Originally Posted by Andrew Dee
Recently I’ve become aware that I’ve been uploading my tracks to Soundcloud in mp3 format when I could’ve been using .wav format. I think I did that out of habit and the misunderstanding that file size might be an issue, rather than SC’s ‘minutes uploaded’ limit. I am going to replace my tracks progressively with wav format.

However, it makes me wonder if it makes a difference at all? I’ve had a few comments on the quality of my mixes, but never had anyone tell me they heard the sampling rate was too low. Was it that forum members were just being kind, or is there no discernible difference to the listener? What format do you upload in?

I’ve recently seen a few comments about the quality being better on SC than YouTube, and assuming that’s true, wondered what causes this? Is it the YT processing engine, or is it because we put the audio through video editor *and* YT processing engine - so in effect it’s getting processed twice? I am assuming - that because each platform has it’s own LUFS standard - that each is compressing/limiting to this different standard?

Andrew

Andrew,

I've used 44,100Hz, 16 bit WAV format for over 10 years with Soundcloud. My reasoning was based on the following...

  • With the Audiophile version, BIAB Realtracks and Realdrums are provided as 44,100/16 and by using this information as settings I can use BIAB files in their native states without any audio manipulation need by the computer software.
  • Every time a file is converted from WAV to MP3 or WAV to STREAMING FORMAT, there is the possibility that some audio data is lost or its quality is affected in some way due to the compression process.
  • As a result, uploading 44,100/16 WAV format to Soundcloud means that there is only a single conversion from WAV to S/cloud's streaming format. I reason that this should help keep my audio files as close to original as possible.


With BIAB's UltraPak and PlusPak updates for Windows, the files are 128kHz wma. This means that PG Music has converted the native WAV format to a lossy format (wma). When using the BIAB UltraPak to create a backing track, as I understand it (I may be wrong, and I also have no idea what happens with FLAC files), the wma files are first converted to WAV and the backing track is assembled from these newly converted WAV files (this is most likely a lossless conversion). Then if BIAB renders the backing and it is rendered to mp3, another conversion happens (this is a lossy conversion). Then, if that mp3 is uploaded to Soundcloud, another conversion most likely happens as the file is put into SC's streaming format. These three lossy conversions each have potential to reduce audio quality. How much the audio is affected, though, I have no idea. It's a bit like taking a photocopy of a photocopy, and I have noticed that when a photocopy of a 3rd generation photocopy is taken, sometimes the degradation in visual quality is noticeable. I suspect that there is a similar behaviour with audio.

With the 44.1kHz, this is how the information is 'grabbed' per second. It means that every second, 44,100 samples of information are processed. In other words, the main pitch of a sound and all the harmonics associated with it (i.e. the tone of the sound), are reduced from a continuous audio-wave flow of data (as in analogue processes like playing a musical instrument) to 44,100 samples of digital data each second. With a frequency of 48,000Hz, there are an extra 3,900 samples taken per second so the quality of the digital representation of a continuous sound wave should be around 9% more accurate compared to 44,100 samples per second ({44,100/3,900} * 100).

From what I understand at the moment, up-scaling audio quality (say, wma to WAV) is most likely a lossless process. It's the down-scaling and repeated down-scaling of quality that creates audio issues.

In regard to the bits... These are all about a digital representation of the height of the sound wave (i.e. the volumes of the main pitch and its harmonics). The more bits there are, the better interpreted volume variations within an audio file are.

A 16-bit audio file has a possible 2^16 (2 to the power of 16 = 65,536) unique levels of volume. This corresponds to around 98dB of audio range. This means that data is stored in 'digital blocks' that are ~0.0015dB high. With a 24-bit audio file, there are 2^24 (= 16,777,216) unique volume levels have around 144dB of audio range. Also, the digital block-size for information storage in 24-bit is around 0.0000087dB. This means that each block of 0.0015db (16-bit) contains around 170 blocks of 0.0000087dB. In other words, the ability of 24-bit audio is that it has the potential to store subtler volume-change information (low dB block-size) and a greater dB range of information (high dB block-range). Since instrumental tone is determined by the relative volumes of individual harmonics, and overall instrumental volume is determined by the total sum of main pitch and the harmonics, the ability to reproduce audio more accurately is significant for 24-bit audio.

--Noel

P.S. This is my interpretation of things that I've read over the years. I'm not a sound expert, although I do have limited experience in the physics of sound. The Wikipedia link below is useful to read.
https://en.wikipedia.org/wiki/Audio_bit_depth

Also, these days in Reaper I use BIAB tracks at 44.1kHz/16bit, Synth V vocals (and my own vocals) at 48kHz/24bit. I then render the file to a 44.1/16 WAV for upload to SC. Reaper automatically works with stems of mixed audio frequencies and bit-depths. To my old ears, what Reaper does sounds seamless. My grandchildren can probably hear a difference but I cannot.

MY SONGS...
Audiophile BIAB 2026