What is aımfi™?

aımfi™ is the result of over twelve years of independent research and development in the field of digital audio recording. It relies on new techniques of restoration and preservation to achieve the same spacious, well-defined, true-to-life sound generally associated with pure analog reproduction, with all the convenience of digital audio—even with sources that were digitized as early as the 1970s.

Restoration is necessary when working from digital sources. This is achieved by means of corrective signal processing that aims to reverse as accurately as possible the loss of integrity caused by previous digital components in the audio chain (particularly analog-to-digital converters). The parameters of the signal processing to be applied are determined on the basis of new findings related to digital audio combined with advanced signal analysis and acute critical listening. Once the signal has been restored, or when working from a pure analog source, aımfi™ ensures the preservation of the integrity of the signal throughout the remainder of the digital audio chain.

aımfi™ is researched and developed entirely at 2-bit Masters. All mastering projects handled at 2-bit Masters benefit from aımfi™ restoration and preservation.

What was the motivation behind this research and development project?

There is a fairly large consensus that many digital recordings—most obviously early ones from the 1970s to the 1990s, but to a certain extent contemporary ones also—sound flatter and more artificial to the human ear than analog recordings. Yet the negative impact of well-known issues such as jitter, aliasing, and quantization noise was already dealt with properly decades ago and is effectively negligible since the mid-1980s, if not earlier.

The hypothesis that digital recordings are practically perfect and that we only prefer analog recordings because we find their weaknesses more euphonic is misguided at best. If it were true, these euphonic qualities should be fully retained when an existing analog recording is converted to digital, which is rarely if ever the case. Conversely, transferring a digital recording to an analog medium does not automatically make it as euphonic as a fully analog recording, which further invalidates this hypothesis.

Some sound engineers have been trying to compensate for the relative lack of depth and the artificiality of digital recordings with various types of processors at the recording, mixing, or mastering stage: exciters, equalizers, compressors, image enhancers, etc. While this approach can have pleasant results, it does not increase the fidelity of the reproduction. It does in fact reduce it—by merely replacing what is lost in the digitization process with something else that was not present in the source signal.

It seemed therefore that a better understanding of the actual causes of the relatively flat and artificial sound of digital recordings might make it possible to prevent these issues by increasing the fidelity of such recordings.

What were the main findings?

The research showed that it was not only possible to prevent these issues, but that the fidelity of existing digital recordings impacted by these issues—going back all the way to the 1970s—can in many cases be restored with corrective signal processing, resulting in a depth and naturalness comparable to that of fully analog recordings—albeit without the typical shortcomings of the latter (wow and flutter, hiss, dropouts, etc.).

Most current analog-to-digital converters—and some earlier ones—are indeed practically perfect within the limits of human hearing, but their performance can be altered under certain circumstances. Other converters are inherently problematic. Luckily, there does not appear to be any major issue with current digital-to-analog converters in general, and so consumers won’t have to make any changes to their setups to enjoy the benefits of aımfi™-processed releases.

It should perhaps be pointed out that none of the findings in this research invalidates mainstream scientific theory on signal sampling and human hearing. The issues at hand are practical and specific to implementations of analog-to-digital conversion in the field of audio engineering. If anything, the possibility to re-create analog-like reproduction from early digital sources confirms that so-called standard resolution is sufficient to convey the extent of what we can hear, as has always been supported by mainstream scientific theory in these fields.

What were the main difficulties that had to be overcome?

The main difficulty, and probably the main reason why such research does not appear to have been attempted—or at least successfully completed—before, was to identify the actual issues at hand. Discarding such well-researched and well-understood limitations of digital audio as jitter, aliasing, and quantization noise and focusing on a scientifically sound understanding of human hearing proved the best approach to formulating working hypotheses and getting research underway.

On the technical side of things, one of the main findings is that some converters may act as expected in a typical testing or measurement setup but behave somewhat differently in an actual recording chain. This makes it difficult to measure the impact of this unexpected behavior on the audio signal itself and workarounds had to be found.

It’s also not practical to test and measure every existing converter, particularly early ones that are currently very expensive and hard to find in a fully functional state and without any modification. Relying on existing digital transfers of otherwise preserved analog recordings made on these different systems proved the only practical approach, but it made it necessary to properly identify what alterations resulted from the analog-to-digital conversion itself and what may have occurred elsewhere in the audio chain—sometimes deliberately on the part of the engineer making the transfer. Unfortunately, the information in some record companies’ archives appears to be incomplete—if not erroneous—in this respect, which further complicates this task.

Another difficulty that has not yet been fully dealt with is the need to understand how digital equipment was used in actual studio settings, especially from the 1970s to the mid-1990s. The practical limitations of these early digital systems often had an impact on the way recording, mixing, and mastering chains were set up, with different approaches being adopted by different studios. Because the way the system is set up can have an impact on the performance of the converters, gaining a better understanding of studio practices at the time will make it easier to predict how these recordings have been impacted and what kind of restoration may be necessary, which should help streamline and shorten the restoration process.

It should also be mentioned that in the case of fully digital recordings, the sonic impact of the converters was already audible in the studio, thereby influencing artistic and engineering decisions at the recording and mixing stages. When restoring such sources, it is necessary to compare the restored version with the orignal and to make sure that the character of the original recording has not been altered by the restoration process—or to compensate for these alterations with appropriate adjustments at the mastering stage (even when such adjustments are necessary, fidelity to the source signal is greatly improved, and the associated analog-like rendition can still be achieved).

What are the practical applications of these findings?

The most obvious one is the ability to restore existing digital recordings, as well as existing digital transfers of analog recordings that may no longer be available as such, preventing the production of new digital transfers (because the analog sources have since been discarded, damaged, or lost). Any digital recording (or digital transfer of an analog recording) mastered at 2-bit Masters will benefit from this restoration process.

This is not only useful when preparing digital releases of these recordings but also for any analog release (on vinyl or tape) made from digital sources. In some cases also, analog sources may need to be cleaned up in the digital domain (to fix issues such as hiss, dropouts, or wow and flutter) even for an analog release, and 2-bit Masters can guarantee that this process will be done without affecting the depth and naturalness of the recording through its digitization.

Another less obvious application, one which was not anticipated, is the ability to identify analog transfers of early digital recordings as such. In many instances, record companies hold digital and analog copies of the same recording in their archives, seemingly with insufficient documentation, and may wrongly assume that the analog copy is the master while the digital one is a backup. In reality many early digital masters were backed up to analog tape because of the risk of competing digital formats becoming obsolete, or to send these recordings to mastering studios not equipped to work from the digital master (even analog copies made from digital backups of original analog recordings have been identified in some instances).

In such cases the digital master, restored with aımfi™, should be preferred to the analog copy as the main source to prepare new releases. In most cases, 2-bit Masters will have the ability to accurately determine whether the analog copy is the master or a copy from an early digital source.

How long does it take to restore a digital recording?

Most recordings (or collections of recordings made on the same setup and digitized with the same converter) take about two days of work to be restored prior to the mastering session itself. There are however still cases where a recording presents unprecedented challenges and requires further research, especially earlier ones, which can take several days or even weeks—if it is necessary to locate other copies of the recording, for instance (this additional time will not be charged to 2-bit Masters’ customers). Projects have to be accepted and planned to accommodate such challenges, which can unfortunately lead to relatively long deadlines.

As mentioned earlier, further research will lead to a better understanding of how digital equipment has been used since the 1970s (as will each project that gets completed at 2-bit Masters), which should greatly reduce the risk of encountering unexpected issues. The average duration of projects should then become easier to predict accurately, making it possible to complete more projects within shorter deadlines.

What further developments can be expected?

As the research advances and more and more projects are completed at 2-bit Masters, it will hopefully be possible to design more advanced software that combines and partly automates the analysis and signal processing required to restore digital sources. Once that becomes a reality, it should also be possible to train other engineers to perform this kind of restoration—here at 2-bit Masters but also at other companies who may be interested in offering this service.

Another possible area of development may be to design software that can re-create the sound of classic early analog-to-digital converters, similar to popular plugins that mimic the sound of analog tape or vinyl records. In some sub-genres of pop and rock music that developed around that time (such as indie rock, post-punk, thrash metal, hip-hop, or electronic dance music), this kind of sound is sometimes sought after. The ability to re-create it in a more controlled manner should appeal to producers currently working in these sub-genres or looking to incorporate similar sounds into their work.

At this time it is not yet possible to give a timeline with regard to these further developments.

With which digital audio file formats and media is aımfi™ compatible?

aımfi™ is handled entirely on the production side and therefore it allows for delivery in every existing digital audio file format and on every medium, both lossless and lossy, at any resolution (whether PCM or DSD). Conversion to the desired format or medium takes place at the final stage of the processing chain, and the benefits of aımfi™ are heard by the end listeners regardless of playback hardware.

aımfi™ digital masters can also be used to produce analog releases (on vinyl or tape). Mono, stereo, and surround (4.0, 5.1, etc.) are also all supported.

Where can I hear samples of projects mastered with aımfi™?

aımfi™ was introduced very recently. Samples should be available on this website some time in late 2025 or early 2026 (as soon as copyright matters have been cleared). The first commercial releases benefiting from aımfi™ are not expected to come out before the second half of 2026 (2-bit Masters is not directly involved in the release process of these projects).