I’m in a FB group related to voice overs, and the moderator of that group said he could hear a difference between recordings in Audacity and Adobe Audition. His statement was something like “the difference is subtle but it is there…” (and he was comparing Audacity with Audition and other paid programs).
My gut was either he’s not comparing apples to apples (different sampling rates or bit depth?) or he simply hasn’t done any sort of double blind testing to see if there is really a difference. I suspect since he’s invested $XX in another program, he thinks he hears a difference (something better).
Alternately the Audition playback could be slightly louder, as volume differences do create a “preference bias”.
I can understand if you apply different effects, plugins or processing, then that’s a different thing. I’m thinking of the raw recording.
Any thoughts on potential differences between recording. While I was thinking there would NOT be a difference, I could be missing something and I’m guessing there are plenty of people here who can provide some decent perspective.
PS: I’m assuming same Windows machine is recording, same mic/interface… just switching between the Audacity and something else.
Basically the differences come down to hardware, drivers and processing (what you do to the audio).
Assuming a level playing field where these are all the same, there will be no difference.
Where there can be a very (very very very) small difference, is with a 24 bit sound card on Windows with Audacity as shipped. Due to licensing restrictions, Audacity cannot be shipped with ASIO support, so Windows users must use standard Windows drivers, or build Audacity themselves from the source code and enable ASIO support (requires registering with Steinberg that own the rights to ASIO). Unfortunately, due to a limitation in one of the libraries (components) that Audacity uses to access the sound system, recording using the Windows drivers is limited to 16 bit. What this means in practice:
In the above scenario (24 bit sound card + Windows drivers), if the recording level is set to peak at around -6 dB (recommended), then the digital noise floor is around 90 dB below the peak level. If recording in other software, or with ASIO, or on Mac or Linux, with 24 bit support, then assuming a peak level of -6 dB, the digital noise floor is much lower (around -130 dB). The latter case is theoretically much better, but considering that in a good home studio with good equipment, the noise floor due to the microphone + ambient noise + pre-amp noise + RF interference +… (any other imperfections) the actual noise floor is likely to be at least 20 dB higher than the 16 bit digital noise floor. the actual difference is similar to listening to someone breathing while standing next to a busy highway.
To illustrate this difference, here is a sample of pink noise with a peak level of around -40 dB. At some point in the track, I’ve added in additional noise at -60 dB. Question: where in the track have I added the additional noise?
I suspect that you are right.
By the way, I’m not suggesting that 24 bit hardware is a waste. It is very handy, and will be supported in Audacity on Windows as soon as possible. The big advantage of 24 bit audio is that you can be much more generous with allowing headroom, which can be very useful in live recording situations, especially when under time pressures. When limited to 16 bit you need to be more careful to get a good recording level (ideally around -6 dB). It is essential that the recording level never hits the top/bottom of the track (0 dB) during recording as that will almost certainly result in some distortion.
THANKS for the details! Very helpful. And I can’t hear any difference in the pink noise demo you provided, I can’t find the point it changes. And love the details on the 24 bit stuff, that does provide great context.
His statement was something like “the difference is subtle but it is there…”
I’m pretty sure he could’t tell the which is which in blind test. Tell him you have switched to Audition and he’ll tell you your recordings sound better!
Whats REALLY IMPORTANT is the performance, the quality of the voice or instrument, acoustics and soundproofing, microphone selection and microphone placemment.
Once you get a digital data stream it doesn’t matter what software you use to “capture” it.
You run-across this kind of thing in the “audiophile community” all the time… Someone will claim that changing to expensive speaker cables “lifted a veil”, or something like that. Then you do a blind test and they can’t hear a difference… Then they start telling you why your blind test was invalid, or they say that all blind listening tests are flawed and this kind of science just doesn’t apply to audio.
There’s a fair amount of nonsense in the pro audio commuity too. They are not as bad as the audiophiles, but they will often tell you that 16-bits is not as good as 24-bits although nobody can hear a difference in a proper blind test. And, they will talk about how terrible MP3 is, when in-fact it’s often very-hard (or impossible) to hear the difference between a high-quality MP3 and the uncompressed original (in a proper blind ABX test).
Of course, if you start using EQ, compression, and noise reduction there will be differences and Audition may do a better job. And, some high-end VST plug-ins may not run in Audacity.
For what it’s worth, there is one exception to this - the Windows WASAPI host in Audacity’s Device Toolbar can record 24-bit audio, but in the current 2.0.5 release of Audacity, WASAPI is limited to recording of computer playback. More work needs to be done before WASAPI can be released for other input sources.