10

I'm new to the music scenario and EDM production. With so many options of FREE samples out there, I'm wondering how to best check the quality of a sample? Bit? Khz? Wav? Are there any software that I can use to analyse the file? If yes, what should I looking for?

Cheers.

New contributor
OldCastle is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
  • 9
    The only way is by listening to it. If it sounds good, it’s good. – Todd Wilcox Jul 12 at 8:56
14

The only thing that matters is if the sample sounds good. You have to listen to it and decide whether you want to use it. No numbers or waveforms or analysis programs are going to tell you whether the sound is good.
But be aware a sample that sounds good by itself might not sound good in a mix, and other sounds that work well in context might sound bad when you listen to them alone.

Obviously if you have the choice of the same sound at different bit-rates, you will want to take the higher bit-rate. And if the sample is compressed 'lossless' formats (.flac etc.) are likely to sound better than 'lossy' formats like .mp3

  • Poster might want to distinguish compressed 'lossy' formats and compressed 'lossless' formats. WAV files can be compressed to FLACs without altering the audio data, for instance. – Flying Soda 2 days ago
  • @FlyingSoda Thank-you for pointing that out. I've tried to make it clearer. – PiedPiper 2 days ago
  • As I mention in my answer, mp3 is also not a good format for e.g. drum loops as the file durations are unlikely to match the musical loop length. – topo morto 2 days ago
  • @topomorto If you want to loop mp3 files, it's trivial to extract them to .wav – PiedPiper 2 days ago
  • @PiedPiper Sure, but won't that generally still give you a wav file the length of the original mp3 (and not necessarily the actual length of the audio you want looped?) Of course you can then do the work to find the loop end point again, truncate and fix the problem - but the problem wouldn't have been there in the first place if the audio were stored as .wav, .flac, or even .ogg. – topo morto 2 days ago
6

While the ultimate indicator of quality is what sounds best. If you're looking for the "cleanest" sound, there are a few things you should look for.

Typically, you want to find audio files in unaltered formats - though it comes at the cost of dramatically larger filesizes. These are audio formats (both compressed and uncompressed) that store the audio data without changing the content inside. Common filetypes are '.wav', '.aiff', '.m4a', and '.flac'. "Lossy" files are those that contain audio that has been altered to be more easily compressed into a small filesize. Common filenames are '.mp3', '.aac', '.m4a', '.wma', and '.ogg'.

'.m4a' is a bit special because both lossy and lossless files can be encoded in this format, so it's a gamble whether or not you'll get unaltered files.

More information on filetypes here: Audio file format.

Lossy encoders generally reduce the complexity of the sound, and thus - technically - the quality, though it does this in hard to discern ways. Most use variations of a psychoacoustic masking trick that removes certain pieces of audio without the listener noticing. Quiet frequencies in audio that are near loud frequencies can be cut out, since they're sounds that are typically very difficult to hear normally. They will also cut out quiet transients (transients are loud volume peaks in an audio track, for instance the initial snap of a snare) that are nearby loud transients, since most people will only hear the the loud transient even if the quiet one came before it.

This effects some major aspects of the sound. Lossy encoders will muddy the stereo image, as the phase and frequencies of audio have been altered. This means that altered stereo orchestral recordings, for instance, can sound flat or hollow. You'll also get some noticeable artifacts for noisy sounds, like cymbals. You may have noticed that in compressed audio, the cymbals might have become more hissy than before. You may also lose some low frequency definition, especially sharp bass attacks, since bass frequencies can be hard to differentiate from transients.

There's a much more in-depth discussion of the effects of lossy encoding here: What Data Compression Does To Your Music.

Frankly, most people don't care about the quality of the samples that you use. And, if you put any amount of reverb on the sound, Few people will ever notice.

New contributor
Flying Soda is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
  • Although this is a pretty good answer, and actually applicable since many samples found on the internet are in the truely bad 128kBit/s mp3 format, it has still aspects that are FUD. As I've argued before here, there is no such thing as lossless audio since audio is in principle an analogue phenomenon, and the only meaningful distinction is “does it sound notably different from the original”? Specifically, the transient behaviour even of a “perfect lossless” system is not actually perfect; the filter in the DAC smears transients out in a quite similar way to gentle lossy encoding. – leftaroundabout 2 days ago
  • While I do agree with your sentiment that 'lossless' is not in the most literal term 'lossless', it is one of the terms used by most circles to describe the kind of work that type of encoder does upon a file. Most people will have an opinion on the coloration from the mic'ing and the minimal effects of the DAC through listening alone, so I felt it was prudent to only explain things which are easily missed - in fact purposefully designed to be missed. – Flying Soda 2 days ago
  • 3
    @leftaroundabout I always thought lossless was generally used to mean "compared to the original stream of sampled values" (i.e. post sampling) rather than "compared to the analogue original" (when you'd be comparing apples with oranges...) – topo morto 2 days ago
  • @topomorto yes, that's exactly what it means correctly, but when people argue why “lossy is bad” this is usually not what they mean. – leftaroundabout 2 days ago
5

While everyone is rightfully emphasizing that the sound is, at the end of the day, the only important metric, there are still a few basic technical things to look out for:

  • Lossy vs. lossless: given the choice, you always want lossless WAV, FLAC, AIFF, etc. (See @Flying Soda's answer for more details.)
  • Stereo vs. mono: Some things, like kicks, snares, or many instrument samples, make perfect sense in mono. But with more atmospheric things like crowd or traffic noise, risers, shaped white noise, and pads, you want the added complexity and depth of stereo. (Unless you have a compelling reason otherwise.)
  • Clipping: Audio can clip (go beyond the "maximum value" and descend into harsh high-frequency distortion). Unless you want this for stylistic reasons, always avoid clipped samples.
  • Bit depth: 24-bit is technically superior, but the differences are small. 16-bit audio will always be good enough until you are much, much deeper in the audio world (as in professional). It matters much more when recording (making gain staging simpler) than in choosing samples.

Everything is subject to artistic judgement, and one can always make an artistic choice against a technical consideration. But you have to be aware of the technical details first.

New contributor
Linuxios is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
4

There are no real indicators like that. Of course, in general you could say for example, that .wav files are higher quality than .mp3 or that during the production phase 24 bit may be better than 16 bit etc., but who actually tells you, that this is the original file?

If someone downloaded a 240p youtube video and converted the terrible audio file into .wav and changed it to 24 bit for example... Does that make the sound quality of the sample all of a sudden great again? For sure not...

So actually, the only way to know if a sample sounds good is to listen to it. Trust your ears!

Usually you can already tell it by just listening to a single sample... But even if not - If you have already a project with a lot of good samples and throw in a sample that isn't really high in quality, you will probably notice it once you imported it into your project and 'compare' or hear it together it with the other sounds.

4

Other answers have given some information about the technical considerations of the audio format, and pointed out that one primary consideration is of course "is this the sound I want?".

There are also some other considerations that affect how easy it will be to get a good quality result:

Loopability of the whole file: If the sample represents a sound like a rhythmic loop or a sustaining pad, does the full length of the file represent a well-cut section that loops seamlessly? If not, is there at least enough sound in the file to select a smoothly-looping section? Note that mp3 files are poor choices for looping, as they are made up of 'blocks' of time which are unlikely to add up to exactly the loop length you need (see https://sound.stackexchange.com/questions/8916/mp3-gapless-looping-help).

Usability of loop points: Perhaps there are loop points defined in metadata within the file, such as within the SMPL chunk in a wav file. Can your software understand these? Again, do they seem well-chosen and represent smooth loops?

Availability of variations: Often it's desirable to be able to use one or more variations of a sound in a piece - e.g. you may want busier and sparser versions of a drum loop.

Multisampling: If you are looking for an instrumental multisample, does the candidate you've found cover all the playing techniques you need? Does it have samples for each note on the keyboard, or does it use one sample to cover a range of notes? Again, are the loop points well-chosen, if looping is used?


Another thing that you should check (that is admittedly nothing to do with quality) is the licencing terms for the samples. Just because a sound can be downloaded for free doesn't necessarily mean that it's licensed for use in your own music. In some cases, the person distributing the sound may not even have the legal right to do so! Just something to be aware of.

1

From a similar question of mine on SuperUser: How to check the sound quality of an audio recording.

I personally use Audacity and look to see if the sound is clipped after ~20 kHz, but if you are looking for other measures, i.e. mp3 compression artifacts, audacity can help you find those as well. You might also want to Google for detailed information regarding spectral analysis of mp3 files.

-1

If we're talking about the quality of the actual wav/mp3 format, the highest standard is 320kbps as far as i am aware. I use samples in my production with thousands of sounds (kicks,snares,percs,cymbals) and some sample kits are far superior in quality than others but in actuality, it doesn't matter about the quality. Why? Well, sound layering is a topic you should watch on youtube and that gives you complete control over the quality. Also when you're mixing your track, using eq among other mixing tools gives you also complete control over the quality. I've been producing for 7 years and the greatest and high quality EDM drums in the market are by "Vengeance". Check their samples online and if you know how to navigate the net you may find them much cheaper than one could believe..

Your Answer

OldCastle is a new contributor. Be nice, and check out our Code of Conduct.

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.