Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Completely confused over these Spectrum readouts?! Encoding MP3  (Read 6035 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Completely confused over these Spectrum readouts?! Encoding MP3

Hi,

I am a DJ and I am wanting to rip some CD's to the highest possible quality MP3.  I wanted to do all of this in Audacity as to what I can see it gives the best quality encoding, well from looking at the attached Spectrum graphs anyway?  Basically I can't use Audacity as it seems when I 'multiple export' it just repeats the 'Artist Name' from one of the tracks and copies it over to all the other tracks. I asked Audacity support and apparently you have to do every track separate, doesn't make sense to me lol!

Long story short, I have started using 'Exact Audio Copy' as this copies all the correct metadata over.  I have set it up to encode externally using the LAME encoder as I was under the impression this was the best to use quality wise.  I have EAC setup to High Quality/320kbps.

When running the final copies through SPEK I am concerned that Audacity is giving a better quality copy than EAC.  See the much more definitive line/shelf around 16kHz.  Obviously being a DJ I am running these tracks though big PA sound systems and I want the better quality copy... Can anyone more experienced with this please advise???

Thank you so much,
Steve

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #1
From our rules:

TOS 8. All members that put forth a statement concerning subjective sound quality, must -- to the best of their ability -- provide objective support for their claims.  Acceptable means of support are double blind listening tests (ABX or ABC/HR) demonstrating that the member can discern a difference perceptually, together with a test sample to allow others to reproduce their findings.  Graphs, non-blind listening tests, waveform difference comparisons, and so on, are TOS not acceptable means of providing support.

https://hydrogenaud.io/index.php/topic,3974.html

Put simply, those graphs are worthless.

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #2
I'm pretty sure someone could make a file compression algorithm that makes great looking spectrums at the cost of sound quality... 

MP3 is lossy, but it tries to throw-away stuff you can't hear.

EAC and Audacity both use LAME so you should be able to set them up identically.

If you have a good-sounding original, good-quality MP3s should sound GREAT through a great-sounding system!   But, nobody is forcing you to use lossy compression.

Quote
I wanted to do all of this in Audacity
Audacity doesn't rip CDs.   You could rip to WAV and then convert the tracks individually to MP3 with Audacity (or some other conversion program).

Quote
it just repeats the 'Artist Name' from one of the tracks and copies it over to all the other tracks.
Frequently, I end-up editing some of the metadata with MP3Tag anyway.   (And of course, most CDs do have the same artist on all tracks and it's the only the various-artist CDs that you'd have to change.)

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #3
Quote

EAC and Audacity both use LAME so you should be able to set them up identically.

Yes I am am aware of this, hence my confusion with the different readouts, I was expecting them to be the same, that's why I'm confused.

Quote
Audacity doesn't rip CDs.   You could rip to WAV and then convert the tracks individually to MP3 with Audacity (or some other conversion program).

Sorry my bad, I'm also aware of this.  I meant I am importing Flac or Wav files in to Audacity.  EAC you can encode to MP3 straight from the CD.. maybe this is what is causing the difference in readout and possible quality difference?!

Quote
Frequently, I end-up editing some of the metadata with MP3Tag anyway.   (And of course, most CDs do have the same artist on all tracks and it's the only the various-artist CDs that you'd have to change.

This just isn't practical for me.  They are compilation CD's and have multiple artists.  This is what caused me to try and find an alternative, i.e. Exact Audio Copy.  Purely because it's much quicker for the quantity I need to do and having to edit all the metadata manually would literally add hours to the task.

BUT I'm still predominantly interested in the final audio quality, hence my hesitation.  Just to clarify I am doing this for pub/bar djing and with what equipment I use and my workflow Lossy is my only option.  For other types of DJing I do I only use Lossless formats.

Thanks for your reply,
Steve

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #4
Let me address some of the concerns here.

First of all, both EAC and Audacity can use the LAME, so as long as they're the same version and are configured with exactly the same parameters, the output should be identical.

Second, with Mp3tag you can use freedb, Discogs or MusicBrainz to try and locate your CD and it'll tag the files appropriately for you, quite often with the correct artist for each track.

Third, you say this is for pub/bar DJing ... In this context, do you really think that with all the hubbub and commotion that's common for such social gathering places, people are going to try and stand still to try and pick out lossy compression artefacts? Unless you're using a ridiculously low quality setting, I say it's not likely anyone's going to ever notice anything that's off.

I'm not trying to be facetious or confrontational here. The thing is, with modern lossy encoders at a high enough setting, people are extremely unlikely to be able to successfully ABX them when comparing with the lossless original (barring killer samples and whatnot), and that's in a controlled environment, with good headphones. In a noisy social environment, such as a pub, it's a non-issue, imho
Lossless: flac --best --verify
Lossy: opusenc --bitrate 160

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #5
Quote
Second, with Mp3tag you can use freedb, Discogs or MusicBrainz to try and locate your CD and it'll tag the files appropriately for you, quite often with the correct artist for each track.

Thank you, I wasn't aware it could do that in batches! I will take a look.

Quote
Third, you say this is for pub/bar DJing ... In this context, do you really think that with all the hubbub and commotion that's common for such social gathering places, people are going to try and stand still to try and pick out lossy compression artefacts? Unless you're using a ridiculously low quality setting, I say it's not likely anyone's going to ever notice anything that's off.

Yes that's exactly what I'm saying.. Maybe not in the literal sense but I've had this happen on more than one occasion, where I've downloaded from supposedly reputable sites (DJ download pools etc, which are stated as 320kbps) and when played through a very high quality loud system in a bar with a dancefloor full of people, it's sounded awful and muffled and this has really killed the mood.  People know something doesn't sound quite right and it really kills the atmosphere.  (that's from real life experience) It's horrible and embarrassing.

Quote
I'm not trying to be facetious or confrontational here. The thing is, with modern lossy encoders at a high enough setting, people are extremely unlikely to be able to successfully ABX them when comparing with the lossless original (barring killer samples and whatnot), and that's in a controlled environment, with good headphones. In a noisy social environment, such as a pub, it's a non-issue, imho

Of course and I completely agree with you!  This post was purely to do with me being a perfectionist/audiophille and obviously if I've got two options in front of me I want to pick the best.  I just wasn't sure whether I was actually getting a very noticeable difference in quality as there is an obvious difference in the graphs.  It just confused me a little.  It may well be to do with the settings and I will continue to play with these.  Through headphones no, maybe I can't tell the difference but if there is, I don't want the poorer quality track for my professional work.

Cheers

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #6
Picking the one with highest frequency response may end up being audibly inferior.

It could be that data is being wasted encoding things you are less likely to hear in favor over things you are more likely to hear.  Furthermore, it is often the case that higher bands of frequencies being encoded require more data than lower bands of frequencies.  When this isn't the case, it is usually because the higher bands frequencies are being synthesized rather than being encoded in order to save on data.

Make sense?

PS: Don't get sucked into the nonsense that stereo encoding is better than joint-stereo encoding.  This is practically never the case; properly implemented joint-stereo is practically always better than stereo encoding for music.

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #7
Quote
Second, with Mp3tag you can use freedb, Discogs or MusicBrainz to try and locate your CD and it'll tag the files appropriately for you, quite often with the correct artist for each track.

Thank you, I wasn't aware it could do that in batches! I will take a look.

Note that for some of these you may also have to register on their website first.

Quote
Third, you say this is for pub/bar DJing ... In this context, do you really think that with all the hubbub and commotion that's common for such social gathering places, people are going to try and stand still to try and pick out lossy compression artefacts? Unless you're using a ridiculously low quality setting, I say it's not likely anyone's going to ever notice anything that's off.

Yes that's exactly what I'm saying.. Maybe not in the literal sense but I've had this happen on more than one occasion, where I've downloaded from supposedly reputable sites (DJ download pools etc, which are stated as 320kbps) and when played through a very high quality loud system in a bar with a dancefloor full of people, it's sounded awful and muffled and this has really killed the mood.  People know something doesn't sound quite right and it really kills the atmosphere.  (that's from real life experience) It's horrible and embarrassing.

All I can think of is, the mastering of the original may have been wonky, or these MP3s were either (accidental, since I'm not attempting to point fingers or anything) reencodes from a lossy source, or used an older, less efficient and fine tuned encoder. As an example, and you can search around these forums for this, the Bandcamp website has provided, on occasion, "lossless" files that almost certainly come from a lossy source, and a user on H-A asked them and they said that they don't force artists to send them genuine lossless files, just a lossless format whose source could very well be lossy.

Which brings me to my point: if you know the source of the MP3, which you seemingly do as you have the original retail CDs, and use a modern encoder with a high enough setting, you should be in the clear.

Quote
I'm not trying to be facetious or confrontational here. The thing is, with modern lossy encoders at a high enough setting, people are extremely unlikely to be able to successfully ABX them when comparing with the lossless original (barring killer samples and whatnot), and that's in a controlled environment, with good headphones. In a noisy social environment, such as a pub, it's a non-issue, imho

Of course and I completely agree with you!  This post was purely to do with me being a perfectionist/audiophille and obviously if I've got two options in front of me I want to pick the best.  I just wasn't sure whether I was actually getting a very noticeable difference in quality as there is an obvious difference in the graphs.  It just confused me a little.  It may well be to do with the settings and I will continue to play with these.  Through headphones no, maybe I can't tell the difference but if there is, I don't want the poorer quality track for my professional work.

Cheers

What you can do is, use both EAC and Audacity, ignore the spectrograms (because as it was pointed out to you, they reveal nothing, and as a matter of fact can be misleading since it could be possible to produce a spectrogram that looks more like the original but audibly sounds worse), and use a double-blind test, comparing the generated MP3s with the lossless original. If they sound indistinguishable to you on a setup involving a quiet room with a pair of quality headphones, they're going to sound just fine out there in the pub.
Lossless: flac --best --verify
Lossy: opusenc --bitrate 160

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #8
I would rip CDs to lossless (do that only once!) and then encode MP3s from those files.

For DJing you would be surprised at how low in bitrate you can go if you want to, but then: is space really an issue? Why are you using MP3 in the first place? Is it that it is the only tag-able format supported by a certain application?


Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #9
Thank you everyone for your advice, appreciate it!

I would rip CDs to lossless (do that only once!) and then encode MP3s from those files.

Does this make a difference??  What I like about EAC is I can encode straight from CD, well it converts to WAV and then encodes using LAME on the fly.  Again this is obviously a big time saver!

For DJing you would be surprised at how low in bitrate you can go if you want to, but then: is space really an issue? Why are you using MP3 in the first place? Is it that it is the only tag-able format supported by a certain application?

Yes software, external players, compatibility and having 1000's of tracks, lossless just isn't practical.  Like I mentioned I DO use strictly only Lossless, either WAV or AIFF for my 'proper' DJing, where I would only have around 20 tracks ready on a memory stick!  (where I'm turning up to an event with high quality equipment that has been specifically set up with high quality sound in mind and people who know music and are expecting a great sound and musical experience) very different to my more regular djing playing 'chart favourites!'  

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #10
I would rip CDs to lossless (do that only once!) and then encode MP3s from those files.

Does this make a difference??  What I like about EAC is I can encode straight from CD, well it converts to WAV and then encodes using LAME on the fly.  Again this is obviously a big time saver!

Well you purport to be a DJ right? Maybe you want to do some remixing of tracks at some point. Surely you'd only want to do that with the original lossless copies. Always good to keep them at hand: Get a good program that lets you encode lots of tracks in batches (I like foobar2000), properly sort and tag all your FLACs and just encode stuff on the fly when necessary. IMHO 320 kbit/s MP3s are a waste... V0 should be more than enough for your uses.

Quote
1000's of tracks, lossless just isn't practical

Eh, memory sticks have huge capacity nowadays... 1000's of FLACs hardly seems impossible. But sure, MP3s are probably most compatible.

Quote
with high quality sound in mind and people who know music and are expecting a great sound

There will probably be no audible difference in such venues as long as you use lossless sources for your MP3s, sensible encoder settings for LAME and quality better than V5.

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #11
Yes software, external players, compatibility and having 1000's of tracks, lossless just isn't practical.

1TB+ 2.5" external drives are not expensive and very easy to carry around since they don't require any power supplies. I have 1000s of lossless tracks on a 1TB drive, 2-4TB drives are available too :)

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #12
[...] Like I mentioned I DO use strictly only Lossless, either WAV or AIFF for my 'proper' DJing, [...]

A question, just out of curiosity ... Is there a reason you use WAV or AIFF? These are uncompressed lossless files; in case your equipment supports it, wouldn't FLAC or ALAC be better? These are compressed lossless files (hence, take up less space) that additionally have better metadata and tagging support.

I did see you mention FLAC in one of your previous posts, but I wanted to make sure you understood that a lossless file will contain the exact same audio information, bit for bit, as the source (in this case, your CD), since there's seemingly a lot of misinformation about this topic on other audio forums.
Lossless: flac --best --verify
Lossy: opusenc --bitrate 160

 

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #13
Quote
Yes that's exactly what I'm saying.. Maybe not in the literal sense but I've had this happen on more than one occasion, where I've downloaded from supposedly reputable sites (DJ download pools etc, which are stated as 320kbps) and when played through a very high quality loud system in a bar with a dancefloor full of people, it's sounded awful and muffled and this has really killed the mood.  People know something doesn't sound quite right and it really kills the atmosphere.  (that's from real life experience) It's horrible and embarrassing.
"Unknown" recordings are a completely different story.    The poor quality could be low-quality MP3 compression or it could be 1000 other things.   And since you downloaded them, I'm guessing you don't have the original CDs to compare to so maybe it's just a lousy recording.  ;)

Quote
Yes software, external players, compatibility and having 1000's of tracks, lossless just isn't practical.  Like I mentioned I DO use strictly only Lossless, either WAV or AIFF for my 'proper' DJing, where I would only have around 20 tracks ready on a memory stick!
Just doing some rough approximations -
160 FLAC (or ALAC) files on a 4GB thumb drive
640 FLACs on a 16GB thumb drive
40,000 FLACs on a 1TB hard drive

...I've got about 18,000 'V0' MP3s taking-up 140GB on my laptops.    I've got three 4TB USB drives with copies all of the MP3s, plus about 2TB of concerts & movies.     (I should have made a lossless music archive, but I didn't have the disk space when I started-out so the original CDs are my only lossless archive/backup. )   


 

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #14
I would rip CDs to lossless (do that only once!) and then encode MP3s from those files.

Does this make a difference??  What I like about EAC is I can encode straight from CD, well it converts to WAV and then encodes using LAME on the fly.  Again this is obviously a big time saver!

The "difference" it makes is that then you have the lossless file and can produce whatever lossy format from the file next time, rather than rip it again. You say you use only lossless for your 'proper' DJing, so why not have the lossless files? Having to re-rip the same CD does not sound like a "big time saver".
A format like FLAC has other advantages, like being checksummed (and thus can be verified for file errors - WAV/AIFF/ALAC-in-MP4 do not have a checksum!) and taking up less space. In my collection, 10 GB stores 24 hrs of music (CD -> FLAC). A 1 TB drive takes 100 days and nights of music. And that is predominantly metal, that takes more than average space.

(On the dancefloor, people would not hear the difference between lossless and MP3s@320 or even a third of that - provided your software guards against digital clipping. Don't confuse bad music for bad format!)

Re: Completely confused over these Spectrum readouts?! Encoding MP3

Reply #15
STEVE16, i think i understand u´re talking about...

As i can see in your grafics, EAC set your mp3, by default, with these command line: -m s -V 4 -q 0 -lowpass 20.5
And Audacity uses these other commands: -m s -V 4 -q 9.

Where -ms (stereo) could be -mj (join stereo) and -q 9 could be -q 7 for example... it´s no matter, all of that are highers n: values.

Just use mediainfo to read the .mp3 at the last line and see those commands.

Proof these command line: "-m s -V 4 -q 9", CBR @ 44.1 KHz and tell me what u listen to... Do not use the LPF command line: - lowpass 20.5. It is not necesary using 48 KHz when ripping CDs, all beyound 44.1 KHz is BLACK or no sound.

U should to listen the highest mp3 possible sound quality.