Support - Sonnox Fraunhofer Pro-Codec FAQ's
Sonnox Fraunhofer Pro-Codec : Frequently Asked QuestionsThe Sonnox Fraunhofer Pro-Codec plug-in provides a wealth of innovative features and new possibilities for mix and mastering engineers alike. Many of the questions we're often asked are answered below. However, if your question is not listed here or you need further help, please contact us at support@sonnoxplugins.com
FAQ's What is the purpose of the Pro-Codec? Which DAW's and operating systems are the Pro-Codec compatible with? Before I use the Pro-Codec, is there anything I need to do first? How should the Input signal be prepared for the Pro-Codec? What are the optimal Host Audio Buffer sizes? What is NMR and how do I use it? How can I use the Pro-Codec to compare 24 bit audio to 16 bit audio? When does the signal need dithering prior to the Pro-Codec? What happens when working at higher sample rates? How good is the sample rate conversion? Why are some codec settings not auditionable? Why is the mp3-HD codec not auditionable? Why is there no control to trim the input level? Why is there no control to reduce the output level, when the encoding can produce a gain increase? How do I prepare a signal for lossless codecs? Why isn't there provision for metadata handling in the Pro-Codec? Why doesn't the Sonnox Fraunhofer Pro-Codec include other free codecs such as FLAC and Ogg Vorbis? Do I have to pay extra mp3 licensing fees if I use the Pro-Codec? Can you give me guidance on the circumstances when mp3 royalties are due on content?
Click here for latest Release Notes and Known Issues.
What is the purpose of the Pro-Codec? The Sonnox Fraunhofer Pro-Codec allows you to :
Which DAW's and operating systems are the Pro-Codec compatible with? We can confirm compatibility with the following : Mac OSX 10.4.11 and above (Intel Mac)
Mac PowerPC computers are not currently supported. At this time, we cannot support DAW's running in 64-bit mode. Native 64-bit versions of Sonnox plug-ins, including the Pro-Codec, are planned for release during 2011. Windows XP, Windows 7 (32-bit and 64-bit)
Currently, we cannot fully support DAW's running in 64-bit mode using VST plug-in bridges. Native 64-bit versions of Sonnox plug-ins, including the Pro-Codec, are planned for release during 2011. Before I use the Pro-Codec, is there anything I need to do first?
How should the Input signal be prepared for the Pro-Codec? The signal input to the plug-in should be prepared as if it were being written to CD or an HD Gold Master. This means that the order of processing should be:
Prior to this chain of processing, you may wish to place limiters and other plug-ins that will allow you to control peaks over 0dB that may occur at the encoding stage of the Pro-Codec.
What are the optimal Host Audio Buffer sizes? For most of the codecs, the following buffer size/sample rate combinations are optimal :
The exceptions are the HE-AAC and HE-AACv2 codecs, which require double the sizes given. However, as the CPU load of these two codecs is light, you may not notice any difference if you do not use an optimal buffer size. Using optimal buffer sizes ensures that the DSP runs most efficiently. Use of non-optimal buffer sizes will cause CPU load spiking/peaking, especially at higher sample rates. This is particularly important for HD-AAC which is a very CPU-intensive process.
I've heard that lossless files sometimes don't sound the same as the original. Is it true that "lossless" isn't lossless? The term "lossless" should strictly be applied to a file-based encoding process. If a wav file is encoded with a lossless codec, and then decoded to an output wav file, the audio content of those two wav files will be identical. It is indeed a lossless audio data transform. Note the actual files might be different, as they might contain different header (RIFF) information. Also note that two lossless codecs are not inter-changeable; mp3-HD is specific to 16-bit wav files and the HD-AAC is specific to 24-bit wav files. If the file-based situation above is replicated inside a workstation using the Sonnox Fraunhofer One very useful feature of both lossless codecs is that they include a buried lossy channel. Some newer players will support the lossless channel, however older players will be incompatible with the lossless channel and will only play back the lossy core. If so, it is entirely possible that what you are listening to will not sound identical to the original. To properly prepare the lossless file, it may be necessary to reduce the level to accommodate bitstream 'overs' on the lossy core. If the original and lossless files are at different levels, then there will be some perceptual difference. mp3-HD is lossless only to 16 bits, ie. CD-quality. Comparing the original 24-bit master to the 16-bit lossless version will show a difference. How do I use the Diff signal? The diff signal is enlightening the first time you hear it. It trains your ears to know the level at which to look for differences below the current signal, what the differences comprise of, and where those differences are, especially in transients. However, the Diff signal does not provide the full picture, as it does not account for perceptual masking effects. This is where the NMR (Noise-to-Mask Ratio) display may provide additional information on where difference may be audible.
What is NMR and how do I use it? The Noise-to-Mask Ratio (NMR) is an indication of the frequency areas where the difference between All lossy codecs will produce a very slightly different output from their input. Sometimes, if you listen to the difference signal it can sound significant; however, the very nature of a perceptual coder is that this difference signal should be inaudible (ie. masked by the output signal). A user can choose to trade off more data compression (and smaller files) against increased audibility of artefacts and codec-induced noise. Theory states that this codec-induced noise should be inaudible when the NMR indicator in the plug-in is green. Under some circumstances (codec, frequency and input signal dependent) this NMR line turns red, indicating that the induced noise is possibly audible. The training and sensitivity of the listener's ears is also an obvious and variable factor to take into account. The NMR display is a useful indication when auditioning the mp3 or AAC codec families. If there are The NMR calculation is less accurate for parametric codecs (those that use enhancements such as
How can I use the Pro-Codec to compare 24 bit audio to 16 bit audio? Working in a 24-bit session, select the mp3-HD codec, which auditions the input at 16 bits. Use the Master In button to switch between input at 16 bits and input at 24 bits.
What is the output signal? Usually the output of the plug-in is the decoded version of the encoded signal you are monitoring. You are therefore auditioning what the material will sound like when played back through, for example, an mp3 player. During an Online Encode process (writing bitstreams to disk) the output signal is switched to the input signal, in case you are bouncing down to disk. When does the signal need dithering prior to the Pro-Codec? Dithering is not required for lossy codecs, only the lossless ones. A bitmeter is displayed in the Pro-Codec's FFT display when you are using lossless codecs. This tells you whether it is necessary to dither. The signal should be dithered at 16 bits prior to the Pro-Codec if you are using the mp3-HD codec. It should be dithered at the 24 bits level if you are using the HD-AAC codec. Both lossless codecs involve a truncation. If you have 'inters' and you are using a lossless codec, then you must dither. You can either dither before the Pro-Codec or use the internal dither control. If you are using the trim dB control (ie. you are not at 0.0dB) for the lossless codecs, then you must dither with the internal dither control.
What happens when working at higher sample rates? Most codecs (with the notable exception of HD-AAC) do not support higher sample rates. At higher sample rates the plug-in will automatically down-sample as necessary before encoding/decoding, and then up-sample again for auditioning and generation of the diff signal. This offers the opportunity to create encoded versions directly, without first having to create a down-sampled version of, for example, a 192K master. The Pro-Codec will never cross-convert from a multiple of 48 to a multiple of 44.1, or vice versa. However, please note that at highest quality settings for lowest bitrates, lossy codecs can cross-convert from 44100 to 32000. The Offline Encode feature does not support down-sampling. The source file given to the Offline Encode page must already be down-sampled to whatever rate you wish to use for the encoding. Only the HD-AAC codec supports sample rates higher than 48KHz.
How good is the sample rate conversion? The plug-in's down-sampling and up-sampling algorithm is lossless to -180dB up to half sample rate. Down-sampling and up-sampling is limited to factors of 2 or 4, so does not support, for example, cross-conversion from 192K to 44.1K.
Why are some codec settings not auditionable? At the highest quality settings, some codecs at some bitrates produce an encoded output which is at a different sample rate to the input. This output is thus not auditionable without cross-converting. Instead of cross-converting, the Pro-Codec auditions a lower quality version, whilst ensuring the higher quality version will be written to file. If there is some doubt about the bitstream 'overs' being different for the highest quality settings, then use Offline Encode. However, please remember that you need to already be down-sampled to use that option. Why is the mp3-HD not auditionable? The mp3-HD codec is not capable of block-by-block encoding, so it cannot be auditioned in real time. Instead, the Pro-Codec auditions the input signal at 16-bit depth, which is theoretically an exact match for the lossless-to-16bit nature of the mp3-HD codec.
Why is there no control to trim the input level? The lack of an input control encourages correct treatment of the signal prior to the plug-in. Likewise the lack of inter-sample peak detection encourages correct monitoring and treatment of the signal prior to the plug-in, for example by using a Sonnox Oxford Limiter. The 5 trim controls indicate that any level adjustment you do make is codec-specific rather than material-specific.
Why is there no control to reduce the output level, when the encoding can produce a gain increase? The lack of an output level trim control encourages the user to lower signal levels using their mastering insert chain setup. Encoding a signal frequently produces a bitstream that has a level a few dB's higher than the original. This is due to HF component changes. Thus an input signal that peaks at 0dB may peak at more than, say, +3dB after encoding. Ideally, the best method to prepare an encoded file is one codec at a time. With this method, you use the bitstream level meters to indicate any overs due to encoding, and then you adjust the mastering chain preceding the Pro-Codec plug-in to reduce the signal level or peaks accordingly. Using this method, your auditioned level matches that of your mp3-player level, and hence you do not need an output level trim. The Pro-Codec plug-in makes it possible to encode using multiple codecs simultaneously in real time. Working in this way, it may be necessary to use the 5 trim dB controls to reduce the individual codec input levels by an appropriate amount, to be sure each version does not produce 'over' conditions when decoded. However, you may notice your master output shows 'overs' as a result of the auditioned output, which is not trimmed by the trim control - It may be necessary to reduce your host's master level control to correct this.
How do I prepare a signal for lossless codecs? Preparing a signal for lossless encoding is not quite as simple as might be expected. You cannot change the level of the lossy core without also changing the level of the lossless signal. Therefore, in order to ensure that your lossy core signal does not generate any overs during playing, we recommend the following :
Why isn't there provision for metadata handling in the Pro-Codec? The Pro-Codec will insert basic metadata into an encoded audio data stream. The basic metadata includes codec type and other codec settings. It also adds a metadata tag specifying the manufacturer. Currently the plug-in has no additional metadata editing or display features; and currently there is no related stand-alone metadata editor. There are several available third party metadata editors that can be used until the situation changes.
Why doesn't the Sonnox Fraunhofer Pro-Codec include other free codecs such as FLAC and Ogg Vorbis? The Sonnox Fraunhofer Pro-Codec is a collaboration between Sonnox Ltd. (who produced the plug-in)
Do I have to pay extra mp3 licensing fees if I use the Pro-Codec? No. It is important to distinguish between the use of codecs and the use of content produced by the codecs; and therefore it is important to note that any mp3 royalty fees are independent of which encoder is used to produce the mp3 file content. If royalty fees are due for a particular project or enterprise, they will be due irrespective of whether the Pro-Codec is used or some alternative encoder is used. There is no circumstance where using the Sonnox Fraunhofer Pro-Codec will increase licence fees or require additional fees of any sort.
Can you give me guidance on the circumstances when mp3 royalties are due on content? Both the mp3 and AAC compression routines are based on technology that is patented. At present no To summarise:
If you are in doubt whether fees are applicable for your project, please contact www.mp3licensing.com
|