Hello all, I am ripping vinyl records into FLAC and I would like to have the vinyl records recorded on 24 bits/96khz. I have installed audacity with ASIO support and I have an external soundcard (Focusrite Scarlett 2i2, latest drivers installed) which supports 24/96 samples and which I have connected to my computer via USB. I have set up audacity to use this soundcard as the recording device.
Then I use an application called ClickRepair in order to repair the vinyl clicks. This application shows the recording as being 16bits.
Now I am not sure if I have recorded the sound on 24bits but ClickRepair reports it as 16 bits or I have indeed recorded the audio on 16 bits. Is there a way to tell if it’s actually 16 or 24 bits? And if it is, how can I set up Audacity to use 24 bits?
Thank you
I am using Audacity 2.0.3, Windows 7 Professional with SP 1 and I got the executable Audacity with ASIO support.
Hello kozikowski, thank you for your prompt reply.
You are right, Audacity is not pre-baked with ASIO, I downloaded and installed it (Audacity+ASIO) from an external source (not from the official Audacity site).
Originally I installed ASIO because I was following a tutorial on how to rip vinyl records and I wanted to follow it to the letter. So I did everything by the book.
But if you say I don’t need the ASIO support I will go with you. Do you know what settings I need to set in order to get the result I need?
I downloaded the MediaInfo and I have run a test. I recorded three different WAVs by setting a different host interface from the device tab. One with ASIO, one with MME and one with Window Direct Sound (for all these three I selected the Focusrite Scarlett as the recording device). I opened these files with MediaInfo and is shows the stream as 16 bits for all of them. Any idea what I need to do in order to get the 24 bits stream?
Please note that distributing Audacity with ASIO breaches the license terms of Audacity, or ASIO, or both.
We are unable to offer technical support for modified versions of Audacity.
Regarding Audacity + ASIO, I know you can’t offer support and I have read that Audacity cannot be distributed with ASIO. But I had to download it somehow since I couldn’t follow all the 7 steps (described here: http://wiki.audacityteam.org/index.php?title=ASIO_Audio_Interface) and compile everything from scratch.
Regarding the Save As… Other Uncompressed format… I will try this. As a side question, can this uncompressed format be anything I want even though a soundcard might not support it? For example I know that the Scarlett supports up to 24bits stream. If I record it can I save it as 32 bits stream without loosing any information? Sorry if this question is trivial but I’d like to understand how this ting works.
Audacity recordings are unfortunately always 16 bit.
Wasapi and Wdm-ks are the exceptions.
Make sure that 24 bit is enabled via Windows control panel.
Asio might work too, I haven’t tested that.
Here’s the bit in the manual specifically about “other uncompressed formats”: Audacity Manual
The “options” provide a wide range of settings from 4 bit (ADPCM) up to 64 bit float format.
The general rule is that going from a high bit format to a lower one will lose some information, but going from a low bit format to a higher one will not. It’s a bit like measuring the length of a piece of string in cm or mm (where millimetres provide greater precision). Converting 16 bit to 24 or 32 bit will pad the data with zeros, so the actual sample values remain exactly the same, just expressed in a different way.
Audacity works internally in 32 bit float format, so it is recommended to use that as the default setting (Edit > Preferences > Quality).
If you then export in 32 bit float format, that guarantees that the exported audio is identical to the audio in the Audacity project. However, 32 bit float format files require 4 times the disk space of 16 bit (CD quality) files, and few applications (other than “pro audio” software) support 32 bit float format.
Robert thank you for your answer. This is the exact reason why I wanted the ASIO, since the official Focusrite states
Is my interface in 16 bit, or 24 bit on Windows?
This article addresses concerns raised by some about 16 bit and 24 bit in Windows systems.
When using the WDM driver, only 16 bit modes are seen as an option in the Advanced section of the interface’s Properties. However, this setting has absolutely no effect when recording using the ASIO driver, which users should always use in DAWs, due it’s its low latency and increased peformance over WDM (Windows’ own audio driver).
The audio stream captured from our USB 2.0 hardware is always 24-bit samples. These samples are not altered or processed in any way by the driver when recording via the ASIO driver. Windows audio control panel may say the device is in 16 bit mode, but the ASIO driver bypasses the Windows audio system, and is therefore not affected by this setting.
Therefore, using 24 bit depth in your recording software, while using the ASIO driver, is still 24 bit when recording/creating, where 24 bit actually matters.
From the above statement I understand that by using the ASIO with the Focusrite soundcard, and if I set the correct sample rate in the Quality tab, then my audio is always 24 bits. Can someone confirm that this is correct?
Thank you for the clarification Steve, now I got it. By the way exporting the file as other uncompressed format worked, thank you very much. Another quick question though. I have set the Quality to 32 bits float but my soundcard can generate only 24 bits (I guess Audacity pads the remaining left-hand bits with zeros?) So sound-wise I don’t actually gain anything from having set the quality to 32 bits, since my soundcard cannot send 32 bits information to Audacity anyway. Is my thinking correct?
Thank you again for your valuable info.
Regards
Panos
If I export it to 24 bits then I loose information (since Audacity removes a few bits and therefore removes some information). But since my soundcard is able to generate 24 bits stream then it
That part is largely irrelevant for Audacity.
Audacity is not a DAW, and does not do real time processing or virtual instruments, so low latency has no benefits for Audacity. There may be a slight advantage when recording projects that have a very large number of tracks, but in most cases the performance of WDM drivers should be more than up to the job.
That point is probably a bit misleading. Although the Focusrite can use 24 bit format, it the bottom few bits will be garbage (noise). The specification for the Focusrite gives an analog dynamic range of around 100 dB (depending on the input used). That means that the bottom 7 bits will essentially be noise. In other words, the best actual recording quality you can expect to get in ideal conditions and perfect set-up is around 17 bit. When recording “real world” sources, such as instruments and vocals, background noise (including self noise from microphones, instruments, etc) is likely to reduce the usable dynamic range to less than 14 bit. The bottom line is that if you are careful about your recording levels, you will be able to capture the full quality of the sound in 16 bit. 24 bit recording allows a bit more space for “sloppy” settings (more headroom), and in the case of top level pro equipment, allows a lot more headroom.
If you use ASIO on Windows and you set Audacity to record in 24 bit or 32 bit float (the latter is recommended), then Audacity should capture the exact audio stream. However, we cannot guarantee what a modified version of Audacity will do as we have no way of knowing how it has been modified - we do know that there are modified versions of Audacity circulated on the Internet that contain viruses and other malware, and there is nothing we can do about that other than strongly recommend that people use the official version. In other words, using a version that has been modified by someone else is entirely at your own risk.
I have set the Quality to 32 bits float but my soundcard can generate only 24 bits (I guess Audacity pads the remaining left-hand bits with zeros?) So sound-wise I don’t actually gain anything from having set the quality to 32 bits, since my soundcard cannot send 32 bits information to Audacity anyway. Is my thinking correct?
First, don’t worry too much about the bit depth. The BIG limitation is the vinyl record itself, which because of the background noise has less than 16-bits of usable resolution.
There’s no harm in using 24/96 but it’s total overkill for analog vinyl. 24/96 is the current pro studio standard, but the people who have done blind listening tests will tell you that you can’t hear the difference between 24/96 and "“CD quality” 16/44.1 (in a proper level-matched blind listening test). Basically, it turns-out that 16/44.1 is better than human hearing. Obviously, vinyl is worse than human hearing.
Your Scarlett interface should have come with Abelton Live. I assume Abelton works with ASIO drivers.
24-bit DACs & ADCs typically have about 20-bits of usable resolution, and they are integer devices limited to 0dB maximum.
32-bit floating-point makes digital signal processing “easier”, and it allows for signals over 0dB, or very-low signal levels (below the least-significant bit with 16 or 24-bit formats). That means you can greatly reduce the volume, and then boost it back up (digitally) without losing resolution. Or you can do something like boost the bass and push the peaks over 0dB without clipping. If you do that, you’ll need to normalize the peaks to 0dB or less before rendering your final file, but there is no permanent damage from going over 0dB temporarily with floating-point.
DVDdoug thank you for your detailed analysis. Can you elaborate on the following?
First, don’t worry too much about the bit depth. The BIG limitation is the vinyl record itself, which because of the background noise has less than 16-bits of usable resolution.
There’s no harm in using 24/96 but it’s total overkill for analog vinyl. > > 24/96 is the current pro studio standard, but the people who have done blind listening tests will tell you that you can’t hear the difference between 24/96 and "“CD quality” 16/44.1 (in a proper level-matched blind listening test). Basically, it turns-out that 16/44.1 is better than human hearing. Obviously, vinyl is worse than human hearing.
What is the background noise? What is the human hearing (I guess the level most humans can hear the sounds)? And why CD is better and vinyl is worse than the human hearing?
In my opinion we should not even be answering your questions. You are using an outdated and illegal version of Audacity.
If you compile Audacity yourself with ASIO support and do not distribute the build, then it is perfectly legal.
Alternatively if you want 24-bit recording and think you can hear the difference, you can use the Nightly Builds for Windows which include support for the full WASAPI and WDM-KS hosts which both support 24-bit recording. WDM-KS will probably have fewer bugs than WASAPI. There is a small chance your inbuilt sound device might not support WDM-KS properly and that the computer could freeze or crash when launching Audacity. Most people won’t see that.
See http://wiki.audacityteam.org/wiki/Nightly_Builds. If you get the r13376 build, that is identical to 2.0.6 release except that it includes full WASAPI and WDM-KS, so should be stable (assuming your built-in sound device can handle WDM-KS).
In case this is of interest to some, I just installed the latest Audacity alpha build (“audacity-win-r13510-2.0.7-alpha-27-oct-14.zip”).
I disabled “stereo mix” in the Windows Sound panel.
I was then able to record streaming audio from online radio (http://tunein.com/radio/Abacusfm-Mozart-Piano-s124549/?streamid=26177830) through Windows WDM-KS in Audacity. This is on Windows 8.1 with all updates and integrated Realtek High Definition Audio.
Strictly, it doesn’t mitigate it (much). Even if he only gave the download URL to you and required a password that no-one else has, someone could intercept the information he gave you or could find and hack into his server.
To be “legal” he should have just helped you compile it.
I have a question about that. Why does Export action keeps saving the audio on a 16 bits basis from a source file recorded on 24 bits while the default is yet set at 32-bit float ? Why Export does not use the default settings from Preferences ?
It’s usually better to post a new question when the old post is soooo old!
When you export, after you select WAV you should see a “Encoding” drop-down box where you can select the bit depth.
The 32-bit float option is for internal/temporary processing. …There are advantages to processing in floating-point no matter the original format or the format you export-to.
24/96 is the current pro studio standard, but the people who have done blind listening tests will tell you that you can’t hear the difference between 24/96 and "“CD quality” 16/44.1 (in a proper level-matched blind listening test). Basically, it turns-out that 16/44.1 is better than human hearing. Obviously, vinyl is worse than human hearing.
You are using an outdated and illegal version of Audacity.