartridge stylus suspension form a resonant system with a resonant frequency and a resonance Q. This resonance is excited by the surface of the record.
The resonance will result on a small (or big) movement in the cantilever, and this movement will cause a negligible, or moderate, or big, or atrocious frequency modulation of the audio signal. This is manifested as FLUTTER as well. This will happen even if the platter mass is extremely, impossibly high.
Thus, the flutter performance will also depend on the tonearm and cartridge used. Which is what I claimed on my previous post.
https://www.diyaudio.com/community/...yl-and-a-decent-turntable.399727/post-7389608
Correction: the files I have been listening to were actually .mp3. The other files from the same site, Internet Archive, were FLAC, and I assumed that these were, too. I wondered why the sound was so awful and distorted: Audacity shows that the track's volume is increased into clipping many many times.
Normalized and mp3 again: much better, but I have already ordered the vinyl from discogs.
You can check the samples version here:
https://archive.org/details/cd_framptons-camel_peter-frampton
Correction: the files I have been listening to were actually .mp3. The other files from the same site, Internet Archive, were FLAC, and I assumed that these were, too. I wondered why the sound was so awful and distorted: Audacity shows that the track's volume is increased into clipping many many times.
Normalized and mp3 again: much better, but I have already ordered the vinyl from discogs.
You can check the samples version here:
https://archive.org/details/cd_framptons-camel_peter-frampton
Given the sheer volume of e-waste, and the fact that someone I knew threw away a turntable because the belt was broken, it seems to make sense to buy an used one. It's not going to stop all waste of resources but there is a difference between throwing one plastic bottle into the ocean and recycling it.
The detail and expansive nature of a proper system looks to be more or less guaranteed with an Japanese vintage turntable. Voltage??
How to convert the audio? Some suggestions have been made: this looks promising. (The other unit suggested earlier was not available)
https://www.whathifi.com/rega/fono-mini-a2d/review
From this article:
https://www.whathifi.com/advice/how-to-digitise-your-vinyl-collection
The detail and expansive nature of a proper system looks to be more or less guaranteed with an Japanese vintage turntable. Voltage??
https://www.google.com/url?sa=t&rct...usg=AOvVaw107KuSMj9Me9V20v3e2pAc&opi=89978449
How to convert the audio? Some suggestions have been made: this looks promising. (The other unit suggested earlier was not available)
https://www.whathifi.com/rega/fono-mini-a2d/review
From this article:
https://www.whathifi.com/advice/how-to-digitise-your-vinyl-collection
Attachments
Last edited:
@BasicHIFI1
Good to see another vinyl newbie from SL...I'm new to the forum, so I can't PM you.
Can share my very short journey getting on the vinyl train with the limited market in Sri Lanka:
1. Turntable - vintage Rotel RT90/new, basic AT cartridge and stylus. It was a lucky find as it's Singapore model, with 230v power.
(bought locally from a known importer of vinyls, amps,speakers and turntables)
2. Speakers ‐ QAcoustics 3030i oversized bookshelf speakers
(yamaha centre)
3. Aiyima tube phono preamp, Aiyima T9 pro DAC w/ tubes (thought I'd try out the Chi-fi route as a start...turned out to be good fun!)
(Amazon + Aramex shipping from US)
All this is MY starter pack, which all came in under LKR 300K/USD 1K...the chi-fi stuff might not gel with some folks ....but I chose to have a bit of (cheap) fun and see if I get more serious later.
Good luck!
Good to see another vinyl newbie from SL...I'm new to the forum, so I can't PM you.
Can share my very short journey getting on the vinyl train with the limited market in Sri Lanka:
1. Turntable - vintage Rotel RT90/new, basic AT cartridge and stylus. It was a lucky find as it's Singapore model, with 230v power.
(bought locally from a known importer of vinyls, amps,speakers and turntables)
2. Speakers ‐ QAcoustics 3030i oversized bookshelf speakers
(yamaha centre)
3. Aiyima tube phono preamp, Aiyima T9 pro DAC w/ tubes (thought I'd try out the Chi-fi route as a start...turned out to be good fun!)
(Amazon + Aramex shipping from US)
All this is MY starter pack, which all came in under LKR 300K/USD 1K...the chi-fi stuff might not gel with some folks ....but I chose to have a bit of (cheap) fun and see if I get more serious later.
Good luck!
Hi. Great to hear from you. Welcome to the forum. You will find much good advice here, and experienced audio enthusiasts who are very helpful with a huge store of knowledge behind them.
You must be aware of the two vintage Hi Fi stores that I posted earlier, I plan to visit them when I visit SL soon, just to check. You probably bought your equipment from one of them. I enjoy experimenting with the extreme low end (cost wise, not audio -wise)
The system I have at home is a Sony ES 1980s vintage system, I got the amplifier serviced at Debug, however there was only so much they could do with it.
I hear chip amps are sufficient to the task, and I have a good set of speakers for the moment.
The Q - Acoustics is a solid looking speaker. You did not mention stands?
https://www.forbes.com/sites/marksp...s-3030i-have-a-big-bass-plus-plenty-of-punch/
The Ayima tube pre-amp - those tubes are decorative I think, according to an YouTube review. Maybe the specs will says something about these.
You did not mention an amplifier?
There are some sellers of vinyl records on ikman.lk and FB, are these good sources? (Large selection and good quality?)
Posted earlier. Could you suggest a range of good old chipamps? Here is what I come up with. I have a PAM 8403 that I need to set up and another one I think LM 3886. Shorting the output will destroy the PAM, so that is a drawback.
Pyle Wireless Bluetooth Home Audio Amplifier - 90W Dual Channel Mini Portable Power Stereo Sound Receiver w/ Speaker Selector, RCA, AUX, LED, 12V Adapter - For iPad, iPhone, PA, Studio Use - PFA330BT
AmazonCorrection: they are functional, 'buffer stage'. These Aiymas can be pretty good, according to reviews.