The spoken article review is a process designed to provide helpful, objective feedback about recording and production quality to producers of Wikipedia spoken articles. It is also intended as a way of maintaining a minimum standard of quality for spoken articles, in order that they align with, and support, the encyclopedic goals of Wikipedia.
A list organized by subject-area can be found at
Wikipedia:Spoken articles.
Assessment
The assessment covers three areas:
- Technical quality: Examples include background noise, breath noise and pops, editing, and bitrate.
- Clarity: How understandable the recording is. Examples include enunciation and consistency of pacing.
- Accuracy: How true the recording is to the source text. Examples include misreading of words and appropriate use of voice inflection to convey the meaning intended in the source text.
Listen to the recording in full and rate the perceived quality in each area as low, medium or high. See the rating criteria below. A rating of low in one or more areas will require the spoken recording to be unlinked from its parent text article until the problem is resolved. The user who produced the article should be notified of the problem in a clear and friendly manner and given appropriate opportunity and assistance to resolve it.
Each spoken article should be judged objectively and purely on its technical merit. Try to avoid letting subjective impressions, such as your personal reaction to the subject matter of the article, influence your judgment when making an assessment.
When you have started reviewing an article, consider adding a line to the table of articles under review to prevent duplication of effort.
- If any of the criteria are low, do not post to the article and inform recorder of the issue.
Place the following template on the spoken article file talk page and fill it in with the results of the assessment:
==Spoken article quality assessment==
{{Spoken article review
|version= (pull link from media file and paste it with [-------])
|technical=
|clarity=
|accuracy=
|notes=
|reviewer=[[user:username|username]]
}} ~~~~
Rating criteria
As the review process is still under development, minor changes may be made to these criteria from time to time.
Technical quality
High
- Noise: Little to no background noise and/or breath noise; no evidence of clipping or distortion
- Edits: Edits are clean, no clicks, pops or other edit-related noise; little or no changes in audio level between edited sections; good edit timing, no cut off or run-together words
- Recording/encoding: The file is recorded with one channel and encoded to Ogg Vorbis at a nominal bitrate of up to 96 kb/s; no filters or effects have been applied that make the narration less clear; minor processing artifacts (superfluous noises caused by, for example, excessive noise-reduction or sampling rate being too low) that don't compromise voice clarity
- Licensing: All components of the recording, i.e. the narration and any sound clips or samples can be licensed under the GNU Free Documentation License, and have clearly stated references or links to the license information
- Format: The recording has at the beginning the Wikipedia introduction, and at the end the license statement, in accordance with the Spoken Wikipedia recording guidelines
Medium
- Noise: Moderate background noise and/or breath noise, but not enough to distract attention from, or obscure the narration; minor evidence of clipping or distortion
- Edits: Minor clicks, pops or other edit-related noise; minor changes in audio level between edited sections; no cut off or run-together words
- Recording/encoding: The file is recorded with one channel and encoded to Ogg Vorbis at a nominal bitrate of up to 160 kb/s; no filters or effects have been applied that make the narration less clear; moderate processing artifacts
- Licensing: All components of the recording, i.e. the narration and any sound clips or samples can be licensed under the GFDL, and have clearly stated references or links to the license information
- Format: Missing Wikipedia intro
Low
- Noise: Significant background noise and/or breath noise that distracts attention from, or reduces the intelligibility of the narration; significant evidence of clipping or distortion
- Edits: Significant clicks, pops or other edit-related noise; significant changes in audio level between edited sections; cut off or run-together words
- Recording/encoding: The file is recorded with two channels and/or encoded to Ogg Vorbis at a nominal bitrate higher than 160 kb/s; filters or effects have been applied that make the narration less clear; significant processing artifacts
- Licensing: Components of the recording, i.e. the narration and any sound clips or samples cannot be licensed under the GFDL, and/or do not have clearly stated references or links to the license information
- Format: Missing Wikipedia intro and/or license statement
Clarity
High
- Enunciation: Enunciation is consistently clear, with easily understood speech
- Pacing: Pacing, or the speed at which the text is read, is consistently even throughout the recording
Medium
- Enunciation: Enunciation is mostly clear, with only minor problems that do not significantly or consistently affect the ability to understand the speech
- Pacing: Minor variations in pacing
Low
- Enunciation: Significant enunciation problems which reduce the ability to understand the speech
- Pacing: Significant variations in pacing
Accuracy
If the originator of the recording hasn't provided a link to the historical version of the article that the recording is based on, make a polite request on the user's talk page, and give the user a reasonable period of time to respond; if, after this time, the user still hasn't provided the link, the recording can't be assessed for accuracy and should be given a default rating of "Low"
High
- Fidelity: Fidelity to the core or main body of source text is high; text is read out substantially as it was written (with the exception of minor alterations to clarify the meaning of text that does not translate well to spoken language), with minimal misreads; what misreads there are do not significantly alter the meaning of the source text; no adlibs, or unnecessary speech that does not derive from, and is not related to the source text
- Inflection: Voice inflection is used in a conventional manner that aids in conveying the meaning of the source text, and not omitted entirely or over-emphasised
Medium
- Fidelity: Minor deviations from the source text and/or a moderate number of misreads that do not significantly compromise the meaning of the source text; no significant adlibs
- Inflection: Voice inflection is not consistently used in a manner that aids in conveying the meaning of the source text
Low
- Fidelity: Deviations from the source text and/or misreads that significantly compromise the meaning of the source text; significant adlibs
- Inflection: Voice inflection is dramatically under or over-emphasised, and/or not consistently used in a manner that aids in conveying the meaning of the source text
The spoken article review is a process designed to provide helpful, objective feedback about recording and production quality to producers of Wikipedia spoken articles. It is also intended as a way of maintaining a minimum standard of quality for spoken articles, in order that they align with, and support, the encyclopedic goals of Wikipedia.
A list organized by subject-area can be found at
Wikipedia:Spoken articles.
Assessment
The assessment covers three areas:
- Technical quality: Examples include background noise, breath noise and pops, editing, and bitrate.
- Clarity: How understandable the recording is. Examples include enunciation and consistency of pacing.
- Accuracy: How true the recording is to the source text. Examples include misreading of words and appropriate use of voice inflection to convey the meaning intended in the source text.
Listen to the recording in full and rate the perceived quality in each area as low, medium or high. See the rating criteria below. A rating of low in one or more areas will require the spoken recording to be unlinked from its parent text article until the problem is resolved. The user who produced the article should be notified of the problem in a clear and friendly manner and given appropriate opportunity and assistance to resolve it.
Each spoken article should be judged objectively and purely on its technical merit. Try to avoid letting subjective impressions, such as your personal reaction to the subject matter of the article, influence your judgment when making an assessment.
When you have started reviewing an article, consider adding a line to the table of articles under review to prevent duplication of effort.
- If any of the criteria are low, do not post to the article and inform recorder of the issue.
Place the following template on the spoken article file talk page and fill it in with the results of the assessment:
==Spoken article quality assessment==
{{Spoken article review
|version= (pull link from media file and paste it with [-------])
|technical=
|clarity=
|accuracy=
|notes=
|reviewer=[[user:username|username]]
}} ~~~~
Rating criteria
As the review process is still under development, minor changes may be made to these criteria from time to time.
Technical quality
High
- Noise: Little to no background noise and/or breath noise; no evidence of clipping or distortion
- Edits: Edits are clean, no clicks, pops or other edit-related noise; little or no changes in audio level between edited sections; good edit timing, no cut off or run-together words
- Recording/encoding: The file is recorded with one channel and encoded to Ogg Vorbis at a nominal bitrate of up to 96 kb/s; no filters or effects have been applied that make the narration less clear; minor processing artifacts (superfluous noises caused by, for example, excessive noise-reduction or sampling rate being too low) that don't compromise voice clarity
- Licensing: All components of the recording, i.e. the narration and any sound clips or samples can be licensed under the GNU Free Documentation License, and have clearly stated references or links to the license information
- Format: The recording has at the beginning the Wikipedia introduction, and at the end the license statement, in accordance with the Spoken Wikipedia recording guidelines
Medium
- Noise: Moderate background noise and/or breath noise, but not enough to distract attention from, or obscure the narration; minor evidence of clipping or distortion
- Edits: Minor clicks, pops or other edit-related noise; minor changes in audio level between edited sections; no cut off or run-together words
- Recording/encoding: The file is recorded with one channel and encoded to Ogg Vorbis at a nominal bitrate of up to 160 kb/s; no filters or effects have been applied that make the narration less clear; moderate processing artifacts
- Licensing: All components of the recording, i.e. the narration and any sound clips or samples can be licensed under the GFDL, and have clearly stated references or links to the license information
- Format: Missing Wikipedia intro
Low
- Noise: Significant background noise and/or breath noise that distracts attention from, or reduces the intelligibility of the narration; significant evidence of clipping or distortion
- Edits: Significant clicks, pops or other edit-related noise; significant changes in audio level between edited sections; cut off or run-together words
- Recording/encoding: The file is recorded with two channels and/or encoded to Ogg Vorbis at a nominal bitrate higher than 160 kb/s; filters or effects have been applied that make the narration less clear; significant processing artifacts
- Licensing: Components of the recording, i.e. the narration and any sound clips or samples cannot be licensed under the GFDL, and/or do not have clearly stated references or links to the license information
- Format: Missing Wikipedia intro and/or license statement
Clarity
High
- Enunciation: Enunciation is consistently clear, with easily understood speech
- Pacing: Pacing, or the speed at which the text is read, is consistently even throughout the recording
Medium
- Enunciation: Enunciation is mostly clear, with only minor problems that do not significantly or consistently affect the ability to understand the speech
- Pacing: Minor variations in pacing
Low
- Enunciation: Significant enunciation problems which reduce the ability to understand the speech
- Pacing: Significant variations in pacing
Accuracy
If the originator of the recording hasn't provided a link to the historical version of the article that the recording is based on, make a polite request on the user's talk page, and give the user a reasonable period of time to respond; if, after this time, the user still hasn't provided the link, the recording can't be assessed for accuracy and should be given a default rating of "Low"
High
- Fidelity: Fidelity to the core or main body of source text is high; text is read out substantially as it was written (with the exception of minor alterations to clarify the meaning of text that does not translate well to spoken language), with minimal misreads; what misreads there are do not significantly alter the meaning of the source text; no adlibs, or unnecessary speech that does not derive from, and is not related to the source text
- Inflection: Voice inflection is used in a conventional manner that aids in conveying the meaning of the source text, and not omitted entirely or over-emphasised
Medium
- Fidelity: Minor deviations from the source text and/or a moderate number of misreads that do not significantly compromise the meaning of the source text; no significant adlibs
- Inflection: Voice inflection is not consistently used in a manner that aids in conveying the meaning of the source text
Low
- Fidelity: Deviations from the source text and/or misreads that significantly compromise the meaning of the source text; significant adlibs
- Inflection: Voice inflection is dramatically under or over-emphasised, and/or not consistently used in a manner that aids in conveying the meaning of the source text