op het forum van Reason is er sinds enige tijd een heftige outburst van discussies over de geluidskwaliteit van Reason..... een paar freaks zijn eens flink met freq analysers enzo tekeer gegaan en dachten rare dingen te zien in het geluid van een onbewerkte ongepitchede sample..... iedereen werd natuurlijk gek, en nu is er een officieel statement van de Props zelf..misschien interessant voor de reason gebruikers hiero..... ikzelf ben van mening dat Reason, net als fruityloops etc. z'n volgens sommigen "typische" geluid vnl krijgt als je de standaard presets en samples gebruikt, en het feit dat er meer headroom is om clippng te voorkomen en dat het geluid dus zachter staat, vinden mensen het vaak ook "slechter" klinken...maargoed.....ieder z'n eigen mening..hier de statement:
Hi there board members!
There's been a discussion on this board regarding Reason's pitch interpolation algorithm. The topic has been whether Reason doesn't unnecessarily alter the sound when playing back samples at nominal pitch (the original pitch the sample was recorded at).
First of all, let us once and for all state that - yes! it is true that our pitch interpolation mechanism is at work when in fact it needn't be, and that this has a very, very slight (unperceivable) effect on the pitch of a sound when it shouldn't (the pitch change is so small that no human ear can detect it). And - yes! for a sine wave with a pitch of 15kHz this results in a fluctuation in level of something between 1 and 3dB, occurring cyclically over a period of around 1.5 seconds. When this was described to us initially we understood it as a problem affecting pitch only, when in fact it turned to rather be spectrally/amplitude related.
Now, let's try to put this information into context.
* This discussion is only about when a sample is played back at *nominal pitch*. We've said this before but we think it is worth repeating. What criteria need to be fulfilled for a sample to be played back at nominal pitch in one of Reason's samplers? The following:
- The recording and playback sample rate must be identical. If for example the sample was recorded at 44.1kHz and your sound card was set to 48kHz, this discussion has no bearing at all.
- The sample must be played back with the same key (root note) as the original. In other words, this only applies to occasional notes in a melody or one note a time in a chord, but more often with for example REX files.
- There must be not even the slightest tuning adjustment in the settings on the sampler, to get it to play in tune with the rest of the music.
- There must be no pitch bend, vibrato, pitch envelope or other pitch adjustment in effect. If there is, you can stop reading here.
So, if in a given moment, the playback of a certain sample lives up to all the criteria above, yes, we do make changes to the sample that we shouldn't. In all other cases, we do the right thing, as all other sample playback devices, we use DSP methods to modify the sample to the required pitch.
Now, let's look at the impact this has on the total sound quality.
* The phenomenon we are discussing is not equal over the entire frequency range. The discussion has been regarding a 15kHz sine wave, a sound almost outside the human hearing range. At more normal pitches, any amplitude fluctuations created by the pitching algorithm are very, very, very small. For a sine wave at 440 Hz (standard "A") it is (at least) less than one hundredth of a dB, which is far outside anything recognizable by the human ear.
* The energy levels at high frequencies such as 15kHz are normally quite low. Sorry if we have to be a bit theoretic here. Let's take the 440Hz example again, but this time we look at a sawtooth wave rather than a sine. The 34th partial of a 440Hz sawtooth occurs at around 15kHz, but has a relative level of around 3% of the fundamental. This means that any fluctuation in this range is by definition not a level change, but a timbral one. However, the change in timbre is extremely slight, we have here at Propeller headquarters set up the most extreme and brutal conditions we could think of (a static isolated pulse wave and a completely fixed listening position), and people here can hear it, at least those with great hearing. With any normal musical sample it is our sincere assessment that this is undetectable by the human ear.
To summarize: Do the phenomena described on this board exist? Yes. Will they be corrected in future versions of the program? Very likely. Do they have any significant impact on the overall sound quality of Reason, whatsoever? Here the answer is a very clear and loud - No. Does this have any perceivable effect on an overall mix of music made in Reason? - Not a chance in hell.
Let us conclude by pointing out a couple of ironies in this discussion.
As stated above, all samplers - be they hardware or software - need to use DSP algorithms to pitch samples. And the fact is that the quality of the methods used in Reason are among the best in this industry. We encourage you to compare us critically to the competition because we know that we will come out favorably. However, when doing such comparisons it is easy to fall in traps that make one product or another come out less favorably than it otherwise should, so make sure you understand exactly what you are doing before you judge any of our competition too hardly.
The other irony is that we perceive ourselves as straight talkers and very open when it comes to deficiencies in our products, again compared to others. We haven't in this case - or in any other - ever mislead our users, nor do we intend to in the future. We are and will remain proud of providing software with a higher technical quality and superior quality of result compared those we compete with, be they hardware or software manufacturers.
Last, if anyone still thinks that the items discussed above constitute an obstacle in the strive for audio excellence, here's a tip. Please note that this method is relevant regardless of the above discussion and applies not only to our products, but to all samplers. If you have any samples in your production that are truly critical (played in complete isolation for example) you can proceed as follows:
1. Use a separate professional sample editor to resample the file at a higher frequency, for example 96 or 88.2 kHz.
2. Load the sample back into Reason (or other sampler). There is no need to adjust the pitch settings for the sample, or to adjust the playback frequency in Reason's preferences.
The reason why this works is that sample editors use off-line resampling algorithms that are much, much slower but of higher quality than any real-time unit (be it hardware or software) can ever implement.
Good luck with your music!
Ernst Nathorst-Böös and Tage Widsell
Propellerhead Software
Hi there board members!
There's been a discussion on this board regarding Reason's pitch interpolation algorithm. The topic has been whether Reason doesn't unnecessarily alter the sound when playing back samples at nominal pitch (the original pitch the sample was recorded at).
First of all, let us once and for all state that - yes! it is true that our pitch interpolation mechanism is at work when in fact it needn't be, and that this has a very, very slight (unperceivable) effect on the pitch of a sound when it shouldn't (the pitch change is so small that no human ear can detect it). And - yes! for a sine wave with a pitch of 15kHz this results in a fluctuation in level of something between 1 and 3dB, occurring cyclically over a period of around 1.5 seconds. When this was described to us initially we understood it as a problem affecting pitch only, when in fact it turned to rather be spectrally/amplitude related.
Now, let's try to put this information into context.
* This discussion is only about when a sample is played back at *nominal pitch*. We've said this before but we think it is worth repeating. What criteria need to be fulfilled for a sample to be played back at nominal pitch in one of Reason's samplers? The following:
- The recording and playback sample rate must be identical. If for example the sample was recorded at 44.1kHz and your sound card was set to 48kHz, this discussion has no bearing at all.
- The sample must be played back with the same key (root note) as the original. In other words, this only applies to occasional notes in a melody or one note a time in a chord, but more often with for example REX files.
- There must be not even the slightest tuning adjustment in the settings on the sampler, to get it to play in tune with the rest of the music.
- There must be no pitch bend, vibrato, pitch envelope or other pitch adjustment in effect. If there is, you can stop reading here.
So, if in a given moment, the playback of a certain sample lives up to all the criteria above, yes, we do make changes to the sample that we shouldn't. In all other cases, we do the right thing, as all other sample playback devices, we use DSP methods to modify the sample to the required pitch.
Now, let's look at the impact this has on the total sound quality.
* The phenomenon we are discussing is not equal over the entire frequency range. The discussion has been regarding a 15kHz sine wave, a sound almost outside the human hearing range. At more normal pitches, any amplitude fluctuations created by the pitching algorithm are very, very, very small. For a sine wave at 440 Hz (standard "A") it is (at least) less than one hundredth of a dB, which is far outside anything recognizable by the human ear.
* The energy levels at high frequencies such as 15kHz are normally quite low. Sorry if we have to be a bit theoretic here. Let's take the 440Hz example again, but this time we look at a sawtooth wave rather than a sine. The 34th partial of a 440Hz sawtooth occurs at around 15kHz, but has a relative level of around 3% of the fundamental. This means that any fluctuation in this range is by definition not a level change, but a timbral one. However, the change in timbre is extremely slight, we have here at Propeller headquarters set up the most extreme and brutal conditions we could think of (a static isolated pulse wave and a completely fixed listening position), and people here can hear it, at least those with great hearing. With any normal musical sample it is our sincere assessment that this is undetectable by the human ear.
To summarize: Do the phenomena described on this board exist? Yes. Will they be corrected in future versions of the program? Very likely. Do they have any significant impact on the overall sound quality of Reason, whatsoever? Here the answer is a very clear and loud - No. Does this have any perceivable effect on an overall mix of music made in Reason? - Not a chance in hell.
Let us conclude by pointing out a couple of ironies in this discussion.
As stated above, all samplers - be they hardware or software - need to use DSP algorithms to pitch samples. And the fact is that the quality of the methods used in Reason are among the best in this industry. We encourage you to compare us critically to the competition because we know that we will come out favorably. However, when doing such comparisons it is easy to fall in traps that make one product or another come out less favorably than it otherwise should, so make sure you understand exactly what you are doing before you judge any of our competition too hardly.
The other irony is that we perceive ourselves as straight talkers and very open when it comes to deficiencies in our products, again compared to others. We haven't in this case - or in any other - ever mislead our users, nor do we intend to in the future. We are and will remain proud of providing software with a higher technical quality and superior quality of result compared those we compete with, be they hardware or software manufacturers.
Last, if anyone still thinks that the items discussed above constitute an obstacle in the strive for audio excellence, here's a tip. Please note that this method is relevant regardless of the above discussion and applies not only to our products, but to all samplers. If you have any samples in your production that are truly critical (played in complete isolation for example) you can proceed as follows:
1. Use a separate professional sample editor to resample the file at a higher frequency, for example 96 or 88.2 kHz.
2. Load the sample back into Reason (or other sampler). There is no need to adjust the pitch settings for the sample, or to adjust the playback frequency in Reason's preferences.
The reason why this works is that sample editors use off-line resampling algorithms that are much, much slower but of higher quality than any real-time unit (be it hardware or software) can ever implement.
Good luck with your music!
Ernst Nathorst-Böös and Tage Widsell
Propellerhead Software
